# Machine House Media — public-page text This reading copy is generated from the canonical pages below. Each source URL retains its layout, imagery and context. Dates and results belong to the periods stated in the source; this file does not update them. --- Source: https://machinehouse.media/ AI-native content studio · Bengaluru # We build the machine behind your content. From the first idea to the audience it deserves. Strategy, storytelling, production and growth—with data behind every creative call. Audience analysis and forecasting sharpen the direction. Human taste shapes the work. Eight years inside India’s biggest creator businesses. [Build something with us ↗ ](https://calendly.com/rishwajeet/intro-call)[See the work ↓ ](https://machinehouse.media/#cases) [100M+ Views across client work ](https://machinehouse.media/rishwajeet#record)[600+ Episodes shipped ](https://machinehouse.media/rishwajeet#record)[15M+ Audience managed ](https://machinehouse.media/rishwajeet#record) Rishwajeet’s career record · June 2026 CONTENT INTELLIGENCE [Explore the stages ↓ ](https://machinehouse.media/#machine-detail) Audience signals Published work Editorial memory THE DECISION ENGINE ## Taste × intelligence. What should we make next? RJ / EDITORIAL 01 / DIRECTION Strategy & prediction 02 / DISCOVERY Guests & research 03 / STORY Scripts & questions 04 / CRAFT Production & editing 05 / ATTENTION Packaging & short-form 06 / AUDIENCE Distribution & growth Results inform the next brief. 01 / STRATEGY & PREDICTION Compare the proposed topics and guests with earlier releases. Read what viewers watched and said, then choose the idea and the question it needs to answer. [Explore content strategy ↗ ](https://machinehouse.media/services/content-strategy) Selected work across Machine House & the Breakout team [Explore the full portfolio ↓ ](https://machinehouse.media/#clients) [Ankur Warikoo ](https://www.youtube.com/@warikoo)[Bhuvan Bam ](https://www.youtube.com/@BBKiVines)[The Ranveer Show ](https://www.youtube.com/@ranveerallahbadia)[Masoom Minawala ](https://www.instagram.com/masoomminawala)[Royal Enfield ](https://www.instagram.com/royalenfield)[Paytm ](https://www.instagram.com/paytm)[Film Companion ](https://www.youtube.com/@FilmCompanionStudios)[Zero1 by Zerodha ](https://www.youtube.com/@Zero1ByZerodha) Where we come in ## Everything it takes to make content matter. Work with us on a specific part of your content, or bring in the team for the full operation. ### [Strategy & creative direction ](https://machinehouse.media/services/content-strategy) We review your channel and audience, develop the subjects and formats, and plan the next slate. You work directly with Rishwajeet on the choices and script reviews. [Content strategy & consulting ↗ ](https://machinehouse.media/services/content-strategy) ### [Content, start to finish ](https://machinehouse.media/services/content-production) Research, scripts, guest preparation, recording, editing and finishing for shows, podcasts, films and short-form. We agree who owns each stage around your existing team. [End-to-end content production ↗ ](https://machinehouse.media/services/content-production)[Podcast production ↗ ](https://machinehouse.media/services/podcast-production) ### [Distribution & growth ](https://machinehouse.media/services/youtube-growth) Titles, thumbnail concepts, trailers, shorts and publishing. After release, we review the videos and audience response to decide what to make or change next. [YouTube growth & channel management ↗ ](https://machinehouse.media/services/youtube-growth) ### [The intelligence underneath ](https://machinehouse.media/services/ai-content-systems) We build a working record of your content: topics, guests, hooks, packaging, retention and audience response. Custom AI tools make that history useful in the next research brief, script, title and editorial review. [AI content systems ↗ ](https://machinehouse.media/services/ai-content-systems) Explore the full scope of our work in [services ↗ ](https://machinehouse.media/services) Selected work ## Big ideas. Machines to match. How we chose the next guest, found the subject after a hit, and took research through to a finished film. Ankur Warikoo / Money Matters ### Ankur Warikoo Money Matters #### Built for human stories. Engineered for scale. Rishwajeet helped launch Money Matters, from strategy through production and distribution, then built its AI guest-selection system: thousands of financial stories screened and scored, with the team making the final call. Days of work. Under an hour. Thousands of submissions · two-stage intelligence · human editorial sign-off Built in four days. Part of the show’s weekly operation—and still doing the work two years on. [Read the Money Matters case study ↗ ](https://machinehouse.media/work/ankur-warikoo) 40M+ show views 2× Apple Best Podcast 4 days to build the AI workflow Masoom Minawala / The show ### Masoom Minawala The Masoom Minawala Show #### The systems behind the whole show. Guest selection, research, questionnaires, trailers, titles, thumbnails and Shorts. We handle the editorial work across The Masoom Minawala Show through systems we’ve built for it, with Rishwajeet working directly with Masoom. From the guest list to the final Short. Research becomes the questionnaire. The recording becomes trailers, packaging and Shorts briefs. What we learn after release shapes the next episode. Built around Masoom’s voice, her audience and the show we’re making together. [Read the Masoom Minawala case study ↗ ](https://machinehouse.media/work/masoom-minawala) 4.66M reported views +102.8% watch hours +64.7% net subscriber adds Masoom: 1 January–3 September, 2025 vs 2026, long-form show catalogue. Money Matters: historical case record, June 2026. Full scope in each case study. Roshan Kotla / The breakthrough ### Roshan Kotla Career & technology essays #### A small channel. Three hits in a row. Roshan’s channel had a median of around 100 views per video. We helped turn his experience in software and hiring into a viral IT-industry essay—and followed it with two more hits. 791K. 93K. 91K. A viral breakout. Then two more essays above 90,000 views. Machine House developed the research, thesis, scripts and packaging for all three. Reading 443 comments on the first hit helped us choose the next subjects. Roshan brought his lived experience and delivery. [Read the Roshan Kotla case study ↗ ](https://machinehouse.media/work/roshan-kotla) 975K+ combined views ~100 median views before the breakout Views as of 7 September 2026 Scaler Technology & education ### The premise made the difference. We developed research, scripts, production briefs and packaging for 17 Scaler releases. The leading essay, about the next valuable software role, passed 91K views. The full slate passed 150K. Public views · 7 September 2026 [Read the Scaler case study ↗ ](https://machinehouse.media/work/scaler) Rocket Health Health & culture ### A health subject became a story. Rocket Health’s protein film reached 72K views. We developed the research, script, visual direction and briefs for the host and crew, working with Studio 73 on the production. Public views · 7 September 2026 [Read the Rocket Health case study ↗ ](https://machinehouse.media/work/rocket-health) [ ### The KariGhars Packaging, scripting and production. A 478% watch-time increase in the first month, reported June 2026. Inside the engagement ↗ ](https://machinehouse.media/work/the-karighars)[ ### Indian Silicon Valley Three agencies across the content operation. Weekly views grew from 30K to 150K in the historical case record. Inside the engagement ↗ ](https://machinehouse.media/work/indian-silicon-valley) Inside an engagement ## A steady rhythm. Time for the work. Most ongoing work runs in quarterly cycles. We keep a limited active roster so strategy and editorial reviews get real attention. - WEEK 01 ### Find the actual problem. Audit the channel, audience and operation. Agree what the next 90 days need to change. - WEEKS 02–04 ### Build the plan and the workflow. Set the editorial direction, first slate, production responsibilities and useful tooling. - EVERY WEEK ### Make, review, learn. Strategy call, script reviews and a numbers note. Production and publishing follow the agreed schedule. - EACH QUARTER ### Decide what comes next. Continue, change or stop formats based on what the work has taught us. ### What we learn changes the next brief. After Roshan’s IT essay broke out, we read all 443 comments. The next commission came from what people were worried about: they had studied and earned the degree, but the job they expected was no longer certain. That became an essay about graduates struggling to find work. Roshan had years of hiring experience to draw on; we developed the research, script and package with him. The follow-up passed 93K views. The full case shows the three releases and the decisions behind them. Rishwajeet Singh Founder · Strategist · Bengaluru Who you work with ## The person behind the machine. I’m Rishwajeet, a content strategist and creative technologist. I’ve spent eight years working with creators and brands—developing stories, leading teams and building the operations behind their content. That work takes me from scripts and shoots to the software we use to make them. I founded Machine House to bring that work together. I also co-founded [The Breakout Company ](https://breakout.company/), where my partners and I run a full-service content agency. [More about my work ↗ ](https://machinehouse.media/rishwajeet)[My personal website ↗ ](https://rishwajeetsingh.com/) The wider body of work ## Creators, shows and ambitious brands. Experience across Machine House and the Breakout team, spanning finance, culture, education, lifestyle and business. Credits below identify the relevant body of work. [Ankur Warikoo Rishwajeet / Machine House ↗ ](https://www.youtube.com/@warikoo)[Astro Arun Pandit Rishwajeet / Machine House ↗ ](https://www.youtube.com/@AstroArunPandit)[Masoom Minawala Rishwajeet / Machine House ↗ ](https://www.instagram.com/masoomminawala)[Roshan Kotla Rishwajeet / Machine House ↗ ](https://www.youtube.com/@RoshanKotla)[The KariGhars Rishwajeet / Machine House ↗ ](https://www.instagram.com/the_karighars)[Scaler Rishwajeet / Machine House ↗ ](https://www.youtube.com/@Scaler)[Indian Silicon Valley Rishwajeet / Machine House ↗ ](https://www.youtube.com/@IndianSiliconValley)[Rocket Health Rishwajeet / Machine House ↗ ](https://www.instagram.com/rocket.health)[Freedom With AI Rishwajeet / Machine House ↗ ](https://www.instagram.com/freedom_with_ai/)[Avinash Mada Rishwajeet / Machine House ↗ ](https://www.instagram.com/meetavinash/)[Bhuvan Bam Breakout team ↗ ](https://www.youtube.com/@BBKiVines)[The Ranveer Show Breakout team ↗ ](https://www.youtube.com/@ranveerallahbadia)[BeerBiceps Breakout team ↗ ](https://www.youtube.com/@BeerBiceps)[Paytm Breakout team ↗ ](https://www.instagram.com/paytm)[Royal Enfield Breakout team ↗ ](https://www.instagram.com/royalenfield)[Rhea Chakraborty Breakout team ↗ ](https://www.instagram.com/rhea_chakraborty)[bigbasket Breakout team ↗ ](https://www.instagram.com/bigbasketcom)[BigBrainCo Breakout team ↗ ](https://www.youtube.com/@raaazofficial)[Humans of Bombay Breakout team ↗ ](https://www.instagram.com/officialhumansofbombay)[WTF · Nikhil Kamath Breakout team ↗ ](https://www.youtube.com/@nikhil.kamath)[Vaibhav Sisinty Breakout team ↗ ](https://www.instagram.com/vaibhavsisinty)[Zomato Breakout team ↗ ](https://www.instagram.com/zomato)[Film Companion Breakout team ↗ ](https://www.youtube.com/@FilmCompanionStudios)[Neeraj Arora Breakout team ↗ ](https://www.youtube.com/@neerajarora)[Zero1 by Zerodha Breakout team ↗ ](https://www.youtube.com/@Zero1ByZerodha)[LearnApp Breakout team ↗ ](https://www.youtube.com/@PrateekSinghLearnApp)[Tanay Pratap Breakout team ↗ ](https://www.youtube.com/@tanaypratap)[Scott Carney Breakout team ↗ ](https://www.youtube.com/@sgcarney)[The Hollywood Reporter India Breakout team ↗ ](https://www.youtube.com/@HollywoodReporterIndia)[CoinDCX Breakout team ↗ ](https://www.youtube.com/@CoinDCX)[Casey Adams Breakout team ↗ ](https://www.instagram.com/casey)[Sai Kumar Bysani Breakout team ↗ ](https://www.instagram.com/datawithsai/) Explore the [Breakout team’s portfolio ↗ ](https://breakout.company/) From the work ## Some things we’ve learned along the way. [Research ### 12 guests, 1,000+ sources, 6 hours ↗ ](https://machinehouse.media/writing/twelve-guests-six-hours)[Guest strategy ### Celebrity guests get views. Experts build the show. ↗ ](https://machinehouse.media/writing/celebrity-vs-expert)[Production decisions ### You can pour 15 lakh into a video and set it on fire ↗ ](https://machinehouse.media/writing/the-15-lakh-question) [All writing ↗ ](https://machinehouse.media/writing) ## Some machines are for everyone. We also make software and games. TubeCount asks you to guess how many views a video got from its thumbnail. Try it in Machine House Labs . [Explore Machine House Labs ↗ ](https://machinehouse.media/labs) LAB / 001 Available to play TubeCount ### How good is your instinct? A real YouTube thumbnail. Three guesses. Find out how well you can judge what people watch. [Play TubeCount ↗ ](https://tubecount.com)[About the game ↗ ](https://machinehouse.media/play) Before we talk ## A few useful answers. Who should work with Machine House? + Creators, founders and brands with something worth building. A show, a channel, a body of work, or an entire content operation. We work across strategy, research, writing, production, packaging, distribution and growth, supported by AI systems built around the work. Can you handle the whole channel? + Yes. Rishwajeet leads strategy, with his team at [The Breakout Company ](https://breakout.company/) covering scripting, production, editing, packaging and channel operations. You can also work with him on strategy and systems through Machine House. We scope the team around what your channel needs. How are Machine House and Breakout connected? + Rishwajeet founded Machine House Media and co-founded The Breakout Company. Machine House brings together his content strategy, production expertise and creative technology. His Breakout team handles full-service engagements. Both begin with a conversation about what you want to build. What makes your approach different? + We study the actual content alongside its results: topics, guest types, hooks, title tests, retention and comments. That analysis changes the next creative decision. For Roshan Kotla, reading 443 comments helped shape two follow-up essays that each passed 90K views. For Masoom, a 12-dimension editorial model connects guest choices, research, packaging and performance. Rishwajeet leads the creative judgment and the team carries the work through production and release. How do you use AI in the work? + The useful part is the context: your channel’s history, audience, source research, earlier creative decisions and actual results. We build systems around that material to support research, scripting, production briefs, packaging and review. Writers and strategists make the creative decisions; the system helps them start with a better understanding of the work. What does an engagement cost, and how does it work? + Most ongoing engagements run on quarterly retainers. After a 30-minute intro call, we prepare a written proposal with the scope, team and investment. The usual rhythm starts with an audit and plan, then weekly strategy, script reviews and a performance note. Can we work together remotely? + Yes. We are based in Bengaluru and work across time zones. Strategy, writing, editorial reviews and channel operations can run remotely. We agree the recording setup and any on-site production needs when we scope the engagement. Start a conversation ## Let’s make something that matters. A story you cannot stop thinking about. A show you want to build. A content operation ready for its next chapter. Bring the ambition. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/work/ankur-warikoo [Home ](https://machinehouse.media/)/ [Work ](https://machinehouse.media/#cases)/ Ankur Warikoo MONEY MATTERS / SHOW & SYSTEMS # Human stories. Engineered for scale. Rishwajeet helped launch Money Matters and worked across its first 58 episodes. He later built the guest-selection system that reduced weekly application review from days to under an hour. Results reported June 2026 · Updated September 7, 2026 01 / The short answer ## Who produces Money Matters? Rishwajeet helped launch Money Matters end to end as Ankur Warikoo’s Chief of Staff and Content Operations Manager. He worked across strategy, production and distribution for the first 58 episodes, while leading a 13-person team within Warikoo’s wider content operation. Each episode needed a guest willing to open up their finances, preparation for that conversation, and a team to record, edit and release it. 02 / The full story ## An ambitious show. A precise operation. - The show Money Matters opens up the financial lives of real people. A strong episode needs more than a set of numbers: a decision, a tension, a story the audience can see itself in. Rishwajeet helped turn that premise into a repeatable show, with 58 episodes and 20M+ views during his run . - The operation The work connected ideation, guest selection, recording preparation, production and distribution. Rishwajeet managed a 13-person team and seven partners across Warikoo’s broader content operation. The role covered the work between choosing an episode and publishing it. - The intelligence Thousands of guest submissions made selection the show’s biggest bottleneck. Rishwajeet built a two-stage AI workflow: first demographic and structural eligibility, then editorial scoring against patterns from earlier episodes. It turns a flood of applications into an informed shortlist. A strategist makes the final selection. - The efficiency The initial system was built in four days . Weekly review fell from days to under 60 minutes . The team could spend more time developing promising stories, with explicit selection criteria and a process they could run themselves. 03 / The numbers ## The show’s results. The work behind them. VIEWS ON SHOW 40M + FOLLOWER GROWTH 9M→13M Wider ecosystem · two-year engagement APPLE BEST 2× #1 finance podcast, India WORKFLOW BUILD 4d Initial build · guest-selection workflow Historical outcomes reported June 2026. Show growth reflects the combined work of the host and wider team. Stack demographic filter editorial scoring strategist sign-off production handoff weekly perf note 04 / The content intelligence ## A shortlist the team could review each week. The workflow filters applications for demographic and structural eligibility, then scores the remaining stories against editorial criteria drawn from earlier episodes. A strategist reviews the shortlist and chooses the guests. Warikoo’s team continued running Money Matters after Rishwajeet’s departure. The June 2026 case record reports that the guest-selection system was still in use two years after it was built. Explore the [intelligence behind a content operation ](https://machinehouse.media/services/ai-content-systems), or the full scope of [building and producing a podcast ](https://machinehouse.media/services/podcast-production). 05 / The work itself ## Watch Money Matters. Ep. 1 — Left with nothing on month end? Where the show started. Ep. 114 — A financial plan for a ₹35,000 salary. 114 episodes in, the pipeline still picks the guests. All episodes → [the Money Matters playlist on YouTube ](https://www.youtube.com/playlist?list=PLLbfvOX935JGUsSj8X5CQtjiehkZCnelR) 06 / Common questions ## Direct answers. Who produces Money Matters, Ankur Warikoo’s podcast? + Ankur Warikoo’s team runs Money Matters. Rishwajeet Singh helped launch and run the show during his time as Warikoo’s Chief of Staff and Content Operations Manager, including its first 58 episodes. He also built its AI-assisted guest-selection system. What was Rishwajeet’s role beyond the AI system? + He worked across show strategy, production and distribution, and led a 13-person team within Warikoo’s wider content operation. The guest-selection system was one part of the work of building and running Money Matters. How did the AI guest-selection system work? + A first stage checks demographic and structural eligibility. A second scores editorial potential using criteria informed by earlier episodes. A strategist reviews the shortlist and makes the final choice. Built in four days, the workflow reduced weekly review from days to under 60 minutes. What results did Money Matters get? + The June 2026 case record reports 40M+ show views and two Apple Best Podcast recognitions. During Rishwajeet’s two-year engagement, Warikoo’s wider creator ecosystem grew from 9M to 13M followers. The first 58 episodes generated 20M+ views during his run. Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/work/masoom-minawala [Home ](https://machinehouse.media/)/ [Work ](https://machinehouse.media/#cases)/ Masoom Minawala THE MASOOM MINAWALA SHOW / EDITORIAL INTELLIGENCE # The systems behind the whole show. Guest selection, research, questionnaires, trailers, titles, thumbnails and Shorts. We handle the editorial work across The Masoom Minawala Show through systems we’ve built for it, with Rishwajeet working directly with Masoom. Results reported June 2026 · Updated September 9, 2026 01 / The short answer ## Who produces The Masoom Minawala Show? Rishwajeet works directly with Masoom on the show’s content strategy and editorial operation. Machine House connects the entire arc: deciding what the show should explore, choosing guests, researching their stories, preparing questionnaires, shaping trailers and Shorts, and developing the episode’s titles and thumbnails. The show starts from a female perspective on work, money, relationships, health and identity. That informs the guests and questions, while the conversations remain open to anyone who finds them useful. 02 / The full story ## The systems across an episode. - Strategy & guest selection We use the show’s episode history, audience response and prediction models to assess possible guests and subjects. Rishwajeet and Masoom choose the conversation: what this person can explain, why it matters to the audience and what the episode should deliver. - Guest research Our research system follows interviews, books, writing and other sources to find the stories, gaps and contradictions worth exploring. For one New York shoot, it researched 12 guests across 1,000+ sources in six hours , producing an 80-page brief per guest. - Episode plans & questions We turn that research into the episode’s premise, question arc and host questionnaire. The brief gives Masoom the context behind a question, places to follow up and a route through the conversation. - Titles & thumbnails After recording, we develop title and thumbnail options from what the conversation actually delivers. Packaging tools draw on earlier episodes and tests; the team reviews the options and chooses the release package. - Trailers & editor briefs We work through the recording to select the cold open and trailer, then give the editors a structured brief. The host-facing brief and Editor Pack carry the decisions into the edit. - Shorts & captions The Shorts workflow identifies candidate clips and develops a cut sheet for the editors, with openings, cut instructions and captions. Each Short needs to make sense to someone who has never watched the full episode. - Performance & the next episode We refresh the show’s data, code episodes across 12 editorial dimensions and compare forecasts with released work. Retention, packaging tests and audience response feed back into the models and research used for the next guest, questionnaire and release. 03 / The numbers ## The show’s results. The work behind them. REPORTED SHOW VIEWS +94.8% 2.39M → 4.66M WATCH HOURS +102.8% 328K → 665K NET SUBSCRIBER ADDS +64.7% 23,093 → 38,029 OLDER CATALOGUE 1.26M views in the 2026 period Matched 1 January–3 September, 2025 vs 2026. YouTube Analytics, retrieved 6 September. Full long-form show catalogue, including older releases and sponsored episodes; personal uploads and Shorts excluded. These are show outcomes from the combined work of Masoom and the team. Measurement note · 9 September 2026: the exported “views” totals above are reported counts. The 2026 window crosses YouTube’s 24 August public-counting change; this API export has not yet been reconciled to a common engaged-view series. The +94.8% comparison should not be read as a verified like-for-like increase in audience viewing. Watch hours and net subscriber additions are separately reported measures. [Read the measurement analysis. ](https://machinehouse.media/writing/youtube-changed-what-a-view-means) The older work is still earning attention. Earlier episodes contributed 1.26M views, or 27% of the 2026 total. The 35 episodes released in the current-year period contributed 3.40M views. The show is building a catalogue people keep finding. The earlier June 2026 case record reported 35M+ views for the show, subscriber growth from 70K to 140K, Apple Best Podcast recognition in 2025 and 3.5M organic Instagram reach for one piece. Those historical figures use a different scope from the long-form comparison above. 12 dimensions topic authority emotional register specificity hook type guest archetype packaging runtime release window tone format subject novelty 04 / The content intelligence ## Why we changed the guest-selection question. Early in the show, we tagged episodes by guest type, topic, mood and fame, then compared viewing with subscriber and comment response. The pattern was useful: a prominent guest could bring reach, while an expert could give viewers a reason to stay with the show. That changed the question behind the next booking. What should this episode build: a bigger audience, a useful understanding, a stronger community or a particular kind of authority? The answer shapes the research, the question arc and the promise in the title. The ongoing model codes episodes across 12 editorial dimensions and checks forecasts against released work. We read it alongside retention, packaging tests and audience feedback. Rishwajeet and Masoom make the creative decisions; the analysis gives them something specific to challenge and improve. Read the original account of [the guest-strategy finding ](https://machinehouse.media/writing/celebrity-vs-expert), the [three-way title test ](https://machinehouse.media/writing/the-15-lakh-question), and [researching 12 guests in six hours ](https://machinehouse.media/writing/twelve-guests-six-hours). 05 / The work itself ## Watch the show. Shloka Ambani & Maniti Shah — unfiltered on ambition, motherhood & building legacy. Mickey — Bollywood's makeup legend on the biggest mistakes Indian women make. All episodes → [The Masoom Minawala Show playlist on YouTube ](https://www.youtube.com/playlist?list=PLrcp6HVr1TNpK7ejP6GlMjuM8QU_jrvCk) 06 / Common questions ## Direct answers. Who works on the content behind The Masoom Minawala Show? + Rishwajeet Singh works directly with Masoom on the show’s strategy and editorial operation through Machine House Media. The work covers strategy, guest selection, research, questionnaires, trailers, Shorts, titles, thumbnails and performance analysis. What agency does Masoom Minawala work with for her podcast? + Machine House Media runs the show’s content strategy and editorial operation, led by founder Rishwajeet Singh. The engagement connects guest selection and research with recording preparation, trailers, Shorts, packaging and performance review. How does the content system improve the show? + The system connects audience research, guest knowledge, editorial decisions and published performance. We compare forecasts with audience response and use the findings to refine guest choices, questions, trailers, Shorts and packaging. Rishwajeet leads the creative judgment throughout. What results did The Masoom Minawala Show get? + The long-form show catalogue recorded 4.66M views and 664,872 watch hours from 1 January to 3 September 2026: a 102.8% increase in watch hours over the same dates in 2025. The reported view-count increase is 94.8%, but the 2026 window crosses a counting change and has not been reconciled to a common view definition. Net subscriber additions attributed to the show rose 64.7%. The comparison includes older catalogue viewing and sponsored episodes, and excludes personal uploads and Shorts. These are outcomes of the show and wider team. Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/podcast-production [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ Podcast production PODCAST STRATEGY & END-TO-END PRODUCTION # Build a show people come back for. Podcast strategy and production, from the first premise to the next release. We help you choose the right guests, prepare better conversations, shape the edit and packaging, and learn from the audience you are building. [Talk about your channel ↗ ](https://calendly.com/rishwajeet/intro-call)[Explore the scope ↓ ](https://machinehouse.media/services/podcast-production#includes) Results and reporting periods are listed with each case. 01 / What we do ## Choose the guest. Prepare the conversation. - The show’s premise Define the audience, the question the show keeps returning to, and the format that suits the host. We shape the first slate around a clear reason to watch and a sustainable recording schedule. - Guest research & questions Find what each guest can teach through their experience. Research becomes an episode promise, a question arc and a recording brief the host can use in the room. Selection workflows help when a show is reviewing applications at scale. - Recording & edit Plan the recording setup and production handoffs, then shape the episode around the actual conversation. The full-service scope can include recording coordination, editing and finishing; we agree the setup and responsibilities with you. - Titles, thumbnails & trailers Develop a package for the episode’s strongest idea. We work from the recording to prepare title options, thumbnail directions, trailers and short-form cuts, so each piece has a reason to exist on its own. - Release & review Coordinate publication across the agreed platforms, then review performance and audience response. What we learn informs future guests, questions, episode structure and packaging. 02 / The engagement ## Plan around the host and the recording. We start with the host, the audience and the show’s recurring question. Then we agree a guest slate, research deadlines, recording setup and who reviews the episode before release. Most ongoing engagements run in quarterly cycles , with weekly strategy, script or episode reviews, and a performance note. We use the quarter’s results to decide what to continue, change or stop. Rishwajeet stays involved in the strategy and reviews. You can work with him through Machine House on strategy and systems, or bring in his Breakout team for the full channel. We agree the team, deliverables and responsibilities in a written proposal. A recording studio and a production partner solve different parts of the job. A studio booking can supply the room, equipment and recording crew, depending on its package. Show development, guest research, editorial decisions, post-production and distribution need explicit owners. We can work around your chosen studio or coordinate the agreed production setup; the proposal spells out each responsibility. Before signing a production scope, ask to see a comparable episode and the work that went into it. Confirm the research, edit reviews, recording location and source-file handoff alongside the number of deliverables. For worked examples, see how to [develop the format ](https://machinehouse.media/guides/show-format-development), [plan a video podcast launch ](https://machinehouse.media/guides/video-podcast-launch-plan), or [build a faithful episode trailer ](https://machinehouse.media/guides/podcast-trailer-editing). 03 / When it helps ## Starting a show or producing the next season. - A creator or founder launching a show: turn the initial idea into a format, guest slate and workable production plan. - An established host: improve research, question arcs and packaging while building a stronger editorial direction. - A show with a growing team: connect guest preparation, recording, edits and distribution through clearer handoffs. - A brand investing in long-form: create a show with a useful audience promise and the team to sustain it. We keep a limited roster of active engagements so there is time to understand the channel, review the work and follow through on the decisions. 04 / Relevant work ## Money Matters and Masoom’s show. [The Masoom Minawala Show: ](https://machinehouse.media/work/masoom-minawala) the long-form catalogue grew 102.8% in year-to-date watch hours. The case also reports view totals, with a qualification about the counting change. Guest analysis, deep research, questionnaires, trailers and packaging all feed into the editorial operation. [Money Matters: ](https://machinehouse.media/work/ankur-warikoo) Rishwajeet helped launch the show and worked across its first 58 episodes, which generated 20M+ views during his run. He also built the selection system that cut weekly guest review from days to under an hour. Masoom: matched 1 January–3 September, 2025 vs 2026, long-form catalogue. Money Matters: Rishwajeet’s historical engagement record. Show outcomes reflect the host and wider team’s work. ## Look closer at the work. [Podcast guest research Guest shortlists, source research, interview questions and a brief for the host. We read the guest’s previous work to find specific experiences worth discussing and questions they have not already answered. Explore ↗ ](https://machinehouse.media/services/podcast-guest-research)[Long-form to Shorts Turn a podcast, film or long-form script into short-form pieces that make sense to a new viewer. We find the strongest ideas, preserve their meaning and rebuild the opening, structure and payoff for the format. Explore ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Content production costs Compare the work covered by a quote: research, writing, recording, editing, packaging and release. Use the sample brief to specify your formats, volume, team and review process. Explore ↗ ](https://machinehouse.media/guides/content-production-costs) 05 / Before we talk ## Questions about working together. What does your podcast production service include? + The scope can cover show strategy, guest selection and research, recording briefs, production coordination, editing, titles, thumbnails, trailers, short-form cuts, publishing and performance review. We agree the exact recording setup and production responsibilities in the proposal. Can you help us launch a podcast from scratch? + Yes. We start with the host, audience and show premise, then develop the format, first guest slate, recording plan and release workflow. Rishwajeet’s experience includes helping launch Money Matters and building the editorial operation behind The Masoom Minawala Show. Who leads the engagement? + Rishwajeet leads strategy and editorial reviews. For full-service production, he works with his team at The Breakout Company, the agency he co-founded. How much does podcast production cost? + The investment depends on recording requirements, episode cadence and the production stages we take on. Most ongoing work runs on quarterly retainers. We prepare a written scope and quote after a 30-minute intro call. Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/youtube-growth [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ YouTube growth YOUTUBE STRATEGY & CHANNEL MANAGEMENT # A channel with a direction. And a team behind it. Machine House helps creators and brands grow on YouTube through channel strategy, scripting, production, packaging and a weekly performance review. Rishwajeet leads the strategy, with the Breakout team bringing the full operation together. [Talk about your channel ↗ ](https://calendly.com/rishwajeet/intro-call)[Explore the scope ↓ ](https://machinehouse.media/services/youtube-growth#includes) Results and reporting periods are listed with each case. 01 / What we do ## Run the channel from idea to release. - Channel diagnosis We read the videos alongside their data: topic, format, hook, title and thumbnail, click-through rate, retention, subscriber response and comments. Comparing similar releases helps identify what needs to change in the next idea, script or package. - Topics & scripts A slate of ideas with a clear audience promise, followed by research, hooks, scripts and editorial reviews. Prediction models help assess the options where your channel’s data supports them. - Production & publishing Recording plans, editing, titles, thumbnails and release schedules. For a full-service engagement, our Breakout team takes these stages through to publication with clear ownership at each handoff. - Packaging tests We develop title and thumbnail directions, test the selected options, and read the result in context. A past three-way title test returned 39.2%, 32.1% and 28.8% watch-time share ; the winning wording informed the next packaging decision. - Performance review A weekly numbers note connects the results to an editorial decision: which topics to revisit, where viewers lost interest, and what to change in the next script or package. 02 / The engagement ## A weekly review of the work and results. We review the existing channel, propose the next slate and agree a release schedule with your team. Weekly reviews cover upcoming scripts, packaging choices and what the published videos tell us. Most ongoing engagements run in quarterly cycles , with weekly strategy, script or episode reviews, and a performance note. We use the quarter’s results to decide what to continue, change or stop. Rishwajeet stays involved in the strategy and reviews. You can work with him through Machine House on strategy and systems, or bring in his Breakout team for the full channel. We agree the team, deliverables and responsibilities in a written proposal. Comparing partners? Read our [guide to choosing a YouTube growth agency ](https://machinehouse.media/guides/choosing-youtube-growth-agency), with a worked proposal comparison and an editable brief. 03 / When it helps ## For new channels and the next phase of growth. - A new channel: establish the audience, formats and first slate before committing to a publishing schedule. - An existing channel that has stalled: diagnose the constraint and test a clearer direction. - A creator with a growing operation: bring strategy, writing, production and measurement into one weekly rhythm. - A brand building an audience: make videos people have a reason to watch and a reason to return to. We keep a limited roster of active engagements so there is time to understand the channel, review the work and follow through on the decisions. 04 / Relevant work ## The channels we have worked on. [Masoom’s long-form show catalogue: ](https://machinehouse.media/work/masoom-minawala) 328K to 665K watch hours across the same January–September period. The view totals are reported separately in the case, with a note on the counting change. An ongoing editorial operation connects guest analysis, research, packaging and review. [Roshan Kotla: ](https://machinehouse.media/work/roshan-kotla) a breakthrough essay at 791K views, followed by two at 93K and 91K. The audience response helped identify the next subjects. Earlier engagements include [Indian Silicon Valley’s weekly views growing from 30K to 150K ](https://machinehouse.media/work/indian-silicon-valley) and [The KariGhars’ 478% first-month watch-time increase ](https://machinehouse.media/work/the-karighars). Masoom: 1 January–3 September, 2025 vs 2026. Roshan: public views on 7 September 2026. Earlier outcomes retain their original reporting periods. ## Look closer at the work. [YouTube channel audit We review your videos, audience, titles, thumbnails and production process. You get a written diagnosis, changes to specific videos and a proposed slate to discuss with Rishwajeet. Explore ↗ ](https://machinehouse.media/services/youtube-channel-audit)[Titles & thumbnail strategy YouTube packaging connects an idea to the person who might watch it. We develop titles and thumbnail directions around the audience, the story and the promise the video can actually deliver. Explore ↗ ](https://machinehouse.media/services/youtube-packaging)[The KariGhars Rishwajeet worked on the YouTube channel’s scripts, packaging and production. His June 2026 work record reports a 478% increase in watch time in the first month. Explore ↗ ](https://machinehouse.media/work/the-karighars)[Indian Silicon Valley Rishwajeet managed agencies across short-form, long-form and production for Indian Silicon Valley, a content-led angel syndicate. The engagement ran from 2024 to 2026. Explore ↗ ](https://machinehouse.media/work/indian-silicon-valley) 05 / Before we talk ## Questions about working together. How much does YouTube channel management cost? + Scope depends on the channel and which production stages we own. Most ongoing engagements run on quarterly retainers. After a 30-minute intro call, we prepare a written proposal with the team, deliverables and investment. Can I hire you just for YouTube strategy? + Yes. You can work directly with Rishwajeet on channel diagnosis, topics, formats, packaging and editorial reviews while your existing team handles production. See the content strategy service for the scope. Who will actually run my channel? + Rishwajeet leads strategy and joins the weekly reviews. For full-service engagements, his team at The Breakout Company covers scripting, production, editing, packaging and publishing. We agree responsibilities when we scope the work. Do you work with brand-new channels? + Yes. We help define the audience, positioning, formats and first content slate, then build a production rhythm. Previous launch work includes Freedom With AI’s growth from zero to 100K subscribers across two channels. Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/rishwajeet FOUNDER · MACHINE HOUSE MEDIA # Rishwajeet Singh Rishwajeet Singh is a content strategist and creative technologist, founder of Machine House Media and co-founder of The Breakout Company. He works on content strategy, scripts, production and the software his teams use to make the work. His work includes Ankur Warikoo’s Money Matters and The Masoom Minawala Show. Before the studio, he spent two years as Chief of Staff to Ankur Warikoo — running a 13-person team and the full production machine behind one of India's biggest creators. Based in Bengaluru. In the work since 2018, with 600+ episodes shipped across podcasts and long-form. 100M+ Views driven 3× Apple Best Podcast 15M+ Audience managed 600+ Episodes shipped Rishwajeet’s career record, reported June 2026. Client and show outcomes span his prior roles and Machine House work; the wider Breakout team portfolio is credited separately. Career totals reported June 2026 · Profile updated September 26, 2026 Rishwajeet Singh FOUNDER · STRATEGIST BENGALURU FIELD Creator economy · in the work since 2018 EDUCATION Ashoka University · magna cum laude SHIPPED 600+ episodes, podcasts & long-form BEST KNOWN Money Matters pipeline · still runs ELSEWHERE [Personal website ](https://rishwajeetsingh.com/) · [LinkedIn ](https://www.linkedin.com/in/rishwajeet-singh/) · [@rjrazzle ](https://x.com/rjrazzle) EMAIL [rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) 01 / The person ## Who is Rishwajeet Singh? Rishwajeet Singh is a content strategist and creative technologist who has spent his career inside the Indian creator economy — building the strategy, scripts, and production systems behind some of its biggest shows. He founded [Machine House Media ](https://machinehouse.media/), the AI-native content studio that builds and runs content systems for creators including Ankur Warikoo, Masoom Minawala, Astro Arun Pandit, The KariGhars, Freedom With AI, Avinash Mada, Scaler, Rocket Health, and Indian Silicon Valley. Work with him on a creative challenge or the full content operation. Rishwajeet leads strategy, writing direction, production, packaging and the AI systems that connect the work. His team at [The Breakout Company ](https://breakout.company/) provides the capacity for [end-to-end content production ](https://machinehouse.media/services/content-production). He founded Machine House and co-founded Breakout; they are distinct businesses connected by his work. The June 2026 career record reports 100M+ views across client work, three Apple Best Podcast recognitions for shows he helped build, and 15M+ audience managed. He studied entrepreneurship and media at Ashoka University (magna cum laude) and ran his first YouTube show while still enrolled. Machine House Media Private Limited was incorporated on 11 February 2026 (CIN U70200MP2026PTC082003), formalizing a practice years in the making. 02 / The story ## His boss told him to quit. Out of Ashoka, Rishwajeet joined wariCrew — Ankur Warikoo's content operation — and became Chief of Staff and Content Operations Manager: a 13-person team, seven partners, and the full pipeline behind an ecosystem that grew from 9M to 13M followers in his two years there. He launched [Money Matters ](https://machinehouse.media/work/ankur-warikoo) end to end — 58 episodes, episodes consistently hitting 200K–500K views , 20M+ views during his run. Then the show's biggest bottleneck — thousands of guest applications reviewed by hand every week — needed fixing. He built an AI pipeline in four days that cut the work from days to under 60 minutes. Ankur looked at what he'd made and said: "This work is a great fit for you. Your current job is not. You should quit." So he did. He moved to Bengaluru, spent months going deep on AI, and started building for other creators — Masoom Minawala's show from scratch ( 70K→140K subscribers, Apple India's Best Podcast of 2025 ), Indian Silicon Valley, The KariGhars, Freedom With AI, Astro Arun Pandit. In February 2026 that practice became Machine House Media. 03 / The record ## Selected work. The results alongside it. - wariCrew · 2023–25 Chief of Staff & Content Operations Manager to Ankur Warikoo. Led a 13-person team across the full pipeline — ideation to distribution — while the ecosystem grew 9M→13M followers . Launched [Money Matters ](https://machinehouse.media/work/ankur-warikoo): 58 episodes, 20M+ views in his tenure (40M+ reported by June 2026). Built the AI guest pipeline in 4 days — review time days → under 60 minutes a week. The June 2026 case record reports it still running. - The Masoom Minawala Show · 2025– Strategy, packaging, and a connected research and production workflow for [The Masoom Minawala Show ](https://machinehouse.media/work/masoom-minawala) — AI guest research, question generation, trailers, clips. For one NYC shoot the system researched 12 guests in 6 hours — 1,000+ sources, an 80-page brief per guest. Subscribers 70K→140K . Apple India's Best Podcast of 2025. - Indian Silicon Valley · 2024–26 Content operations for a content-led angel syndicate — three agencies managed across short-form, long-form, and production. Weekly views 30K→150K ; 7M+ views and +50K subscribers in five months. - The KariGhars · 2025–26 Packaging, scripting, and production overhaul for the YouTube channel. +478% watch time in the first month of the engagement. - Freedom With AI · 2024 Content production systems built from the ground up. 0→100K across two channels; 100K→500K for the brand. - Astro Arun Pandit · 2024–25 AI-powered operational systems for a 100-person creator business — workflows, org design, and strategy, working directly with the founding team. - Earlier Strategy and scripting for top Indian brands at OneUp Content Co.; growth roles at Ditto Insurance and YC-backed Synth ; podcast production for The One with Shabd in 2020. - Roshan Kotla · 2026 Research, editorial direction, scripts and packaging. The first three breakout essays reached 791K, 93K and 91K views. [Read how the audience response shaped the next subjects ↗ ](https://machinehouse.media/work/roshan-kotla) - Scaler · 2026 Research, scripts, production briefs and packaging. Seventeen verified releases passed 150K views; the leading essay passed 91K. [Inside the slate and its editorial lessons ↗ ](https://machinehouse.media/work/scaler) - Rocket Health · 2026 Editorial development and production-ready film scripts. The protein film reached 72K views, with research, visual direction and production briefs behind it. [Inside the film ↗ ](https://machinehouse.media/work/rocket-health) New video-view snapshots: 7 September 2026. See the [updated Masoom case ](https://machinehouse.media/work/masoom-minawala#metrics) for the matched year-to-date growth comparison. 04 / How he works ## "Content is a science. You can crack it. " Rishwajeet reads the content alongside its numbers: topic, guest type, hook, click-through, retention, subscriber response and comments. The aim is a better creative decision. For Roshan Kotla, 443 comments revealed a direction for the next essays. For Masoom, a 12-dimension model brings guest and topic analysis into the production. For Scaler, a review of the actual released slate sharpens what to write next. On where the line sits, from his own writing: "Taste is the rarest thing in the room now that everything else is free" — AI can render fifty options in a minute; knowing which one to bet on is the job. And on the work itself: "You can't scale taste by explaining it. You have to make the thing first." He stays involved in titles and thumbnails, including the choices tested after a video is ready. A real three-way title test on one video — same content, three titles — swung watch-time share by seven points; the winning call was a single word. He described that decision in his essay about the ₹15 lakh video. 05 / Timeline ## College show to studio. 2019–23 Ashoka University — entrepreneurship and media studies, magna cum laude. Produces podcasts and runs a YouTube show while at university; in the work since 2018. 2023 Joins wariCrew as Content Operations Manager, then Chief of Staff to Ankur Warikoo — 13-person team, full production pipeline. 2024 Launches Money Matters end to end — India's biggest personal finance show. Parallel operator roles begin: Indian Silicon Valley, Freedom With AI, Astro Arun Pandit. 2025 Builds the Money Matters AI pipeline in 4 days. Ankur tells him to quit — he does. Moves to Bengaluru, goes deep on AI, builds The Masoom Minawala Show to Apple India's Best Podcast of 2025. 2026 Machine House Media Private Limited incorporated — 11 Feb 2026, CIN U70200MP2026PTC082003. Launched publicly 10 March. The operator becomes a studio. Ankur Warikoo Masoom Minawala Astro Arun Pandit The KariGhars Freedom With AI Avinash Mada Scaler Rocket Health Indian Silicon Valley 06 / In conversation ## Beyond the brief. ### WiredIN with Saanand Warrier · Episode 37 Rishwajeet appeared on Saanand Warrier’s podcast for a conversation about YouTube strategy, creator careers, AI channels and building an audience. Published June 2, 2026, the episode introduces him as a YouTube strategist and links to Machine House. [Listen to the WiredIN conversation ↗ ](https://podcasts.apple.com/us/podcast/secret-to-grow-on-youtube-in-2026-masterclass-for-creators/id1650534461?i=1000770787356&l=ar) 07 / In his own words ## Writing from the work. - Editorial judgment [One video did 7.7 lakh views. Then we did it twice more. ↗ ](https://machinehouse.media/writing/did-it-twice-more) Why reading the comments after a hit changed the next story. - Packaging [You can pour ₹15 lakh into a video and set it on fire ↗ ](https://machinehouse.media/writing/the-15-lakh-question) A three-way title test, a seven-point swing in watch-time share, and the choices behind a thumbnail. - Guest strategy [Celebrity guests get views. Experts build the show. ↗ ](https://machinehouse.media/writing/celebrity-vs-expert) The difference between earning a burst of attention and giving people a reason to return. On running the work: [why agencies miss deadlines ](https://machinehouse.media/writing/why-agencies-miss-deadlines) and [the Tuesday report from his time at Ankur Warikoo ](https://machinehouse.media/writing/the-tuesday-report). For the craft behind a release, follow the complete [research-to-script example ](https://machinehouse.media/guides/research-to-script) and its [editor handoff ](https://machinehouse.media/guides/video-editor-brief). [Read all of Rishwajeet’s essays ↗ ](https://machinehouse.media/writing) Elsewhere: [Rishwajeet’s personal website ](https://rishwajeetsingh.com/), [Rishwajeet Singh on LinkedIn ](https://www.linkedin.com/in/rishwajeet-singh/), [@rjrazzle on X ](https://x.com/rjrazzle), and [Machine House Media on LinkedIn ](https://www.linkedin.com/company/machine-house-media). Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/play [Home ](https://machinehouse.media/)/ [Labs ](https://machinehouse.media/labs)/ TubeCount LIVE — NEW VIDEO DAILY # TubeCount: can you tell 200K from 200M? TubeCount is a free daily game built by Machine House Media — the studio behind [Ankur Warikoo's Money Matters ](https://machinehouse.media/work/ankur-warikoo) and [Masoom Minawala's #1 podcast. ](https://machinehouse.media/work/masoom-minawala) You see a real YouTube thumbnail. You guess the view count. Three guesses — the title and channel unlock as clues — then the reveal shows exactly how far off you were. Play long enough and you earn an eye rating: a number on the instinct this whole industry runs on. [Play TubeCount ](https://tubecount.com) [Why we built it ](https://machinehouse.media/play#why) Last updated: June 2026 01 / The game ## Three ways to find out how good your eye is. - The Daily One video, three guesses, same video for everyone. First guess from the thumbnail alone, second after the title unlocks, third after the channel and subscriber count. Every guess is scored on how close you landed in log space — being 2× off of 1M scores the same whether you said 500K or 2M. New video at midnight IST. - Streak Two videos — which got more views? One wrong pick ends the run. It starts easy and tightens: past streak ten you're choosing between videos from the same channel, where channel fame can't help you at all. - Binge Five videos, one guess each, up to 25,000 points. Finish a run and send it as a challenge link — your friend plays the exact same five videos and you compare scores head to head. VIDEO POOL 790+ real videos · verified counts REGIONS 25 US to India to Brazil to Japan GUESSES PER DAILY 3 clues unlock between guesses PRICE free no account needed 02 / Why we built it ## The game tests the instinct we sell. Machine House Media's day job is forecasting YouTube performance before production — prediction models that score topic, format, hook, and audience fit so our clients stop betting production budget on gut. That work runs on one underlying skill: look at a piece of packaging and know what it will do. TubeCount is that skill turned into a game. The data inside it is real and genuinely humbling: The Duck Song has 687 million views on a 2-million-subscriber channel. ChuChu TV has 98 million subscribers — and uploads that do 34 thousand views. Subscriber count predicts almost nothing. Packaging decides, and the eye that reads packaging is measurable. If you work in marketing, growth, or content, your eye rating is worth knowing. And if you'd rather have the forecast than the guess, that's the [YouTube growth practice ](https://machinehouse.media/services/youtube-growth) the game came out of. 03 / Common questions ## Direct answers. Q.01 Who made TubeCount? + TubeCount was built by Machine House Media , the AI-native content studio founded by [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) — the studio behind [Ankur Warikoo's Money Matters ](https://machinehouse.media/work/ankur-warikoo) and [The Masoom Minawala Show ](https://machinehouse.media/work/masoom-minawala). The game grew directly out of the studio's day job: forecasting YouTube performance from packaging before a video is produced. Q.02 Is TubeCount free? + Yes — completely free, in any browser, at [tubecount.com ](https://tubecount.com). No account needed; an optional Google sign-in syncs your stats across devices. Q.03 What does an eye rating mean? + It's the geometric mean of your first-sight misses — how far your first guess typically lands from the real number, as a multiplier. "2.1× off" means your thumbnail-only guesses usually land within about 2× of the truth. The first guess is made before any clues unlock, so the rating isolates pure packaging instinct. Q.04 Why did a content studio build a game? + Because it's the most honest demo of what we do. Our [channel management work ](https://machinehouse.media/services/youtube-growth) runs on forecasting what packaging will do before production. TubeCount makes that same instinct playable — and trainable. Every reveal recalibrates you a little. ## Today's video is waiting. How far off will you be? [Play TubeCount ](https://tubecount.com) [Talk growth instead ](https://calendly.com/rishwajeet/intro-call) or write to [rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/labs [Home ](https://machinehouse.media/)/ Labs MACHINE HOUSE LABS / TOOLS, GAMES & EXPERIMENTS # We make content for a living — and software because we can. Plan a team’s weekly capacity, check a quote against its source or play TubeCount. These public tools and experiments come from the content work. Updated September 18, 2026 01 / The bench ## What's live, what's in build. LAB-003 ### Content capacity planner Live Estimate complete releases per week from each person’s hours, task load and buffer. See which person limits the plan and how much time remains. Free · no account · runs in your browser [Plan capacity → ](https://machinehouse.media/labs/content-capacity-planner) LAB-004 ### Quote checker Live Find an exact or typography-normalized text match in a supplied source, then inspect the surrounding passage. Matching text does not verify truth or attribution. Free · no account · runs in your browser [Check a quote → ](https://machinehouse.media/labs/quote-checker) LAB-001 ### TubeCount Live A daily game: see a real YouTube thumbnail, guess the view count. Three guesses — the title and channel unlock as clues — then the reveal shows exactly how far off you were. Play long enough and you earn an eye rating: a number on the packaging instinct this whole industry runs on. 790+ videos · 25 regions · free, no account [Play → ](https://tubecount.com) [About the game → ](https://machinehouse.media/play) LAB-002 ### YouTube CPM calculator — India In build A planned calculator for exploring YouTube earnings assumptions for an Indian audience. The rate inputs and calculation method are still being developed. In build · 2026 02 / Why a lab ## Tools from the content work. Machine House combines content and creative technology. We build research tools, content intelligence and production workflows as part of the work. Labs gives the public experiments a home of their own. We list each project’s availability here. New releases will appear alongside TubeCount when they are ready to use. ## Have a content project you want to discuss? [Book a call ](https://calendly.com/rishwajeet/intro-call) [Play TubeCount ](https://tubecount.com) or write to [rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing [Home ](https://machinehouse.media/)/ Writing THE NOTEBOOK # Field notes from inside the machine. Essays by Rishwajeet on making content, running teams and the decisions behind the published work. [ 12 JAN 2026 Before the proposal What a client should be able to see before deciding to work together. 2 MIN ](https://machinehouse.media/writing/before-the-proposal) [ 9 SEP 2026 YouTube changed what a view means A 95% growth claim on our own homepage is a good place to start. 4 MIN ](https://machinehouse.media/writing/youtube-changed-what-a-view-means) [ 14 NOV 2024 The Tuesday report A November 2024 account of the weekly delays report at Ankur Warikoo. The 348 tasks and 20 delays describe the week discussed in the original post. 2 MIN ](https://machinehouse.media/writing/the-tuesday-report) [ 15 JAN 2026 Why agencies miss deadlines On ownership, coordination and the delivery promise a creative agency makes. 2 MIN ](https://machinehouse.media/writing/why-agencies-miss-deadlines) [ 2 JUL 2026 One video did 7.7 lakh views. Then we did it twice more. What 443 comments on a breakout video told us to make next. 2 MIN ](https://machinehouse.media/writing/did-it-twice-more) [ 5 JUN 2026 You can pour ₹15 lakh into a video and set it on fire Three titles went head to head — the winner moved watch time share by seven points. 2 MIN ](https://machinehouse.media/writing/the-15-lakh-question) [ 7 APR 2026 12 guests, 1,000+ sources, 6 hours Preparing research and questions for twelve guests ahead of Masoom’s NYC shoot. 1 MIN ](https://machinehouse.media/writing/twelve-guests-six-hours) [ 10 MAR 2026 My boss told me to quit. So I built Machine House. The post that launched Machine House Media, written the morning the site went live. 3 MIN ](https://machinehouse.media/writing/my-boss-told-me-to-quit) [ 20 JAN 2026 I spent 40 hours doing work I don't charge for A strategy client, a video I could see in my head, and 40 hours in Premiere. 2 MIN ](https://machinehouse.media/writing/forty-hours-below-my-rate) [ 16 DEC 2025 Celebrity guests get views. Experts build the show. We tagged every episode by guest type, topic, mood, and fame level, then ran the numbers — the data disagreed with the playbook. 2 MIN ](https://machinehouse.media/writing/celebrity-vs-expert) --- Source: https://machinehouse.media/writing/celebrity-vs-expert [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 16 DEC 2025 · 2 MIN READ # Celebrity guests get views. Experts build the show. We tagged every episode by guest type, topic, mood, and fame level, then ran the numbers — the data disagreed with the playbook. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/in/rishwajeet-singh/recent-activity/all/) Our podcast just won Apple's Best of 2025. And we got there by doing the opposite of what every content strategist tells you to do. I would know — I am one of them. I do strategy for [The Masoom Minawala Show ](https://machinehouse.media/work/masoom-minawala). When I started, I spent weeks inside the analytics, tagging every episode by guest type, topic, mood, and fame level, then running the numbers. I expected to confirm what everyone believes about podcast growth. Book a famous guest, their fans tune in, some percentage subscribes. Book an even bigger name. Repeat. The data said I was wrong. ## The Discovery The pattern was clear — celebrity episodes weren't growing the channel. They brought reach, but it was expert episodes that built the community. The difference wasn't subtle. Here are some recent examples: Shloka Ambani episode — 300K views, press coverage everywhere, multiple pages writing about the podcast. Mitesh Rajani episode — celebrity makeup artist, not a household name — 200K views but significantly more subscribers gained and a LOT more comments. 100K fewer views. More people who actually wanted to come back. ## Why This Happens In a space full of fluff — generic advice, surface-level content, the same recycled tips (and even guests!) — real expertise is rare. Most women will never sit across from a makeup artist who does Bollywood faces. Never get 30 minutes with a specialist who's spent decades mastering one craft. Never have a wealth expert walk them through exactly what to do. The podcast becomes that access point. Not fame — access. Real expertise you can't get anywhere else. Celebrity episodes bring people through the door. Expert episodes give them a reason to stay. ## The Principle Masoom's goal was never "biggest podcast." It was "best podcast for women." Not a tagline for her — an operating principle. If you want to be the best podcast for women, you cover everything women care about. All of it. Almost all women guests, not for optics, but because that's who the show is for. Deep expertise over surface-level fluff. Comprehensive coverage over chasing viral moments. ## What That Looks Like Deep-dive guides — "Our wealth guide," "Our skincare guide," "Our fitness guide" — featuring practitioners who actually do the work, not just talk about it. A 30-minute episode on eyebrows. Just eyebrows, with a specialist eyebrow stylist. "If we're the best podcast for women, we cover everything women care about. Eyebrows included." Masoom's clarity here is incredible. That's the difference between chasing reach and serving an audience. You don't skip topics because they're small. You cover them because your audience needs them. ## The Result In my tenure: 70K to 120K subscribers. Apple's Best of 2025. Celebrity episodes put you on the map. But what builds a community is giving people access to experts they'd never reach otherwise — real value in a space full of fluff. DMs open if you need content help, btw. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/forty-hours-below-my-rate [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 20 JAN 2026 · 2 MIN READ # I spent 40 hours doing work I don't charge for A strategy client, a video I could see in my head, and 40 hours in Premiere. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/in/rishwajeet-singh/recent-activity/all/) I recently spent 40 hours on work that wasn't worth my hourly rate. It made no sense, but it felt incredible. A client had hired me for content strategy. I ended up directing the shoot and editing the final cut myself. The video got 500K views. I don't direct. Haven't edited in three years. I charge a premium for thinking, not timeline work — the whole model is scale the thinking, delegate the doing. The truth is that most video projects die at "close enough." The production pipeline just demands it. Timelines, budgets, feedback rounds, stakeholder alignment — everything pushes toward acceptable. You learn to live with it. Tell yourself "this works" and move on. But I knew what this video needed to look like. Not the abstract version one might write in a strategy doc — "energetic, modern, premium." The actual version. I could see the specific pacing, where cuts would land, how long a shot should hold before it tips into indulgent — in my head. The client had a team — great people, but no process to translate what I was seeing into frames. That gap between my head and their execution would have landed us at "close enough" if we were to follow the deadline. So I directed. Then edited. 40 hours in Premiere for a strategy client. When I finished I just sat there watching it back. Every frame exactly what I saw in my head. I forgot what that felt like — being actually proud of something I made. If you know me, you've heard me talk about this a million times now haha. Once it existed, I could teach from it. Point at a specific frame and say "this is what I mean by letting a shot breathe." Build process from something real instead of a brief nobody can interpret. It is hard and it takes time, but you can't scale taste by explaining it. You have to make the thing first. The unscalable work is how you eventually scale (thanks PG!) This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/my-boss-told-me-to-quit [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 10 MAR 2026 · 3 MIN READ # My boss told me to quit. So I built Machine House. The post that launched Machine House Media, written the morning the site went live. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/in/rishwajeet-singh/recent-activity/all/) I'm launching my company today. It exists because my boss told me to quit. Introducing: Machine House Media Spent two years as [Ankur Warikoo's Chief of Staff ](https://machinehouse.media/work/ankur-warikoo) — strategy, operations, the full production machine behind one of India's biggest creators. Then Money Matters with Warikoo (India's biggest finance show) needed its production pipeline fixed. So I built it myself — Claude, a handful of APIs, four days. We went from having to talk to hundreds of applicants to an AI system that selected guests using insights from the channel data, automated outreach + full booking flow. What took the team days each week now took under 60 minutes. I loved building this! Ankur looked at what I'd made and said: "This work is a great fit for you. Your current job is not. You should quit." There's something weird about building a system in four days that outperforms months of manual work. It forces a question you'd rather not answer — what if the thing you're hired for isn't the thing you're actually good at? I quit. Moved to Bangalore. Spent months learning + building every AI workflow I could think of until I understood where it's genuinely useful and where it falls apart. Eight years working with creators meant I knew every bottleneck. AI turned out to be unreasonably good at exactly those bottlenecks. Then [Masoom Minawala ](https://machinehouse.media/work/masoom-minawala). Incredible vision to make the #1 podcast for women — and by now, I had the playbook down. Guest selection, guest research, question generation, trailer drafts, thumbnail and title — all automated. Not replacing taste. Scaling it. My direction, my instinct for what works, but now running through systems that let me operate at a level that wasn't possible even a year ago. Grew the show from 70K to 140K subscribers. The Masoom Minawala Show won Apple India's Best Podcast of 2025. 35M+ views. Just the start. Something Ankur taught me keeps proving true — content is a science. You can crack it. The difference now is that AI gives you the analytical firepower to actually do it at scale. Episodes analyzed across 20 dimensions. Performance predicted before a frame is shot. Every result feeding back into a custom model. Here's what I know: making content has never been easier. Standing out has never been harder. Most have the same tools, the same formats, the same playbooks. The noise is deafening. And the creators and brands that win from here won't be the ones who just produce more — they'll be the ones with systems that know what to produce, when, and why. Taste that scales. Data that compounds. Signal through the noise. That's what I build. machinehouse.media is live. Strategy, scripting, production, growth — powered by custom AI systems trained on your actual data. The system gets sharper every week you work with it. 100M+ views driven. 3x Apple Best Podcast. Creators, founders, brands, agencies — if you make content and something about it is slower or messier or less informed than it should be (it definitely is), let's talk! This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/twelve-guests-six-hours [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 7 APR 2026 · 1 MIN READ # 12 guests, 1,000+ sources, 6 hours Every interview a guest has ever given, every gap nobody's explored — before the host walks on set. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/in/rishwajeet-singh/recent-activity/all/) I researched 12 guests in 6 hours for a podcast shoot — every interview they've ever given, every gap nobody's explored, questions ranked by clip potential. The content was better than episodes we'd spent a week prepping. ## The problem You can't ask a question nobody's asked if you don't know what everyone's already asked. Most podcast producers don't. They Google the guest, skim a Wikipedia page, maybe watch one previous interview if they're feeling diligent. Then they walk in and ask "tell me your story" — the same question 50 other hosts asked. The guest clicks into autopilot, gives the rehearsed version, and everyone pretends this was a conversation. A guest who's done 40 interviews has answered hundreds of questions publicly. Somewhere in there are gaps — things they've never been asked, recent work that hasn't been covered. But finding those gaps means consuming all 40 interviews. No producer has time for that. ## What the system does For every guest on [The Masoom Minawala Show ](https://machinehouse.media/work/masoom-minawala), it consumes 100s of sources — not just their interviews, but their books, academic work, social media, news coverage, even the people around them. Family, co-founders, collaborators. It runs in spirals: every source it reads surfaces new leads, which surface new sources, until there's nothing left to find. The output is an 80-page brief. What they've said everywhere, what they've never been asked, where they've contradicted themselves across interviews, which stories have emotional weight, which angles map to our audience. Every claim fact-checked before Masoom walks on set. 12 guests. 1,000+ sources. 6 hours. A team of researchers couldn't match the depth in a month. Most producers think they need better questions. They need better information. The questions write themselves after that. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/the-15-lakh-question [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 5 JUN 2026 · 2 MIN READ # You can pour ₹15 lakh into a video and set it on fire Three titles went head to head — the winner moved watch time share by seven points. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/in/rishwajeet-singh/recent-activity/all/) You can pour ₹15 lakh into a video. If the title and thumbnail don't earn the click, you just set ₹15 lakh on fire. Our biggest video yet this year did 7.5 lakh views. Good script, clean edit, but none of that is why it worked. It worked because people clicked — and the only thing that decides a click is the title and the thumbnail. That's the packaging, and it carried the whole video. Both halves of it used to be a bottleneck. You needed a designer who knew the software for the thumbnail, and hours of rewriting to land the title. AI erased that. I can generate the gravestone scene in thirty seconds and draft fifty titles in a minute — the making is free now. Which means the making was never the hard part. The hard part is knowing which version is right, and that's what AI can't touch. Take the title. I had a few I believed in, so I put three head to head: "India's IT Dream Is Falling Apart" → 32.1% "Why India's IT Dream Is Falling Apart" → 28.8% "The Indian IT Dream is Dead" → 39.2%, winner The winning title won on one thing: finality. "Dead," where the others said "falling apart." That small shift, from a slow decline to a flat-out death, was a seven-point swing in watch time share — on YouTube, the difference between a video that travels and one that just disappears. The thumbnail was the same kind of call, made a dozen times over. The gravestone, the marigold garland, the little tricolour, each one chosen to make an Indian story land for a stranger who's never heard the word Infosys. AI could have rendered any of it in seconds. It could never have told me which version to bet on. That's the taste, and taste is the rarest thing in the room now that everything else is free. So I make my own packaging now. I'm no designer, but once you have the eye, AI is just the hands. I still bring in designers I love on some projects, and I'd happily do it again tomorrow. They were never the bottleneck. And here's what surprises people: I'm slower than ever. I'll sit with a title and a thumbnail longer than some people spend on the entire edit. Because that's the most exciting thing I know in this work. Months of effort, lakhs spent, and whether any of it gets seen comes down to a title and an image you can change in an afternoon. Get them right, and a video that would have died at a few hundred views reaches seven and a half lakh people. I don't know a bigger lever anywhere. If you work in content, you gotta start designing. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/did-it-twice-more [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 2 JUL 2026 · 2 MIN READ # One video did 7.7 lakh views. Then we did it twice more . A channel that averaged ~100 views a video, and the engine that made a viral hit repeat instead of fade. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/in/rishwajeet-singh/recent-activity/all/) We made a video that did 7.7 lakh views. This was for a channel with a median of ~100 views. And once the high faded, I started to doubt it. Was it luck? Chance? Could I do it again on purpose, or was that the only time? There was one way to know: make another one. And right away, every instinct I had was screaming to do the obvious thing — make the sequel. Same topic, part two. Ours was about the Indian IT dream dying, so "IT Dream is Dead 2" basically writes itself. The TCS layoffs breakdown. Milk it while it's hot. That move feels like the easy win, and it's how most creators burn their one hit to the ground. So we didn't make it. To figure out what actually made the first one land, I went back to the comments, all 443 of them. Almost none were arguing about IT. A 17-year-old asking if he should pay 15 lakh for a tier-3 engineering seat when there might be no job at the end. A 2025 graduate: "no company came." Parents pushing safe banking exams while their kid watched the floor give way. Underneath every one of them was the same thing: the sense that the deal a whole generation was sold, study hard and get the safe degree and you're set, has stopped being true. And people are scared. That feeling is what travels. But it matters who's saying it. Plenty of people feel it; almost none have the standing to say it out loud and be taken seriously. Roshan does. Fifteen years in software, runs his own firm, hasn't had a reason to hire a junior engineer in years. So when he says the dream is dead, you believe him — he built his whole career in that exact world. The feeling, and the person who's actually lived it. That's the engine. So for the next video we started from the same unrest and found where it met something else Roshan knows cold: the graduate who can't find any work at all. "Why India's Graduates Can't Find Jobs Anymore." It had production problems and went up a month late, long after the wave should have died. It still did 87k. The third, same method, is climbing past it as I write this. It was never luck. Every time, it's a feeling people are sitting with and a person who's earned the right to put words to it. Only the topic ever changes. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/content-strategy [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ Content strategy CONTENT STRATEGY / CREATIVE DIRECTION / AUDIENCE GROWTH # Know what to make next. And why it should work. Work directly with Rishwajeet Singh on the direction of your content. Audience research, positioning, formats, topics, storytelling and packaging—turned into a plan your team can make, measure and improve. [Talk about your channel ↗ ](https://calendly.com/rishwajeet/intro-call)[Explore the scope ↓ ](https://machinehouse.media/services/content-strategy#includes) Updated September 7, 2026 · Results below reported June 2026 01 / What we do ## Choose the subjects. Develop the formats. - A useful diagnosis We read the videos alongside their data: topic, format, hook, title and thumbnail, click-through rate, retention, subscriber response and comments. Comparing similar releases helps identify what needs to change in the next idea, script or package. - Positioning & format Agree what viewers should come to you for and how that promise works on camera. We develop repeatable formats that fit your expertise, personality and production capacity. - Topics & episode premises Build a content slate with a reason behind each choice. Research and performance models help us assess audience fit, specificity, hooks and format before you spend time recording. - Packaging & editorial reviews Develop titles and thumbnail directions alongside the idea, then review scripts or question arcs so the episode delivers on the promise. Your team leaves with concrete decisions and usable briefs. - A plan that changes with evidence After release, we review the results against the original idea and forecast. The next slate reflects what we learned about the audience, the format and the way you presented it. 02 / The engagement ## Your team makes it. We work on the direction. We begin by reviewing your channel and the question you want to resolve. The first plan sets out the audience, formats and proposed subjects, with briefs your writers and producers can work from. Most ongoing engagements run in quarterly cycles , with weekly strategy, script or episode reviews, and a performance note. We use the quarter’s results to decide what to continue, change or stop. Rishwajeet stays involved in the strategy and reviews. You can work with him through Machine House on strategy and systems, or bring in his Breakout team for the full channel. We agree the team, deliverables and responsibilities in a written proposal. Your team can handle recording, editing and release while Rishwajeet leads strategy and reviews. If you need production too, the proposal sets out what his Breakout team will deliver. For a worked example, see how to [develop a repeatable series ](https://machinehouse.media/guides/show-format-development), or [find a founder’s point of view ](https://machinehouse.media/guides/founder-thought-leadership). 03 / When it helps ## When you need an editorial partner. - You have a production team but need editorial direction. We give the team a clear slate, sharper briefs and regular reviews. - You are publishing consistently but growth has stalled. We examine where audience promise, topic selection and packaging have drifted. - You are starting a new channel or show. We establish the audience, premise and formats before building the first slate. - You have several content workstreams. We help connect long-form, short-form and distribution around a coherent editorial direction. We keep a limited roster of active engagements so there is time to understand the channel, review the work and follow through on the decisions. 04 / Relevant work ## How a review changed the next idea. [Roshan Kotla: find the reason behind the hit. ](https://machinehouse.media/work/roshan-kotla) We read 443 comments after the IT essay broke out, then built the next topics around the audience’s concern with education, jobs and getting ahead. The first three essays have passed 975K views together. [Scaler: make the technology matter to someone. ](https://machinehouse.media/work/scaler) Reviewing 17 releases put career and industry stakes at the centre of the strongest work. The leading essay passed 91K views; the verified slate passed 150K. [Masoom: give each guest a clear job. ](https://machinehouse.media/work/masoom-minawala) A 12-dimension content model supports the decisions behind an entire show, from the guest slate to the next title test. Public video views as of 7 September 2026. Periods, roles and scope in each case. ## Look closer at the work. [YouTube channel audit We review your videos, audience, titles, thumbnails and production process. You get a written diagnosis, changes to specific videos and a proposed slate to discuss with Rishwajeet. Explore ↗ ](https://machinehouse.media/services/youtube-channel-audit)[Titles & thumbnail strategy YouTube packaging connects an idea to the person who might watch it. We develop titles and thumbnail directions around the audience, the story and the promise the video can actually deliver. Explore ↗ ](https://machinehouse.media/services/youtube-packaging)[Founder & brand content Content strategy and production for founders, experts and brands with something useful to say. We turn real experience into a point of view, repeatable formats and work an audience has a reason to return to. Explore ↗ ](https://machinehouse.media/services/founder-brand-content) 05 / Before we talk ## Questions about working together. What does a YouTube strategist do? + A YouTube strategist helps decide who the channel serves, what it should make, how videos are packaged and how results should change the plan. Our work includes channel diagnosis, positioning, formats, topics, script reviews and performance analysis. Can we keep our existing editors and production team? + Yes. Rishwajeet can lead strategy and editorial reviews while your team handles recording, editing and publishing. We agree the briefs, reviews and handoffs around your existing operation. Can you also handle production? + Yes. Rishwajeet co-founded The Breakout Company. His team can cover scripting, production, editing, packaging and channel operations as part of a full-service engagement. How do prediction models fit into the strategy? + Models help assess ideas using the channel’s available performance data and editorial criteria. They support a strategist’s decision. Published results are compared with the forecast so later choices can improve. How do we start, and what does it cost? + Start with a 30-minute intro call about your channel, goals and team. We then prepare a written proposal with the scope, working rhythm and investment. Most ongoing strategy engagements run on quarterly retainers. Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services [Home ](https://machinehouse.media/)/ Services THE FULL CONTENT OPERATION # Everything around a remarkable piece of content. Work with Rishwajeet on strategy, writing and reviews, or bring in his Breakout team for production and channel management. We scope the research, recording, editing, packaging and release around the work you want to make. [Talk about what you’re building ↗ ](https://calendly.com/rishwajeet/intro-call) 01 / Where we come in ## A part of the work. Or the whole machine. - Strategy & direction Find the audience, point of view and formats worth building around. We turn research and performance into a slate, creative briefs and editorial decisions. [Explore content strategy → ](https://machinehouse.media/services/content-strategy) - Content production From research and writing through recording, editing, packaging and release. Films, shows, podcasts and short-form work, with a team around the creative direction. [Explore content production → ](https://machinehouse.media/services/content-production) - Distribution & growth Build a channel people return to. Topics, packaging, publishing and performance reviews form a single working rhythm. [Explore YouTube growth and management → ](https://machinehouse.media/services/youtube-growth) - Content intelligence Build the systems that connect your audience, research, creative decisions and results. Prediction, production workflows and a feedback loop around the actual work. [Explore AI content systems → ](https://machinehouse.media/services/ai-content-systems) 02 / Shows, channels & formats ## One creative ambition. Many ways to make it. A format brings these capabilities together. For a [podcast ](https://machinehouse.media/services/podcast-production), that means show positioning, guest selection, research, questionnaires, production, trailers, Shorts and release. For a [YouTube channel ](https://machinehouse.media/services/youtube-growth), it means a consistent editorial direction across topics, scripts, formats and packaging. A research-led film or brand content programme starts with its own audience and creative brief. The scope follows the work you are trying to make. 03 / The engagement ## Close to the work. Built around your team. Rishwajeet leads strategy and creative direction. His team at The Breakout Company provides the production capacity for full-service engagements. If you already have a strong team, we can work alongside it on strategy, writing, systems and reviews. Most ongoing work runs on quarterly retainers, with a written scope, team and investment. We keep a limited active roster, set a regular review rhythm, and use the audience’s response to decide what to make next. 04 / Selected work ## See the operation behind the outcome. [Roshan Kotla: ](https://machinehouse.media/work/roshan-kotla) 443 comments helped reveal why the first hit worked. Three essays have now passed 975K views together. [Scaler: ](https://machinehouse.media/work/scaler) 17 verified releases, 150K+ views, and a sharper read on the premises worth pursuing. [The Masoom Minawala Show: ](https://machinehouse.media/work/masoom-minawala) 102.8% more year-to-date long-form watch time. The view-count comparison is qualified in the case study. [Rocket Health: ](https://machinehouse.media/work/rocket-health) a researched protein film at 72K views. [Money Matters: ](https://machinehouse.media/work/ankur-warikoo) a show built end to end and a weekly guest-selection bottleneck cut from days to under an hour. Read our thinking on [what makes a production investment worthwhile ](https://machinehouse.media/writing/the-15-lakh-question) and [research at the scale a show demands ](https://machinehouse.media/writing/twelve-guests-six-hours). ## A specific challenge? Start here. [YouTube & video scriptwriting Develop the premise, source research and spoken script, with visual notes for the shoot and edit. Explore our work for Roshan Kotla, Scaler and Rocket Health. Explore ↗ ](https://machinehouse.media/services/video-scriptwriting)[YouTube channel audit We review your videos, audience, titles, thumbnails and production process. You get a written diagnosis, changes to specific videos and a proposed slate to discuss with Rishwajeet. Explore ↗ ](https://machinehouse.media/services/youtube-channel-audit)[Titles & thumbnail strategy YouTube packaging connects an idea to the person who might watch it. We develop titles and thumbnail directions around the audience, the story and the promise the video can actually deliver. Explore ↗ ](https://machinehouse.media/services/youtube-packaging)[Podcast guest research Guest shortlists, source research, interview questions and a brief for the host. We read the guest’s previous work to find specific experiences worth discussing and questions they have not already answered. Explore ↗ ](https://machinehouse.media/services/podcast-guest-research)[Long-form to Shorts Turn a podcast, film or long-form script into short-form pieces that make sense to a new viewer. We find the strongest ideas, preserve their meaning and rebuild the opening, structure and payoff for the format. Explore ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Founder & brand content Content strategy and production for founders, experts and brands with something useful to say. We turn real experience into a point of view, repeatable formats and work an audience has a reason to return to. Explore ↗ ](https://machinehouse.media/services/founder-brand-content) ## Understand the decision before the engagement. [Content performance prediction Choose what to forecast, record it before production and compare it with the release. Masoom’s guest-selection example shows why a forecast needs a specific audience goal. Explore ↗ ](https://machinehouse.media/guides/content-performance-prediction)[Content production costs Compare the work covered by a quote: research, writing, recording, editing, packaging and release. Use the sample brief to specify your formats, volume, team and review process. Explore ↗ ](https://machinehouse.media/guides/content-production-costs) Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/content-production [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ Content production CONTENT PRODUCTION / LONG-FORM / SHORT-FORM # From a promising idea to a body of work. We develop and produce films, shows, podcasts and short-form. Rishwajeet leads creative direction, with his Breakout team covering the agreed research, writing, production, editing and release. [Talk about what you’re building ↗ ](https://calendly.com/rishwajeet/intro-call)[Explore the scope ↓ ](https://machinehouse.media/services/content-production#includes) Updated September 7, 2026 · Case results reported June 2026 01 / What we do ## Research, writing, shooting and editing. - A creative direction We start with what you want to be known for, who you want to reach and what the content needs to accomplish. The direction becomes a slate of ideas and formats with a clear audience promise. - Research & writing Subject research, episode premises, scripts, questionnaires and recording briefs. We develop the argument or story, shape the opening, and give the production team the context to make the work well. - Recording & post Shoot planning, production coordination, editing and finishing, scoped to the format. Rishwajeet leads creative direction; his Breakout team brings the people and production capacity required for the engagement. - Packaging & distribution Titles, thumbnail directions, trailers, original short-form pieces and a release plan. We design each output for the audience and platform where it will appear, while keeping the central idea intact. - A better next release We review what people clicked, watched, shared and responded to, then use that evidence in the next slate, script and edit. The [content intelligence system ](https://machinehouse.media/services/ai-content-systems) keeps that learning connected to the daily work. 02 / The engagement ## Agree the format, then build the team. We agree the format, release schedule and production stages you need. The proposal names the team, what you supply, what we make and when each review happens. Rishwajeet stays involved in strategy and editorial reviews. Most ongoing engagements run in quarterly cycles, with regular reviews and a performance note that informs the next decisions. We keep a limited active roster to make that involvement useful. The written proposal sets out the team, deliverables, responsibilities and investment. [See how the services fit together ](https://machinehouse.media/services). 03 / When it helps ## A film, a show, or an ongoing operation. You need a complete content operation. We can own the work from strategy through publication, with responsibilities agreed around your team and production setup. You have a specific format to build. A founder-led channel, a research-led film, an interview show or a sustained short-form programme needs its own production rhythm. Explore [podcast production ](https://machinehouse.media/services/podcast-production) or [YouTube channel management ](https://machinehouse.media/services/youtube-growth) for those scopes. You already have a production team. Work with Rishwajeet on [strategy, writing direction and editorial reviews ](https://machinehouse.media/services/content-strategy) while your team produces the work. 04 / Relevant work ## From the sources to the finished release. [Rocket Health: a researched film at 72K views. ](https://machinehouse.media/work/rocket-health) We developed the protein story from source research through a full script, visual direction and production briefs, working with Studio 73 on the production. [Scaler: 17 verified releases, 150K+ views. ](https://machinehouse.media/work/scaler) Technology research becomes an argument, a script, a visual brief and a package. The published slate then gives the next round of writing something to learn from. [The Masoom Minawala Show: ](https://machinehouse.media/work/masoom-minawala) strategy, guest selection, research, questionnaires, trailers, Shorts and packaging, connected across the release. Public views as of 7 September 2026. Production roles and the work covered are detailed in each case. ## Look closer at the work. [Long-form to Shorts Turn a podcast, film or long-form script into short-form pieces that make sense to a new viewer. We find the strongest ideas, preserve their meaning and rebuild the opening, structure and payoff for the format. Explore ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Founder & brand content Content strategy and production for founders, experts and brands with something useful to say. We turn real experience into a point of view, repeatable formats and work an audience has a reason to return to. Explore ↗ ](https://machinehouse.media/services/founder-brand-content)[Content production costs Compare the work covered by a quote: research, writing, recording, editing, packaging and release. Use the sample brief to specify your formats, volume, team and review process. Explore ↗ ](https://machinehouse.media/guides/content-production-costs) Follow the craft through a complete [research-to-script example ](https://machinehouse.media/guides/research-to-script) and its [editor brief with source audio and cut sheet ](https://machinehouse.media/guides/video-editor-brief). On the operation behind delivery, read [why agencies miss deadlines ](https://machinehouse.media/writing/why-agencies-miss-deadlines). 05 / Before we talk ## Questions about working together. What types of content do you produce? + We work across long-form videos, podcasts, interview shows, research-led films, short-form video, trailers and the supporting packaging and distribution. The proposal defines the formats, volume, production setup and stages we own. Can you take an idea all the way to publication? + Yes. The engagement can include strategy, research, writing, shoot planning, production coordination, editing, packaging, publishing and performance review. Rishwajeet leads creative direction with his Breakout team delivering the agreed production scope. Can we hire you for one part of the production? + Yes. We can work on strategy, research and writing, packaging or editorial reviews alongside your existing team. We define the handoffs and approval responsibilities before work begins. How do you price content production? + Pricing depends on formats, release frequency, recording requirements and the stages we own. Most ongoing work runs on quarterly retainers, with a written scope, team and investment after the intro call. Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/ai-content-systems [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ AI content systems AI CONTENT SYSTEMS / EDITORIAL INTELLIGENCE # Your best thinking. Built into the operation. We build tools for guest selection, research, writing, production briefs and performance reviews. They use your source material and channel history, with editorial decisions assigned to the people doing the work. [Talk about what you’re building ↗ ](https://calendly.com/rishwajeet/intro-call)[Explore the scope ↓ ](https://machinehouse.media/services/ai-content-systems#includes) Updated September 7, 2026 · Case results reported June 2026 01 / What we do ## Put the channel’s history to work in the next brief. - The channel’s memory We connect each published video to its topic, format, guest, script, packaging and results. Your team can ask what worked, for whom and under which conditions—and bring the answer into the next assignment. - Strategy & prediction We define editorial dimensions that matter for your formats and use available performance data to assess topics, guests, hooks and packaging. Forecasts make the assumptions explicit; a strategist makes the creative decision. - Research to production Research becomes a script, questionnaire or recording brief. The content’s central idea stays connected through trailers, Shorts and packaging, with source material and editorial direction available at each stage. - Workflows with ownership We map the actual handoffs, recurring decisions and bottlenecks. Then we build the automation, review points and working tools that let the team move faster while keeping responsibility clear. - A learning operation After release, the system adds the results to the production record. The team can compare the forecast with what happened and use that history when choosing the next guest, subject or package. 02 / The engagement ## Start with a task the team repeats. Show us a recurring task, the material it uses and the result your team needs. We map the steps, choose what to automate and test the output against real examples before expanding the system. Rishwajeet works with the people who use the system to define the editorial criteria and review points. The proposal sets out integrations, testing, handoff and any ongoing support. The written proposal sets out the team, deliverables, responsibilities and investment. [See how the services fit together ](https://machinehouse.media/services). For a worked example, see how to [compare manual work, fixed automation and agents ](https://machinehouse.media/guides/ai-content-workflows). 03 / When it helps ## Where a system can save repeated work. The team repeats too much work. Research, briefs, publishing tasks and reporting often consume time without carrying useful context forward. We look for the parts that can become reliable systems. You need stronger editorial decisions at scale. Guest selection, topic planning and packaging can benefit from explicit criteria, relevant history and well-placed human review. You want the system and the content operation together. We can combine this work with [content production ](https://machinehouse.media/services/content-production) or lead [strategy with your existing team ](https://machinehouse.media/services/content-strategy). 04 / Relevant work ## Systems built inside working shows. [Money Matters: thousands of applications, a usable shortlist. ](https://machinehouse.media/work/ankur-warikoo) Rishwajeet built the AI guest-selection system in four days. It reduced weekly review from days to under an hour, with the team making the final editorial call. [Masoom: an editorial model the production can use. ](https://machinehouse.media/work/masoom-minawala) Twelve dimensions connect guest and topic choices to actual performance. The finding reaches the research, questionnaire, trailer and package. [Scaler: connect the script to the released result. ](https://machinehouse.media/work/scaler) We track the releases we worked on, read their performance alongside their premises and packaging, and use the review to choose the next slate. ## Look closer at the work. [Podcast guest research Guest shortlists, source research, interview questions and a brief for the host. We read the guest’s previous work to find specific experiences worth discussing and questions they have not already answered. Explore ↗ ](https://machinehouse.media/services/podcast-guest-research)[Content performance prediction Choose what to forecast, record it before production and compare it with the release. Masoom’s guest-selection example shows why a forecast needs a specific audience goal. Explore ↗ ](https://machinehouse.media/guides/content-performance-prediction) 05 / Before we talk ## Questions about working together. What is an AI content system? + It is a connected set of knowledge, tools and workflows built around a content operation. It can support audience research, strategy, guest selection, writing, production briefs, packaging and performance analysis, with human editorial judgment at the relevant decisions. Do we need a large amount of historical data? + We first inspect the material and performance data available. Established channels can support richer comparisons and prediction models. Newer operations begin with audience research, explicit editorial criteria and a system that learns as more work is published. Does the system make the creative decisions? + Rishwajeet and the editorial team make the creative decisions. The system provides research, context, forecasts and production support, then carries the results of those decisions into later work. Can you improve an existing workflow? + Yes. We review how your team currently works, identify the recurring bottleneck, and build around the tools and responsibilities that make sense for the operation. The first proposal defines the workflow, integrations, review points and handoff. Start a conversation ## Tell us what you’re trying to build. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides [Home ](https://machinehouse.media/)/ Guides # Work through the content decisions. Scripts, source packets, audio cuts and worksheets. Follow a worked example, inspect the decisions and adapt the files to your own project. ## Choose the direction [ ### Show format development Use Money Matters to study how a show finds its next story, then work through an invented series: its format, six premises and a pilot plan. Read the guide ↗ ](https://machinehouse.media/guides/show-format-development)[ ### Podcast guest selection Masoom’s guest-strategy example and a three-candidate exercise show how the audience goal changes a shortlist. The worksheet records the evidence, access and reason for each choice. Read the guide ↗ ](https://machinehouse.media/guides/podcast-guest-selection)[ ### Editorial slate planning See how Roshan’s comments changed the next essay, then work through a six-premise practice slate. Each idea has a question, evidence to find and a production route. Read the guide ↗ ](https://machinehouse.media/guides/editorial-slate) ## Research and write [ ### Research to script Follow a map-projection video from its source packet to the finished script. See the argument, a passage we cut and the evidence behind each section. Read the guide ↗ ](https://machinehouse.media/guides/research-to-script)[ ### Founder thought leadership Read Rishwajeet’s 40-hours essay alongside the decisions behind its argument. Use the interview worksheet to find equally specific material in a founder’s own experience. Read the guide ↗ ](https://machinehouse.media/guides/founder-thought-leadership)[ ### Interview research brief A completed interview brief with six questions, sources, follow-ups and gaps to resolve. See how Masoom’s research process informs preparation, then adapt the teaching example to your guest. Read the guide ↗ ](https://machinehouse.media/guides/interview-research-brief)[ ### Fact-checking a video script Work through six fact-checking decisions: unsupported figures, misleading comparisons, an invented quote and graphics that claim too much. Download the register with corrections and source notes. Read the guide ↗ ](https://machinehouse.media/guides/fact-checking-content)[ ### Hinglish scriptwriting Compare an English explanation, a literal adaptation and a spoken Hinglish script. The task is to preserve the idea while choosing words, examples and a rhythm the host can use. Read the guide ↗ ](https://machinehouse.media/guides/hinglish-scriptwriting) ## Produce and edit [ ### Video editor brief Listen to the source audio, follow the proposed cut and compare the two timelines. The example includes a full transcript, edit instructions and a blank brief. Read the guide ↗ ](https://machinehouse.media/guides/video-editor-brief)[ ### Podcast trailer editing Listen to two trailers cut from the same teaching conversation. Compare the openings, check every cut against the full recording and see why three clips were left out. Read the guide ↗ ](https://machinehouse.media/guides/podcast-trailer-editing)[ ### Content shoot planning A proposed half-day shoot for a heat-map explanation, with a schedule, shot list, crew responsibilities, fallbacks and editor handoff. Use the worksheet to plan your own recording. Read the guide ↗ ](https://machinehouse.media/guides/content-shoot-plan)[ ### Clips versus original Shorts A good passage can become a confusing clip when its setup disappears. Compare a recorded excerpt with a newly written short from the same source, then use the decision brief on your own episode. Read the guide ↗ ](https://machinehouse.media/guides/clips-vs-original-shorts)[ ### Accessible content production A caption file cannot explain a graphic the host never describes. Use this worked production plan to assign captions, transcripts, visual description and the final player checks before a release reaches the publisher. Read the guide ↗ ](https://machinehouse.media/guides/accessible-content-production)[ ### Video podcast launch plan Work backwards from the first release: test the show, finish a usable episode buffer and give every handoff an owner. A worked schedule and an editable launch plan. Read the guide ↗ ](https://machinehouse.media/guides/video-podcast-launch-plan) ## Distribute and learn [ ### Content performance prediction Choose what to forecast, record it before production and compare it with the release. Masoom’s guest-selection example shows why a forecast needs a specific audience goal. Read the guide ↗ ](https://machinehouse.media/guides/content-performance-prediction)[ ### Content distribution planning A worked release schedule with three companion pieces, their audiences, destinations and owners. Use it to plan what must be ready before each piece goes out. Read the guide ↗ ](https://machinehouse.media/guides/content-distribution-plan)[ ### Audience retention analysis Investigate three moments on an invented retention curve: an opening drop, a dip and a spike. Use the video’s content and audience context to choose an explanation to test. Read the guide ↗ ](https://machinehouse.media/guides/audience-retention-analysis)[ ### Content postmortems Start with Masoom’s catalogue totals, then compare three invented videos at the same age. Work through the arithmetic, possible explanations and a decision for the next release. Read the guide ↗ ](https://machinehouse.media/guides/content-postmortem)[ ### Measuring content business impact Follow six example records through deduplication, qualification, signed work and payment. See what the record can tell you about content’s contribution and where attribution remains uncertain. Read the guide ↗ ](https://machinehouse.media/guides/content-business-impact)[ ### Content catalogue review An old recording can still answer a current question. Review recent demand, outdated advice and missing follow-through before deciding what to keep, update or make again. Read the guide ↗ ](https://machinehouse.media/guides/content-catalogue-review) ## Run the operation [ ### Content production costs Compare the work covered by a quote: research, writing, recording, editing, packaging and release. Use the sample brief to specify your formats, volume, team and review process. Read the guide ↗ ](https://machinehouse.media/guides/content-production-costs)[ ### Content approval workflow Use a weekly delays report, an approval map and three worked review loops to decide who gives feedback, who resolves disagreements and what happens after a late change. Read the guide ↗ ](https://machinehouse.media/guides/content-approval-workflow)[ ### AI content workflows Examine Money Matters’ guest-selection process, then compare manual work, fixed automation and a model step. Try the evaluation exercise with an answer that sounds convincing but outruns its sources. Read the guide ↗ ](https://machinehouse.media/guides/ai-content-workflows)[ ### Evaluating AI content systems A polished brief can contain a false quote, miss a required question or hide a source conflict. Evaluate the whole job against explicit acceptance checks, then count the human work needed to finish it. Read the guide ↗ ](https://machinehouse.media/guides/evaluating-ai-content-systems)[ ### A knowledge system for a content team A team needs to know which decision is current, why it was made and who can change it. Use a worked decision record to carry that context from research into scripts, graphics and the next edit. Read the guide ↗ ](https://machinehouse.media/guides/content-team-knowledge-system)[ ### Choosing a YouTube growth agency Examine the diagnosis, check the claimed results and compare what each team will own. A worked proposal comparison and an editable brief for founders and marketing teams. Read the guide ↗ ](https://machinehouse.media/guides/choosing-youtube-growth-agency) ## More from the work. [What a channel audit should change ↗ ](https://machinehouse.media/services/youtube-channel-audit)[A real title test, with the numbers ↗ ](https://machinehouse.media/services/youtube-packaging)[What belongs in a guest research brief ↗ ](https://machinehouse.media/services/podcast-guest-research)[From a long-form source to a standalone short ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Rishwajeet’s essays ↗ ](https://machinehouse.media/writing) --- Source: https://machinehouse.media/services/founder-brand-content [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ Founder & brand content # Build a body of work only you could make. Content strategy and production for founders, experts and brands with something useful to say. We turn real experience into a point of view, repeatable formats and work an audience has a reason to return to. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Service · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## What happened, and what did you do? A product decision, a difficult customer conversation or a mistake you keep seeing in your industry can give you something useful to say. We interview you about what happened, ask for the details and work out which part deserves a piece of its own. ## Give the point of view a repeatable form. | Format | When it earns its place | Founder or expert essays | You can explain a subject through a strong argument and direct experience. | Interviews and podcasts | Access to other people’s expertise advances a clear show premise. | Demonstrations and breakdowns | The work, product or process is better understood by seeing it. | Customer questions | Recurring questions reveal what the audience needs help understanding. | Short-form stories | A focused idea or decision can deliver a useful payoff in a smaller piece. The format has to fit the speaker and the production capacity. We build a sustainable slate before committing the team to a volume of output. Written founder content can include LinkedIn posts, owned essays and newsletters. We interview for decisions and examples, develop the argument, then draft in the founder’s voice for the intended reader. A B2B founder can build a useful body of work around customer questions and operating experience without committing to daily posts. We agree a cadence that leaves time for the founder to contribute and review. ## The experience and the audience have to meet. In [a published account of our work ](https://machinehouse.media/writing/did-it-twice-more), the insight behind a technology video was the audience’s anxiety about education and employment. The creator could speak credibly about it because of his years running a software business. That relationship guided the next subject. For a brand, the same editorial question is useful: which part of your experience can help a particular audience understand something they care about? This is a way to choose material, not a promise that every founder should comment on the same news. ## Make it work in the real calendar. We connect the strategy to interviews with the founder, research, scripts or talking points, shoot preparation, edits, packaging and release. Rishwajeet leads the creative direction. His Breakout team can provide the production capacity for the agreed engagement. If you already have a capable team, we can own [strategy and editorial reviews ](https://machinehouse.media/services/content-strategy) while they produce. For a full operation, explore [content production ](https://machinehouse.media/services/content-production) and [YouTube management ](https://machinehouse.media/services/youtube-growth). ## Define what the content should do. Audience growth, reputation and qualified business demand are different outcomes. We agree which ones matter, then read the available evidence accordingly. Views show reach. Returning viewers can help assess an ongoing audience. Inquiries and customer conversations can tell a different part of the business story. An introduction or booking should be recorded as such. It becomes a qualified opportunity only when the actual conversation establishes fit. Clear measurement helps the next content decision stay connected to the business. ## Different businesses. Specific editorial work. [Roshan Kotla: ](https://machinehouse.media/work/roshan-kotla) his experience in software and hiring became essays about work that the audience recognised. Three releases passed 975K views, with research, scripts and packaging from Machine House. [Scaler: ](https://machinehouse.media/work/scaler) technology essays with a concrete career or industry stake; 150K+ views across 17 verified releases. [Rocket Health: ](https://machinehouse.media/work/rocket-health) researched health and culture films, with the protein release at 72K views. [The KariGhars ](https://machinehouse.media/work/the-karighars) and [Indian Silicon Valley ](https://machinehouse.media/work/indian-silicon-valley) show earlier work across brand storytelling and content operations. Each case sets out the responsibility and the outcome. New video counts are public snapshots from 7 September 2026. ## Questions about the work. Do you work with founders who are new on camera? + Yes. We can develop formats, talking points and a recording plan around the founder’s experience and comfort on camera. The production approach follows the person and the audience. Can this include both the founder’s voice and the company brand? + Yes. We define the role of each presence and how they connect. The founder’s point of view, company expertise and individual formats need a coherent editorial direction. Can your team handle production and publishing? + Yes. The engagement can include research, writing, production, editing, packaging and publishing with Rishwajeet and his Breakout team. The proposal defines responsibilities and release frequency. How do we know if the content is helping the business? + We agree the desired outcomes and available measurement before production. Audience response, qualified inquiries and commercial outcomes are recorded separately, then reviewed alongside the content decisions. ## Keep exploring. [Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Content production ↗ ](https://machinehouse.media/services/content-production)[Indian Silicon Valley ↗ ](https://machinehouse.media/work/indian-silicon-valley)[From founder experience to an argument ↗ ](https://machinehouse.media/guides/founder-thought-leadership)[Plan the distribution ↗ ](https://machinehouse.media/guides/content-distribution-plan) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/youtube-channel-audit [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ YouTube channel audit # Work out why the channel has stalled. We review your videos, audience, titles, thumbnails and production process. You get a written diagnosis, changes to specific videos and a proposed slate to discuss with Rishwajeet. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Service · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## What needs to change? Bring us the channel and the question you cannot resolve. It might be why a successful video has been hard to repeat, why people click but leave, or what to make after a change in direction. We compare the releases and watch the videos before recommending more production. Rishwajeet reviews the channel as both a strategist and a content operator: what viewers are being offered, what the videos deliver, and what the team can sustainably produce. Your existing editors and production team can stay in place. The outcome is a direction they can act on. ## Five connected parts of the channel. | Part | What we examine | What the decision changes | Audience | Who watches, what they came for, returning viewers and the questions in comments. | Which viewer the channel should prioritize. | Topics and formats | The audience promise, the creator’s authority and recurring structures across stronger and weaker videos. | The subjects and formats in the next slate. | Packaging | Titles and thumbnails together, the traffic context and available test results. | The promise made before someone clicks. | The viewing experience | Openings, story structure, information density, pace and retention alongside the actual video. | The next script, recording brief or edit. | The operation | Research, approvals, production capacity and the release rhythm. | What the team needs to change to deliver the plan. ## A hit can reveal the wrong next move. In a [July 2026 account of our work ](https://machinehouse.media/writing/did-it-twice-more), a channel with a median of roughly 100 views had a video reach 7.7 lakh views. The obvious next idea was a sequel about the same industry. Rishwajeet instead read all 443 comments to understand what had travelled. The response was about a broken promise: study hard, get the safe degree, and a secure career will follow. The creator had the experience to speak credibly about that anxiety. That combination became the basis of the next topic, about graduates struggling to find work. The follow-up reached 87K views in the published account. September update: those three Roshan Kotla essays have now passed 975K views together: 791K, 93K and 91K. [Read the full case, with the current video snapshots and our role. ](https://machinehouse.media/work/roshan-kotla) ## Metrics need a viewing context. A high click-through rate on a small, familiar audience can mean something different from a lower rate as a video reaches new people. We compare videos with attention to their age, format, traffic source and audience, then watch the content to understand what the comparison might mean. YouTube explains that CTR varies with content, audience and the surface where an impression appears. Its guidance also notes that high CTR and average view duration can coexist with low impressions. [Read YouTube’s explanation ](https://support.google.com/youtube/answer/7628154?hl=en). ## A diagnosis your team can use. - A written diagnosis with the priority problems and the evidence behind them. - A review of representative videos, including concrete changes to ideas, openings or packages. - A proposed direction for audiences, subjects and repeatable formats. - A prioritized next slate and tests that can confirm or challenge the diagnosis. - A review call with Rishwajeet and agreed responsibilities for the next phase. We agree the review depth and material at the start. Authorized YouTube Analytics access or exports allow a deeper assessment. A public-only review is possible, with its conclusions limited to what is visible. ## Take the diagnosis into production. You can hand the plan to your own team, continue with [content strategy and editorial reviews ](https://machinehouse.media/services/content-strategy), or bring in our team for [channel management ](https://machinehouse.media/services/youtube-growth). Where the constraint is repeated work or lost context, we can build the [systems supporting the operation ](https://machinehouse.media/services/ai-content-systems). ## Questions about the work. What does a YouTube channel audit include? + Audience and content diagnosis, review of topics and formats, titles and thumbnails, the viewing experience, and the production operation. We turn the findings into priorities, a proposed slate and concrete next tests. Do you need access to YouTube Analytics? + Authorized access or exports help us assess audience, traffic and performance in context. We can also review public videos and packaging, while making the limits of that evidence clear. Will you work with our existing team? + Yes. Rishwajeet can lead the audit and strategy while your writers, editors and producers execute the plan. We agree the briefs and review rhythm around your operation. Does an audit guarantee growth? + An audit gives you a diagnosis and a plan to test. Results depend on the audience, ideas, execution and how the work develops after the review. We do not promise a specific number of views or subscribers. ## Keep exploring. [Roshan Kotla: from one hit to an editorial direction ↗ ](https://machinehouse.media/work/roshan-kotla)[Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Channel management ↗ ](https://machinehouse.media/services/youtube-growth)[The decision behind a repeat hit ↗ ](https://machinehouse.media/writing/did-it-twice-more)[Investigate the retention curve ↗ ](https://machinehouse.media/guides/audience-retention-analysis)[Compare releases at the same age ↗ ](https://machinehouse.media/guides/content-postmortem) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/long-form-to-shorts [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ Long-form to Shorts # A long conversation. Several complete ideas. Turn a podcast, film or long-form script into short-form pieces that make sense to a new viewer. We find the strongest ideas, preserve their meaning and rebuild the opening, structure and payoff for the format. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Service · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## A recorded cut or a new script? A story told well in the recording may work as a cut. An explanation that takes ten minutes to establish may need a new script and another recording. We review the complete source and recommend which approach each idea needs. ## Start with the complete source. A memorable sentence may rely on a qualification two minutes earlier. A striking claim may be challenged later in the conversation. We read or review the complete source before selecting a premise, then retain the context necessary to make the shorter piece accurate. The working brief identifies the source passage, the short’s promise, the opening, the development and the close. For recorded cuts, it includes usable source locations. For new scripts, it identifies what comes from the source and what has been rewritten for the new viewer. ## The same lesson. A new piece of writing. Rishwajeet’s [published packaging essay ](https://machinehouse.media/writing/the-15-lakh-question) describes three title variants with watch-time shares of 32.1%, 28.8% and 39.2%. Its larger argument is about taste and the importance of packaging. A short can give the viewer one complete lesson from that material. New illustrative short script, derived from the published essay We tested three titles on the same video. The best one led the first by 7.1 percentage points of watch-time share. The first said India’s IT dream was falling apart. The winner said it was dead. My read was that finality made the story more compelling. But this test does not tell us to put “dead” in every title. It gives us a creative hypothesis to take into the next package. This is a new demonstration script, not an already published client Short or a claim about its performance. It retains the original test figures while supplying the context a viewer needs without reading the essay. ## A production brief with a clear job. | For a recorded cut | For an original short script | Source locations and the exact spoken material. | Source insight and a new standalone script. | Context required before or after the selected moment. | Opening, development and payoff for a new viewer. | Edit order, on-screen text and necessary supporting visuals. | Recording notes, visual direction and text treatment. | Checks that the cut preserves the speaker’s meaning. | Checks that the rewrite preserves the source’s claim and limits. ## Connect the shorts to the larger work. On [The Masoom Minawala Show ](https://machinehouse.media/work/masoom-minawala), short-form sits alongside strategy, guest research, questionnaires, trailers and packaging. A strong short has a reason to exist by itself, while the body of work builds the show’s identity. We review the published response by piece and premise. Those findings can change what we look for in the next recording, how we open a short, or which subject deserves a longer episode. The intelligence flows in both directions. See the wider [content production scope ](https://machinehouse.media/services/content-production). ## Questions about the work. Do you only cut clips from existing videos? + We can create recorded cuts or original short scripts derived from a complete long-form source. The choice depends on the material, voice and production needs. How many Shorts can you make from an episode? + That depends on how many strong, distinct ideas the source supports. We agree a target and production scope, then select pieces that can deliver a complete promise to a new viewer. Can the shorts work without sending viewers to the full episode? + Yes. Each piece should make sense on its own. Its relationship to the full episode and any call to action depend on the release plan. What source material do you need? + The complete recording, transcript or finished script, along with the creator’s voice and visual references. For recorded cuts we need usable media and source locations. ## Keep exploring. [Rocket Health: research into a film and original shorts ↗ ](https://machinehouse.media/work/rocket-health)[Content production ↗ ](https://machinehouse.media/services/content-production)[Podcast production ↗ ](https://machinehouse.media/services/podcast-production)[A connected show operation ↗ ](https://machinehouse.media/work/masoom-minawala)[Compare a recorded clip with a rewritten Short ↗ ](https://machinehouse.media/guides/clips-vs-original-shorts) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/podcast-guest-research [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ Podcast guest research # Know what to ask before the guest arrives. Guest shortlists, source research, interview questions and a brief for the host. We read the guest’s previous work to find specific experiences worth discussing and questions they have not already answered. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Service · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## Find the conversation in the research. The guest may have told their founding story in ten other interviews. We read those conversations so the host can ask about the decision everyone passed over, follow up on a contradiction or return to a story with a more useful question. For an ongoing show, guest research sits within its editorial direction. We consider audience fit, expertise, a useful episode promise and the host’s own curiosity. [Our work with Masoom ](https://machinehouse.media/work/masoom-minawala) connects those decisions to the show’s strategy, questionnaires, trailers, Shorts and packaging. ## Build the information behind the questions. We examine the relevant source material: previous interviews, writing, books, research, recent work and other public accounts. The aim is to know which stories have already been told, where the guest has been specific, and what remains worth asking. Useful findings become source-linked notes. Direct quotes stay attached to their source. An interpretation remains an interpretation. Where accounts conflict, the host needs a precise, fair question rather than an invented resolution. ## A brief built for the host. | Part | What it does | Episode promise | Defines what the audience should understand by the end. | Guest context | Summarizes the experience and work relevant to this conversation. | Strong material | Collects specific stories, claims, decisions and examples with sources. | Unexplored territory | Identifies useful gaps in earlier coverage or new developments. | Question arc | Moves from an accessible opening into depth, tension and a satisfying close. | Follow-ups | Gives the host ways to ask for a scene, an example, a tradeoff or an explanation. | Recording priorities | Separates essential questions from optional routes through the conversation. ## Twelve guests. One demanding shoot. For a NYC shoot of The Masoom Minawala Show, the research system covered 12 guests and 1,000+ sources in six hours , producing an 80-page brief per guest. Rishwajeet described the work in [his April 2026 essay ](https://machinehouse.media/writing/twelve-guests-six-hours). That was the research volume. For the recording, we turn it into a shorter questionnaire with priorities and follow-ups the host can find during the conversation. ## Ask for a moment. Then understand the decision. Consider an illustrative interview with a founder discussing a difficult product decision. “Tell us about your journey” invites a biography. “Take me to the meeting where you decided to stop building it” invites a scene. “What evidence changed your mind?” opens the decision. “Who disagreed, and what were they seeing?” reveals the tradeoff. These are example questions, not excerpts from a client questionnaire. The actual questions should emerge from the guest’s history and the show’s audience. We write for the host’s voice, then review the brief with the person who will use it. ## Carry the research into the release. When the scope includes the wider show, the conversation informs the [titles and thumbnail directions ](https://machinehouse.media/services/youtube-packaging), trailer and [short-form work ](https://machinehouse.media/services/long-form-to-shorts). After release, audience response helps refine the next guest choice and question arc. That connection is part of our [podcast production service ](https://machinehouse.media/services/podcast-production). ## Questions about the work. What do you deliver for a podcast guest? + The agreed scope can include an episode promise, source-linked research, guest context, story and angle notes, a question arc, follow-ups and a recording questionnaire. Can you help choose guests as well as research them? + Yes. We can connect guest selection to the show’s audience, editorial direction and performance history, then research the selected guests. Will the host receive an 80-page document every time? + Research depth and the recording brief depend on the episode. The published NYC example used an 80-page brief per guest. We scope the material around what the host and production actually need. Can you work with our existing podcast production team? + Yes. We can own research and questionnaires while your team records, edits and publishes, or connect the work to a wider production engagement. ## Keep exploring. [Podcast production ↗ ](https://machinehouse.media/services/podcast-production)[The Masoom Minawala Show ↗ ](https://machinehouse.media/work/masoom-minawala)[Inside the research system ↗ ](https://machinehouse.media/writing/twelve-guests-six-hours)[Choose the guest ↗ ](https://machinehouse.media/guides/podcast-guest-selection)[A complete host research brief ↗ ](https://machinehouse.media/guides/interview-research-brief) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/youtube-packaging [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ Titles & thumbnail strategy # The work deserves a way in. YouTube packaging connects an idea to the person who might watch it. We develop titles and thumbnail directions around the audience, the story and the promise the video can actually deliver. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Service · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## What should the viewer want to find out? A viewer sees the title and thumbnail before any of the work inside the video. We decide which question, consequence or surprise to put there, and what the image can add without repeating the title. We develop packaging alongside the premise and script, then check it against the finished work. The title, opening and episode need to agree on what the viewer will get. A beautiful image cannot repair an idea that has no reason to exist. ## What the engagement covers. - A review of the channel’s existing packages, audience and available performance context. - Distinct title directions, each with an editorial reason behind it. - Thumbnail concepts, reference direction and a brief a designer can execute. - Review of the title and image as a pair, including what is legible at feed size. - Testing and interpretation where the channel and video are eligible, followed by the next packaging decision. Rishwajeet stays involved in the creative call. Production can run with your designer or our team, depending on the scope. ## Three titles. A measurable difference. In Rishwajeet’s [June 2026 account of a packaging test ](https://machinehouse.media/writing/the-15-lakh-question), three title variants were compared on the same video: | Title | Reported watch-time share | India’s IT Dream Is Falling Apart | 32.1% | Why India’s IT Dream Is Falling Apart | 28.8% | The Indian IT Dream is Dead | 39.2% — winning variant The creative interpretation was that “dead” gave the story a stronger sense of finality. The winning variant led the first by 7.1 percentage points of watch-time share . That is the result of this test; it is not a measured 7.1% increase in total video views, and it does not mean the same word wins for every subject. The thumbnail made the argument visually: a gravestone, a marigold garland and the tricolour. Those choices made an Indian story recognizable at a glance. The example shows why packaging combines judgment with testing. ## Test a creative hypothesis. A useful comparison changes something consequential: the angle, emotional register, specificity or image concept. Several near-identical variants can tell you very little. We define what the test is meant to teach before choosing the options. YouTube’s native tool supports up to three titles, thumbnails or paired combinations on eligible videos. It evaluates results using watch time and can report no clear winner. We keep the test’s actual verdict in the review. [YouTube’s A/B testing guidance ](https://support.google.com/youtube/answer/16391400?hl=en). ## The opening must keep the promise. Packaging is part of the full viewing experience. Once someone clicks, the opening has to establish the situation, stakes and reason to keep watching. That is why a packaging engagement often connects to [strategy and script reviews ](https://machinehouse.media/services/content-strategy). After publication we look at the result in context: the test, audience response, traffic and the video itself. The useful finding returns to the next idea and package. See how that operates across [The Masoom Minawala Show ](https://machinehouse.media/work/masoom-minawala). ## Questions about the work. Do you design thumbnails or just advise on them? + The scope can include strategy, title development, thumbnail concepts, designer briefs, creative reviews and production with our team. We can also work with your existing designer. Can you improve packaging for already published videos? + Yes. We can review existing videos, develop stronger packages and plan eligible tests. We first check that the new promise is supported by the actual video. How do you judge a winning title or thumbnail? + We use the platform test’s actual result and review it alongside the creative hypothesis and audience context. Watch-time share, CTR and total views are different measurements and should not be treated as interchangeable. Can packaging guarantee that a video goes viral? + No. Packaging earns attention for the idea, but the subject, audience, viewing experience and distribution all matter. We make a reasoned creative choice and learn from the result. ## Keep exploring. [Scaler: the premise, the package and the results ↗ ](https://machinehouse.media/work/scaler)[Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Channel audit ↗ ](https://machinehouse.media/services/youtube-channel-audit)[The full packaging test ↗ ](https://machinehouse.media/writing/the-15-lakh-question) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-production-costs [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Content production costs # What goes into a content production quote? Compare the work covered by a quote: research, writing, recording, editing, packaging and release. Use the sample brief to specify your formats, volume, team and review process. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated October 13, 2026 ## Start with the scope. Machine House scopes content production around the formats, release rhythm and stages we own. Most ongoing engagements run on quarterly retainers. We prepare a written proposal with the team, deliverables and investment after an intro call. We do not publish a universal per-video rate because two videos of the same length can involve very different work. A hosted interview, a research-led film and a founder’s camera-to-person explanation need different preparation, people and post-production. ## What changes the quote. | Driver | What to specify | Strategy | A defined brief, a new show direction, a channel audit or ongoing editorial planning. | Research and writing | Source depth, subject complexity, script length, questionnaires and review needs. | Recording | Locations, shoot days, crew, equipment, travel and who arranges the setup. | Post-production | Edit complexity, graphics, sound, finishing and the condition of the source media. | Packaging | Title directions, thumbnail concepts, design production and testing responsibilities. | Additional outputs | Trailers, recorded clips, original Shorts, language versions and other adaptations. | Publishing and review | Uploads, metadata, release coordination, analytics and editorial reviews. | Working process | Turnaround, revision rounds, approvers and changes to the agreed brief. ## Compare complete responsibilities. These are illustrative scope shapes, not price packages. They show why “four videos a month” is too little information to compare proposals. | Scope | The production responsibility | Edit support | You provide the footage, selected material and a clear brief. The engagement covers the agreed edit and finishing. | Editorial and production | The team develops the premise, researches and writes, plans recording, then edits and packages the release. | Full content operation | Strategy, slate, research, production, packaging, publishing and performance review run as one ongoing engagement. A lower quote may cover fewer stages. Put the responsibilities side by side before deciding which proposal offers the better fit. ## An example of a quote-ready brief. Illustrative monthly scope Four interview episodes, recorded across two planned shoot days. Guest research and questionnaires for each. Four long-form edits, four trailer edits and eight standalone short-form pieces. Titles and thumbnail production for each full episode. Publishing on the agreed channels and a monthly performance review. The host approves guest selection and question direction. One person consolidates edit feedback. The quote identifies recording costs, travel, revision rounds, source-file delivery and any work outside this scope. This example is a way to make a proposal comparable. It is not a recommended output quota or a description of a named client’s contract. The right release rhythm follows your audience, team and creative ambition. ## Budget for research, writing and packaging. Rishwajeet’s essay [“You can pour ₹15 lakh into a video and set it on fire” ](https://machinehouse.media/writing/the-15-lakh-question) uses a production-budget thought experiment to explain the importance of packaging. That amount is not a Machine House price or a disclosed client invoice. The practical point is that the budget should account for the idea, script and package as well as the shoot and edit. Those decisions determine what the production is trying to accomplish. A complete scope gives them an owner. ## Agency, in-house team or a shared operation? An in-house team gives the business dedicated capacity and close access to its people and product. It also needs management, an editorial lead and enough sustained work for the roles hired. An agency engagement can assemble several disciplines around an agreed scope, but it still needs a client owner, timely access and clear decisions. A shared operation is often worth comparing too: keep the people who know the business, then commission the research, writing, production or strategy capacity that is missing. Compare the same release plan across all three options. Include the founder’s review time, internal coordination, software, equipment, freelance specialists and the work needed between releases. Ask each prospective partner to walk through a relevant project: the initial problem, their actual contribution, the decisions they made, what the client supplied and what was delivered. A show’s total views alone do not establish which part of the work that partner owned. Use the [editor brief ](https://machinehouse.media/guides/video-editor-brief) and [research packet ](https://machinehouse.media/guides/research-to-script) as examples of the level of handoff detail you can ask to inspect. Comparing YouTube partners? Use the [agency selection guide and editable brief ](https://machinehouse.media/guides/choosing-youtube-growth-agency) to examine their diagnosis, case evidence and the work left with your team. ## What to bring to the first conversation. - Your existing channel or a few examples of what you want to make. - The audience and business outcome the work should serve. - The formats and approximate release rhythm you have in mind. - Your current team, recording setup and the stages you want us to own. - Timeline, budget constraints and who approves the work. We can then decide whether you need [strategy ](https://machinehouse.media/services/content-strategy), a focused production engagement or [the full content operation ](https://machinehouse.media/services/content-production). ## Questions about the method. How much does Machine House charge for content production? + We prepare a written quote after understanding the formats, release rhythm, recording requirements and stages we own. Most ongoing engagements run on quarterly retainers. Is recording included in a retainer? + It depends on the agreed scope. The proposal identifies shoot planning, crew, location, equipment, travel and other recording responsibilities so you can see what is included. Are Shorts and thumbnails included with a full episode? + Only when they are included in the written scope. We specify the number and kind of additional outputs, creative work and production responsibilities. Can we start with strategy before committing to production? + Yes. You can begin with a channel audit or strategy engagement, then decide how to produce the agreed direction with your team or ours. ## Keep exploring. [Content production ↗ ](https://machinehouse.media/services/content-production)[Podcast production ↗ ](https://machinehouse.media/services/podcast-production)[Begin with an audit ↗ ](https://machinehouse.media/services/youtube-channel-audit)[Make reviews decide something ↗ ](https://machinehouse.media/guides/content-approval-workflow)[Content capacity planner ↗ ](https://machinehouse.media/labs/content-capacity-planner) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-performance-prediction [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Content performance prediction # Forecast the result before you choose the idea. Choose what to forecast, record it before production and compare it with the release. Masoom’s guest-selection example shows why a forecast needs a specific audience goal. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## What a forecast is for. A celebrity guest might bring more views; another guest might bring more subscribers. Before ranking the next guests, decide which result you want. The forecast should help with that booking decision, and you need to save it before the episode goes out. At Machine House, prediction sits inside the content operation. Rishwajeet brings the creative judgment. The system supplies relevant history, research and a structured way to assess the idea. The [Masoom Minawala case study ](https://machinehouse.media/work/masoom-minawala) shows how that connects to guest selection, questionnaires, trailers, Shorts and packaging. ## Choose the outcome before scoring the idea. “Will this do well?” leaves too much undefined. A launch designed for reach and an episode designed to deepen a returning audience have different jobs. Even a views forecast needs a measurement window: the first week, first month or a longer period. Write down the metric, window and relevant baseline. Keep episode views, subscribers gained, watch time and business inquiries separate. A composite score can help organize a review, but it should not hide which outcome the team actually wants. September 2026 update: a views forecast needs a definition as well as a date range. YouTube’s public counter changed on 24 August. Confirm that historical baselines and the outcome being forecast use the same measurement; do not apply a historical view target mechanically to a different counter. [Read the view-counting analysis. ](https://machinehouse.media/writing/youtube-changed-what-a-view-means) ## Describe the content in editorial terms. Guest names and upload dates do not explain the whole episode. We code relevant features so the team can compare ideas more meaningfully. The Masoom case record describes 12 dimensions: topic, authority, emotional register, specificity, hook, guest archetype, packaging, runtime, release window, tone, format and subject novelty. The categories need stable definitions. Two reviewers should know what “specific” means in this dataset. And the inputs for a pre-production forecast must be information available before the outcome; including later performance in the inputs would defeat the purpose. ## A larger name may serve a different goal. In [Rishwajeet’s December 2025 essay ](https://machinehouse.media/writing/celebrity-vs-expert), the Shloka Ambani episode was reported at 300K views, while a Mitesh Rajani episode was reported at 200K. The latter generated more subscribers and comments in that account. Exact subscriber-conversion counts were not published. That observation gives guest planning a better question: are we trying to maximize this episode’s reach, or develop the audience that wants to return? It does not prove that fame causes weaker community growth. The topic, package, timing and audience also differ. The editorial lesson is to make the objective explicit before ranking the next guest. ## Keep an honest prediction record. | Before release | After the agreed window | Save the premise, inputs, forecast and why the team chose it. | Record the actual outcome using the same metric and window. | Compare against a simple baseline from relevant earlier work. | Check whether the forecast was more useful than that baseline. | Record a range or level of confidence where justified. | Review the misses, including systematic over- or underestimation. | Separate the model’s assessment from the final creative decision. | Identify which learning should change the next brief or decision. A few successful predictions do not establish reliable accuracy. Evaluation should include unsuccessful ideas and forecasts, use work that was not used to fit the model, and preserve the order in which information became available. ## A finding needs somewhere to go. Audience response becomes useful when it changes the work: a guest shortlist, a research priority, a question, an opening, a trailer or a package. We connect those decisions through [content intelligence systems ](https://machinehouse.media/services/ai-content-systems) so the learning remains available to the team. ## Questions about the method. Can you predict exactly how many views a video will get? + We use forecasts to support editorial decisions, not to promise an exact outcome. Their usefulness depends on the data, the content and how they perform against real published results. Is an AI score the same as a performance model? + No. A score can summarize editorial criteria. A performance model needs a defined outcome and evaluation against results. The distinction matters when deciding how much confidence to place in either. What if the channel has little historical data? + Start with audience research, explicit editorial criteria and a simple baseline. Record the decisions and outcomes as you publish, then assess whether richer modeling becomes useful. How does prediction connect to production? + The assessment informs the chosen premise, guest, structure and package. After release, the response returns to strategy, research, writing and packaging for the next piece. ## Keep exploring. [Masoom: the updated show results ↗ ](https://machinehouse.media/work/masoom-minawala#metrics)[AI content systems ↗ ](https://machinehouse.media/services/ai-content-systems)[The Masoom Minawala Show ↗ ](https://machinehouse.media/work/masoom-minawala)[Celebrity guests and expert episodes ↗ ](https://machinehouse.media/writing/celebrity-vs-expert) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/work/indian-silicon-valley [Home ](https://machinehouse.media/)/ [Work ](https://machinehouse.media/#cases)/ Indian Silicon Valley # Three agencies. One content operation. Rishwajeet managed agencies across short-form, long-form and production for Indian Silicon Valley, a content-led angel syndicate. The engagement ran from 2024 to 2026. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Case study · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## Content as part of the business. Indian Silicon Valley’s content was made by three agencies. Rishwajeet managed their work across short-form, long-form and production during the 2024–26 engagement. ## Three agencies. Connected work. The role covered coordination across the three agencies and their releases. [Rishwajeet’s published work history ](https://machinehouse.media/rishwajeet#record) records that management scope alongside the audience results below. ## The reported growth. | Measure | Historical result | Weekly views | 30K → 150K | Views over five months | 7M+ | Subscribers over five months | +50K These figures are recorded in the founder’s public work history and the June 2026 case record. They describe the channel’s outcomes during the engagement. They do not isolate an individual agency’s contribution or establish a measured investment-business return. ## See the channel and the wider practice. [Watch Indian Silicon Valley on YouTube ](https://www.youtube.com/@IndianSiliconValley). Read about [YouTube channel management ](https://machinehouse.media/services/youtube-growth), or see the complete editorial operation behind [The Masoom Minawala Show ](https://machinehouse.media/work/masoom-minawala). ## Questions about the work. What was Rishwajeet’s role at Indian Silicon Valley? + His published work record describes content operations during 2024–26, managing three agencies across short-form, long-form and production. What growth is reported for Indian Silicon Valley? + The June 2026 case record reports weekly views growing from 30K to 150K, with 7M+ views and 50K additional subscribers over five months. Are these current channel totals? + No. These are historical engagement results from the published work record. They are not presented as current subscriber or lifetime view totals. ## Keep exploring. [YouTube channel management ↗ ](https://machinehouse.media/services/youtube-growth)[Founder and brand content ↗ ](https://machinehouse.media/services/founder-brand-content)[Content production ↗ ](https://machinehouse.media/services/content-production) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/work/the-karighars [Home ](https://machinehouse.media/)/ [Work ](https://machinehouse.media/#cases)/ The KariGhars # A brand channel. A stronger reason to watch. Rishwajeet worked on the YouTube channel’s scripts, packaging and production. His June 2026 work record reports a 478% increase in watch time in the first month. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Case study · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## The work around the viewing experience. The engagement covered the title and thumbnail, the script and the production of the video. These responsibilities are recorded in [Rishwajeet’s 2025–26 portfolio ](https://machinehouse.media/rishwajeet#record). ## Three connected responsibilities. | Responsibility | Its role in the content | Packaging | Give the intended viewer a clear reason to open the video. | Scripting | Build a structure that develops the idea and delivers the promise. | Production | Make the viewing experience support that structure and creative direction. ## 478% more watch time in month one. The existing case record, reported June 2026, records a 478% increase in watch time during the first month of the engagement. This is a historical result for the channel, not a forecast for another brand. The public record does not break out the contribution of packaging, scripting and production individually. It also does not publish a sales or lead-attribution result for this period. Watch time is the outcome reported here. ## Look at the work in context. [Explore The KariGhars ](https://www.instagram.com/the_karighars), and read the [packaging approach ](https://machinehouse.media/services/youtube-packaging) or [full channel-management scope ](https://machinehouse.media/services/youtube-growth). ## Questions about the work. What work did Rishwajeet do for The KariGhars? + The published portfolio records packaging, scripting and a production overhaul for the YouTube channel during the 2025–26 engagement. What result is reported for The KariGhars? + The June 2026 case record reports a 478% increase in watch time in the first month of the engagement. It is a historical channel result. Does this case include sales or lead-generation results? + No. The outcome publicly reported here is watch time. The case does not claim a measured sales or lead-generation effect. ## Keep exploring. [Founder and brand content ↗ ](https://machinehouse.media/services/founder-brand-content)[Titles and thumbnails ↗ ](https://machinehouse.media/services/youtube-packaging)[YouTube channel management ↗ ](https://machinehouse.media/services/youtube-growth) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/work/roshan-kotla [Home ](https://machinehouse.media/)/ [Work ](https://machinehouse.media/#cases)/ Roshan Kotla # A small channel. Three hits in a row. Roshan’s channel had a median of around 100 views per video. We helped turn his experience in software and hiring into a viral IT-industry essay—and followed it with two more hits. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Case study · Updated September 27, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## From roughly 100 views to a viral breakout. Before the breakout, Roshan’s videos had a median of around 100 views, as recorded in [Rishwajeet’s July 2026 account ](https://machinehouse.media/writing/did-it-twice-more). He had experience running a software business and hiring people, but his essays had struggled to reach an audience. The first essay we made together was “The Indian IT Dream is Dead.” Published on 16 April 2026, it became the channel’s viral breakthrough, reaching 791,165 views by 7 September. ## 791K. 93K. 91K. Three consecutive hits. | Published essay | Release | Public views | [The Indian IT Dream is Dead ](https://www.youtube.com/watch?v=o321t-rfScA) | 16 April 2026 | 791,165 | [Why India’s Graduates Can’t Find Jobs Anymore ](https://www.youtube.com/watch?v=yTF_aczKJzg) | 22 May 2026 | 93,128 | [The Middle-Class Job Dream Is Dead ](https://www.youtube.com/watch?v=FKc_IY2oQK4) | 19 June 2026 | 91,065 975,358 combined views as of 7 September 2026. These are cumulative public views for the three named releases, at different ages. The next two essays each passed 90,000 views. A small channel had landed a breakout and sustained its reach across two more releases. ## How we built on the first hit. After the breakout, an IT-industry sequel was the obvious option. We went back to the discussion under the first video to decide what the next one should answer. The responses pulled in different directions. Experienced IT workers challenged the idea that their industry was finished. Students and recent graduates asked whether paying for an engineering degree would still lead to a job, and what they should do if campus recruiters never arrived. We chose to follow the graduate's problem. The next essay examined campus placements: the route from a degree to a first job, and the expectations a family builds around it. Roshan could approach that subject as someone who had been through placements himself and now made hiring decisions at his own firm. The third essay widened the subject to the middle-class career promise. ## Roshan’s memory became part of the script. Roshan recalled preparing for placements: finding the right sky-blue shirt, shaving for the first time and copying a résumé from a senior who had already been hired. He also remembered the gap between the salary announced before the interview and the amount that reached his bank account. Those details gave us a scene to write. A viewer could picture the day a student expected adult life to begin, before the essay examined what happened when that route failed. The shirt and résumé passage appears in the [delivered graduate video ](https://www.youtube.com/watch?v=yTF_aczKJzg). Research helped us choose the question; Roshan supplied a memory that made it his story to tell. ### About the comment sample The comment sample behind this decision was collected on April 19, 2026, a little over three days after the first video's publication. Its 443 records include 351 viewer comments and 92 replies from the channel. We used the discussion to develop an editorial hypothesis; it was not a representative survey. The working materials explain the choice, while the dated viewing totals above show what happened after publication. They do not isolate how much the research contributed to those results. ## The work behind all three. Machine House developed the research, argument, outline, script and packaging. Roshan’s examples and feedback shaped the drafts; he recorded the essays. After each release, we read the response before choosing the next subject. ## The account from inside the work. Rishwajeet wrote about the decision while the third video was still growing. Read [“One video did 7.7 lakh views. Then we did it twice more.” ](https://machinehouse.media/writing/did-it-twice-more) The essay preserves its July numbers; this case updates the view counts to September. ## Questions about the work. What did Machine House do for Roshan Kotla? + Machine House worked on research, editorial thesis, outlines, scripts, original Shorts and packaging. Roshan supplied his lived experience, hiring stories, feedback, final creative judgment and on-camera delivery. What results does this case cover? + The IT Dream, Graduates and Middle-Class essays had 791,165, 93,128 and 91,065 public views respectively on 7 September 2026: 975,358 in total. The figures cover those three releases, not the entire channel. ## Keep exploring. [Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Content production ↗ ](https://machinehouse.media/services/content-production)[YouTube growth ↗ ](https://machinehouse.media/services/youtube-growth)[More client work ↗ ](https://machinehouse.media/#cases)[Research and scriptwriting for your next video ↗ ](https://machinehouse.media/services/video-scriptwriting) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/work/scaler [Home ](https://machinehouse.media/)/ [Work ](https://machinehouse.media/#cases)/ Scaler # Make the technology matter to someone. Machine House developed research, scripts, production briefs and packaging for 17 Scaler releases. Together they passed 150,000 views; the leading essay, about the next valuable software role, passed 91,000. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Case study · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## A technology subject needs a human stake. The leading essay in this slate asked what the next valuable software role would look like. Other strong releases examined pressure on AI startups and India’s place in the AI industry. The subjects gave viewers a way to think about work and the industry they were entering. Our work for Scaler connects source research, the central argument, the script, visual and production briefs, and packaging. It then returns to the published video to see how that promise landed. ## Read the slate. Keep the premise in view. We reviewed 17 releases alongside their editorial origins, packaging and audience response. The review combined public performance histories with available Studio data and 146 comments across six leading videos. ## 150K+ views. Seventeen releases we can point to. | Selected release | Published | Public views | [The Next Million Dollar Software Job Isn’t Coding ](https://www.youtube.com/watch?v=vhGbd0fQMIo) | 11 August 2026 | 91,178 | [OpenAI Keeps Killing Its Customers’ Startups ](https://www.youtube.com/watch?v=HfAxbWWWuTY) | 20 August 2026 | 18,841 | [India Has Zero Frontier AI Models. Here’s Why ](https://www.youtube.com/watch?v=ip7bm9jrqpA) | 22 August 2026 | 16,012 150,373 public views across 17 verified Machine House releases , published 3 July–24 August 2026. Snapshot: 7 September 2026. The total counts only the releases matched to our work, at their current ages. ## Keep the argument intact through the handoff. The script needs evidence, a sequence and a voice that can carry the argument. The production brief translates that into visual beats, examples and the material the team needs to make the piece. Titles and thumbnail directions develop around the same premise. After publication, the review comes back to the writing: whether the opening reached the promise, whether viewers understood the argument, and what deserves a stronger explanation next time. ## Questions about the work. What work does Machine House do for Scaler? + The work includes technology research, editorial development, scripts, visual and production briefs, packaging and review of the published slate. Which videos are included in the 150K+ total? + The total covers 17 published releases verified against Machine House’s Scaler work record, released from 3 July to 24 August 2026. They had 150,373 cumulative public views on 7 September 2026. What did the performance review change? + The review sharpened the editorial brief around concrete career and industry stakes, a clear argument and delivering the answer promised by the title. Those findings inform later ideation, scripts and packaging. ## Keep exploring. [Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Content production ↗ ](https://machinehouse.media/services/content-production)[YouTube growth ↗ ](https://machinehouse.media/services/youtube-growth)[More client work ↗ ](https://machinehouse.media/#cases)[A complete research-to-script example ↗ ](https://machinehouse.media/guides/research-to-script)[An editor brief with source audio and cut sheet ↗ ](https://machinehouse.media/guides/video-editor-brief)[Research and scriptwriting for your next video ↗ ](https://machinehouse.media/services/video-scriptwriting) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/work/rocket-health [Home ](https://machinehouse.media/)/ [Work ](https://machinehouse.media/#cases)/ Rocket Health # From protein research to a finished film. Rocket Health’s protein film reached 72,000 views. Machine House developed the research, script, visual direction and production briefs, working with Studio 73 on the production. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Case study · Updated September 9, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## Give a health brand something worth watching. Machine House worked with Studio 73 on films about health and culture for Rocket Health. On the protein film, our work covered the research, script and documents for the host, crew and editor. ## Make the research part of the viewing experience. The protein film moved from source research to a sustained argument, then into a full script. The production material developed the visual beats, host prompts, props and edit instructions needed to carry that argument on camera. ## One researched film. 72,000 views. [Watch the protein film on Rocket Health’s channel ↗ ](https://www.youtube.com/watch?v=cEd6wSsp0Ak) | Release | Published | Public views | Rocket Health’s protein film | 17 July 2026 | 72,000 Public view count as of 7 September 2026. This result belongs to the named film, made with the host and production team. ## The brief reaches beyond the words. The work included a full-length script, supporting evidence, visual direction and practical host and production documents. That lets the people shooting and editing understand both what to make and why the story is built that way. Original short-form scripts were also developed from the longer work. Each short needs its own premise and payoff; the long-form film remains the source of the underlying insight. ## Questions about the work. What did Machine House produce for Rocket Health? + Machine House developed editorial direction, source research, scripts, visual direction and production-ready briefs for health and culture films, working with Studio 73 on production. What does the 72K result refer to? + It is the cumulative public view count on 7 September 2026 for Rocket Health’s protein film, published on 17 July 2026. ## Keep exploring. [Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Content production ↗ ](https://machinehouse.media/services/content-production)[YouTube growth ↗ ](https://machinehouse.media/services/youtube-growth)[More client work ↗ ](https://machinehouse.media/#cases)[A complete research-to-script example ↗ ](https://machinehouse.media/guides/research-to-script)[An editor brief with source audio and cut sheet ↗ ](https://machinehouse.media/guides/video-editor-brief)[Research and scriptwriting for your next video ↗ ](https://machinehouse.media/services/video-scriptwriting) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/research-to-script [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Research to script # Turn research into a script worth hearing. Follow a map-projection video from its source packet to the finished script. See the argument, a passage we cut and the evidence behind each section. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 24, 2026 ## Give the research a job. “Map projections” is a subject you could research for weeks. “If every flat world map distorts something, how should we choose one?” gives the writer a film to make. The example below follows that question through the sources, outline and spoken script. Our worked example: a curious viewer recognizes a world map but has never chosen a projection. They need to understand why a map suited to navigation can mislead an area comparison. The argument is that a map should be judged by its task and the property that task needs to preserve. This guide follows an original teaching example, created for this page. It is not a client commission. You can [download the complete research packet and script ](https://machinehouse.media/resources/research-to-script/research-packet.txt), then follow its [handoff to an editor ](https://machinehouse.media/guides/video-editor-brief). Working on your own film? [Download the blank research-to-script brief ](https://machinehouse.media/resources/research-to-script/blank-research-brief.txt) or [see how to use it ](https://machinehouse.media/guides/research-to-script#your-film). ## Collect evidence around claims. Build the source packet around the statements the film must earn. For each claim, retain the original source, the passage or section that supports it, the date checked and the boundary of the claim. Keep observations, interpretations and proposed lines separate. That makes it possible to change a sentence without losing the evidence behind it. | Claim to establish | Source and boundary | One flat world map cannot preserve every property. | [USGS: uses of map projections ](https://www.usgs.gov/faqs/how-are-different-map-projections-used). This is a constraint on representing the curved world on a plane, not a claim that every regional map is equally distorted. | Standard Mercator represents constant-bearing routes as straight lines and preserves local angles. | [Esri: Mercator ](https://doc.esri.com/en/arcgis-pro/latest/help/mapping/properties/mercator.html). Do not call that route the shortest route. Do not silently equate standard Mercator with Web Mercator. | An area comparison needs a projection that preserves relative area. | [Esri: selecting a projection ](https://support.esri.com/en-us/knowledge-base/select-a-suitable-map-projection-or-coordinate-system-1-000006113). The purpose determines which property matters; changing shapes is still a tradeoff. | There is no universally best flat projection. | [USGS: Map projections ](https://www.usgs.gov/publications/map-projections). Choose for the purpose and geographic extent. These sources were checked on September 7, 2026. The packet assigns them IDs so each factual paragraph can point back to the right support. The opening choice and the language connecting the facts are original writing. A source can support the science without supplying the story. ## Turn the outline into an argument. A topic outline tells you what subjects appear. An argument tells you why one part follows another. Read your outline as a sequence of changes in the viewer’s understanding. If two sections can swap places without affecting anything, investigate whether the relationship between them is clear enough. | Topic outline | Argument for this film | Definition of a map projection | Give one viewer two jobs: compare land area and follow a compass bearing | History of Mercator | Establish the impossible requirement: preserve every property on one flat world map | Types of projection | Explain why the Mercator tradeoff can be useful for one job | Advantages and disadvantages | Show why that same tradeoff interferes with an area comparison | Conclusion | Return to the viewer's task and make the selection The revised outline uses the same small body of research. It introduces a choice, explains the constraint, makes one compromise useful, shows where that compromise causes a problem, and returns to the choice. The writer has done the work of deciding which fact the viewer needs next. For a research-heavy YouTube script, this is also where you decide what deserves a scene, a demonstration, a comparison or an expert explanation. Do not automatically convert each source into its own section. A section earns its place by changing the answer. ## Write the version someone can hear. Spoken explanation needs its own structure. A viewer cannot glance back at the previous paragraph as easily as a reader can. Introduce the object before its qualification; give a comparison a clear purpose; read the transitions aloud. The [BBC’s writing-for-broadcast lesson ](https://downloads.bbc.co.uk/academy/youngreporter/Lesson%20Plans%20-%20Lesson%20Five%20-%20Transcript%20English.pdf) is a useful reminder to write for the medium and test the words by speaking them. Here is the opening from the example. Notice that the technical term arrives after the viewer has a reason to care about it. Imagine choosing a world map for a classroom. You want students to compare how much land different countries occupy. A sailor planning a route along a constant compass bearing needs a map for a different reason. The same projection will not serve both tasks equally well. Then remove passages that delay the question. This paragraph was written as a deliberately weak alternative: Cartography has a rich history, from early representations of the world to today's sophisticated digital mapping systems. Over the centuries, mapmakers have developed many projections, each with advantages and disadvantages. Understanding these projections is essential to appreciating the maps we use every day. It promises importance without demonstrating it. Removing it gets the viewer to the actual constraint sooner. This is an editorial decision about this film, not a rule that historical context is always unnecessary. ## Read the complete script. ### A01 — Put the viewer inside a choice Imagine choosing a world map for a classroom. You want students to compare how much land different countries occupy. A sailor planning a route along a constant compass bearing needs a map for a different reason. The same projection will not serve both tasks equally well. ### A02 — Explain the constraint The Earth is curved, and the map is flat. When a projection turns one into the other, some properties change. You can preserve relative area, or preserve angles locally, but a flat map of the whole world cannot preserve every property at once. The choice depends on what the map needs to show. ### A03 — Make the apparent flaw useful On a standard Mercator map, a straight line can represent a constant compass bearing. That makes it useful for navigation, although the line is not generally the shortest route. Mercator also preserves angles locally. But as you look towards the poles, the areas become increasingly exaggerated. A map with a useful navigation property can give a misleading impression of how large a country is. ### A04 — Return to the first job For the classroom comparison, choose a projection that preserves relative area. The countries may have unfamiliar shapes, because shape is one of the properties that changes. But their relative sizes will be suitable for the comparison you asked the students to make. ### A05 — Keep the explanation honest You also need to specify the region. A projection suited to one small area can keep the relevant distortions small there, even if it would be a poor choice for a world map. Tell the mapmaker which part of the Earth you need to show, as well as what people will use the map for. ### A06 — Resolve the opening choice The classroom map and the navigation map can therefore look different for good reasons. Put their purpose beside them: this one lets students compare land area; this one lets a sailor draw a constant-bearing route. Someone looking at either map can then see what it preserves and what they should not use it to judge. Scope note: Mercator here means the standard projection. Web Mercator is a distinct variant. A constant-bearing route is not generally the shortest route. The example teaches a selection principle; it is not a navigation tutorial. ## Send the decisions with the script. Before recording, separate the spoken copy from visual instructions and factual support. A script full of parenthetical production notes is hard to perform. A clean script with no explanation of its evidence is hard to check. Keep them connected through paragraph or scene IDs. For this example, A03 needs the audience to understand the constant-bearing advantage and area distortion. A decorative rotating globe would not establish that relationship. A useful visual instruction names the two properties to show, keeps the map labels legible and asks for a factual check on the diagram. If an illustration cannot be produced accurately within the budget, revise the explanation before the edit depends on it. Give the editor the audience question, the argument, the complete script, the source packet, the intended shape of the cut and the qualifications that must survive shortening. Our companion [video editor brief ](https://machinehouse.media/guides/video-editor-brief) includes actual source audio, a timecoded transcript and an assembled example cut. The final read-through asks a practical question: could another person explain why each paragraph is here, verify its factual claims and understand what the viewer is meant to see? If that still depends on a call with the writer, the handoff needs another pass. ## Build a brief for your own film. Copy the [blank research-to-script brief ](https://machinehouse.media/resources/research-to-script/blank-research-brief.txt) into your document editor. Start with the question your film will answer and the evidence that could change your proposed answer. Leave the spoken draft empty until you have decided what the sources support. Repeat the claim record for each factual statement the argument depends on. Then use the passage records to connect the spoken words to their evidence and required visuals. If a claim is still unresolved before recording, name who will check it and decide whether to qualify it, remove it or delay the recording. For a short film, one claim record and a few passages may be enough. For a longer essay, repeat the argument sequence wherever the viewer needs another step. The completed brief can travel with the script into the [editor handoff ](https://machinehouse.media/guides/video-editor-brief#template). ## Questions about the method. What belongs in a research packet for a video script? + The audience question, proposed argument, claim-by-claim sources, dates checked, factual limitations, complete script and unresolved questions. Use stable paragraph or scene IDs to connect them. How do you avoid a video script that sounds like a report? + Organize the explanation around the viewer’s question and the changes in their understanding. Choose the next fact because of what the viewer now needs, then test the spoken draft aloud. Can I adapt the script and template? + Yes. Machine House permits adaptation of the original script and blank templates for your production work. Linked third-party sources retain their own rights. The example is teaching material, not a client result. ## Keep exploring. [The example’s editor handoff ↗ ](https://machinehouse.media/guides/video-editor-brief)[Research, scripts and end-to-end production ↗ ](https://machinehouse.media/services/content-production)[Educational video work for Scaler ↗ ](https://machinehouse.media/work/scaler)[Mental-health education for Rocket Health ↗ ](https://machinehouse.media/work/rocket-health)[Check the script’s claims ↗ ](https://machinehouse.media/guides/fact-checking-content)[Research and scriptwriting for your next video ↗ ](https://machinehouse.media/services/video-scriptwriting) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/video-editor-brief [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Video editor brief # Give the editor enough to start cutting. Listen to the source audio, follow the proposed cut and compare the two timelines. The example includes a full transcript, edit instructions and a blank brief. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Tell the editor what the cut must achieve. “Make it engaging” leaves the editor to work out what the video is saying. Start the brief with the audience’s question and the answer this cut needs to give. Then identify the source passages, what can go and which qualifications need to stay. The assignment: shorten “Which mistake should a map make?” into a spoken explanation of the area-versus-navigation tradeoff. Preserve the opening choice, the useful property of standard Mercator, the return to area comparison and the final decision. The geographic-extent explanation remains available in the full source and accompanying transcript. This is an original teaching exercise with synthesized demonstration narration. It is not Rishwajeet’s voice, a real interview or a released client film. The source file and cut below exist so you can inspect the handoff against actual media. For your own project, [download the editable editor brief template ](https://machinehouse.media/resources/editor-brief/blank-editor-brief.txt). The worked cut below shows how to fill its source, selection and review instructions. ## Give every instruction a stable source. Name the recording and version before writing timecodes. Use source time for locating material and sequence time for where it lands in the proposed edit. A timestamp without that distinction can send an editor to the wrong sentence. For multiple cameras or separate sound, specify the synchronized timeline and its frame rate before anyone cuts. This exercise uses elapsed audio time in minutes:seconds.milliseconds. The editable JSON also carries exact PCM frame offsets at 22,050 Hz. These audio sample positions are not video-frame timecode. A real video handoff should state the actual timeline format, including frame rate and any drop-frame convention. [Download the full source audio (WAV) ](https://machinehouse.media/resources/editor-brief/map-source.wav) · [Download the timecoded transcript (TXT) ](https://machinehouse.media/resources/editor-brief/source-transcript.txt) · [Download the cut sheet (JSON) ](https://machinehouse.media/resources/editor-brief/cut-sheet.json) The full source runs 01:44.772. The transcript retains every segment and the scientific scope notes. Each selection below points to that file. ## Make the proposed cut reconstructable. The cut keeps four of the six source segments. It removes the general flat-map constraint and the paragraph about geographic extent to concentrate on one comparison. Those omissions narrow what the short edit explains; they do not license a broader claim about every map. | Selection and purpose | Source in → out | Edit in → out | A01 — Put the viewer inside a choice | 00:00.000 → 00:14.648 | 00:00.000 → 00:14.648 | A03 — Make the apparent flaw useful | 00:33.342 → 00:55.918 | 00:15.148 → 00:37.724 | A04 — Return to the first job | 00:56.918 → 01:11.432 | 00:38.224 → 00:52.738 | A06 — Resolve the opening choice | 01:28.645 → 01:44.772 | 00:53.238 → 01:09.365 There is a half-second gap between selections. The resulting narration is 01:09.365 long; this is an audio assembly for inspecting the structure, with no finished visual treatment or music. Assembled example — synthesized narration [Download the example edit ](https://machinehouse.media/resources/editor-brief/map-edit.mp3) [Download the example edit (MP3) ](https://machinehouse.media/resources/editor-brief/map-edit.mp3). Its spoken text is A01, A03, A04 and A06 in the [transcript ](https://machinehouse.media/resources/editor-brief/source-transcript.txt). For an interview, quote the actual words and preserve the surrounding answer. Reordering two genuine sentences can still manufacture a connection the guest never made. The editor needs enough context to spot that, plus an explicit instruction to flag uncertainty before a context-changing cut is approved. ## Describe the work a visual has to do. “Add B-roll” specifies an asset category. It does not tell the editor what the viewer needs to understand. Tie visual directions to the passage they support, with separate notes for required facts and creative choices. | Passage | Useful direction | A01: two jobs | Present the classroom area comparison and the constant-bearing route as distinct tasks. Give the viewer time to read each. The exact graphic style is open. | A03: Mercator tradeoff | Show a checked standard Mercator diagram with the constant-bearing line and area distortion explained separately. Never label the line “shortest route.” Do not use a random online-map screenshot as equivalent evidence. | A04: area comparison | Use a properly labelled equal-area example for the comparison. Keep geographic extent and labels comparable; do not casually stretch an image to fit a frame. | A06: decision | Return to task, geographic area and protected property. The final visual should help the viewer apply the question, not introduce another fact. These are proposed visual instructions, not assets included in the download. The editor should flag the missing map artwork before finishing. Agree who commissions or sources it, who checks its accuracy, and whether its usage terms allow the intended release. ## Separate story approval from finishing. Review the first assembly for the argument, omissions and meaning of each cut. Review a later cut for pacing, graphic comprehension and the way sound and picture work together. Then check the approved export, captions, names, source credits and delivery settings. Keep the sequence proportional to the job: a small edit can combine stages when the reviewer can still make a clear decision. Consolidate comments into one decision list with an owner. “The opening feels slow” is a useful observation but an incomplete instruction. A stronger note identifies the location, the problem and the intended result: “At sequence 00:00, keep both jobs; remove the repeated setup before the comparison so the choice arrives once.” The exact change should follow the material rather than a universal rule to cut faster. Name who can approve story changes and who can approve the release. A polished export does not answer whether the argument was accepted. [Frame.io’s approval guidance ](https://blog.frame.io/2022/04/18/10-steps-for-client-approval/) also emphasizes defining the review process and keeping feedback clear. Plan accessibility at the brief stage. Decide who owns captions, a transcript and any necessary description of visual information; check that on-screen labels can actually be read in the intended format. [W3C’s media planning guidance ](https://www.w3.org/WAI/media/av/planning/) explains why these requirements belong before production, when choices are easier to make. A caption file alone is not proof that every accessibility need has been met. ## Use the blank brief on a real project. The [editable editor brief template ](https://machinehouse.media/resources/editor-brief/blank-editor-brief.txt) includes the decision, source inventory, cut instructions, visual requirements, sound, accessibility, review ownership and deliverables. Fill it with actual filenames and agreed settings. Remove fields that do not help the editor make a decision. For a YouTube long-form film, include the opening promise and where it gets paid off. For a podcast, preserve the guest’s meaning and the relationship between questions and answers. For a short, specify the standalone question and the context a new viewer lacks. For an educational film, make the evidence and diagram checks easy to locate. Finish the brief by naming unresolved choices. A missing graphic, an unverified quote or an undecided aspect ratio should be visible before edit time is spent on it. The handoff is ready when the editor can start with confidence, identify the real constraints and propose improvements without guessing what the piece is for. ## Questions about the method. What should a video editor brief include? + Audience, intended outcome, stable source files, the story or argument, proposed cuts, visual and sound requirements, factual and rights limitations, accessibility needs, reviewers and exact agreed deliverables. Are the example timecodes real? + Yes. They are measured from the supplied synthesized source audio, with source and edit positions shown separately. They use elapsed audio time and PCM sample offsets, not video-frame timecode. Should the editor follow every proposed cut exactly? + Treat the proposed cut as an inspectable starting point. The editor should propose improvements and flag missing context, assets or factual issues; identify which changes need story approval. ## Keep exploring. [The full research and script example ↗ ](https://machinehouse.media/guides/research-to-script)[End-to-end content production ↗ ](https://machinehouse.media/services/content-production)[Podcast production ↗ ](https://machinehouse.media/services/podcast-production)[Standalone shorts from long-form material ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Build a faithful trailer ↗ ](https://machinehouse.media/guides/podcast-trailer-editing)[Resolve approvals ↗ ](https://machinehouse.media/guides/content-approval-workflow)[Plan what the editor needs captured ↗ ](https://machinehouse.media/guides/content-shoot-plan)[Compare a recorded clip with a rewritten Short ↗ ](https://machinehouse.media/guides/clips-vs-original-shorts)[Accessible content production ↗ ](https://machinehouse.media/guides/accessible-content-production) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/why-agencies-miss-deadlines [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 15 JAN 2026 · 2 MIN READ # Why agencies miss deadlines On ownership, coordination and the delivery promise a creative agency makes. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/posts/rishwajeet-singh_every-creative-agency-will-lie-to-you-they-activity-7417505973759774720-aEUY) Every creative agency will lie to you. They don't want to, but they will. Client wants a video in 3 weeks. Agency says yes. Week 1 feels good — kickoff call, excited energy, timeline looks clean. Week 2, things get quiet. You follow up. "All on track!" Week 3, silence. You follow up again. The editor got slammed. The scriptwriter had a family thing. They need a few more days. A few more days becomes a week. Then two. Then something ships that isn't what anyone wanted. You've either lived this or you will. They're not incompetent. That's what's frustrating. The agency has talent. They've shipped good work before. The portfolio is real. They genuinely believed they could deliver in 3 weeks when they said yes. So what keeps happening? The freelancers are working with 4 other clients. Their incentives are split. Nobody owns the outcome — everyone's shepherding their piece. The agency isn't really producing. It's coordinating. And coordination without ownership falls apart the moment anything goes wrong. Something always goes wrong. Fast. Good. Cheap. Pick two. Everyone knows this. But agencies can't say it out loud because the agency down the street will promise all three. So the whole industry operates on a lie that everyone knows is a lie but nobody can stop telling. The structure forces the lie. Not creativity — creativity exists everywhere. You're buying coordination plus institutional knowledge. Someone who can get multiple people moving together and who's seen enough projects to know what breaks before it breaks. Most agencies are bad at the first and don't have the second. I run a small operation. Boutique — deliberately. Only work with people who treat this project like it's theirs — on the client and freelancer side. That means I don't pretend we're fast AND cheap. But we deliver. On time. At quality. This does not mean failures don't exist, but less than your massive content mill. The trifecta is real. The trick isn't breaking it — it's being honest about which two you're choosing. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/the-tuesday-report [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 14 NOV 2024 · 2 MIN READ # The Tuesday report A November 2024 account of the weekly delays report at Ankur Warikoo. The 348 tasks and 20 delays describe the week discussed in the original post. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/posts/rishwajeet-singh_every-tuesday-i-send-our-team-at-ankur-warikoo-activity-7262763794358018048-5vS3) Every Tuesday, I send our team at Ankur Warikoo a report that would make most managers uncomfortable: a transparent list of everyone’s delays, visible to all. Most companies treat delays like secrets — hidden in private messages, hushed in meetings, or smoothed over in reports. We put them on full display. It sounds harsh until you see the results: Last week, our team handled 348 tasks. Total delays? Just 20. The surprising truth is that exposing delays creates excellence — not through pressure, but through systems. Here’s what we learned from building our transparency machine: ➤ Most Teams Solve the Wrong Problem When tasks start slipping, companies often react with more meetings, check-ins, and “urgent” follow-ups. They try to solve an information problem with attention. It doesn’t work. ➤ Transparency Over Authority Every task lives in Asana. Every delay is tracked. Every Tuesday, everyone sees everything. There’s no need for constant updates or reminders — the system handles it. ➤ Excellence Lives in the Margins We build buffers and failsafes into every process. A two-day delay doesn’t derail our content engine because our systems are designed to absorb minor setbacks. When you expect imperfection, you create space for consistency. ➤ Data, Not Drama Delays aren’t failures — they’re data points. If someone consistently delays certain tasks, it’s a sign to refine processes or provide training. We use transparency as a tool for optimization, not blame. This level of transparency scares most leaders. They worry it’ll hurt morale or foster blame. The reality? It builds ownership. When everyone can see the machine working — or where it’s slipping — they become invested in making it run better. Most content operations run on chaos and hope. We built ours on systems and transparency. The result? A team that doesn’t need constant management, because they can see exactly what needs to be done. Want to know the real secret to scaling content? Stop hiding problems. Build a machine that makes them impossible to ignore — and watch them start solving themselves. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/show-format-development [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Show format development # Can this idea make six different episodes? Use Money Matters to study how a show finds its next story, then work through an invented series: its format, six premises and a pilot plan. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Start with the repeatable question. A series about small businesses could go in hundreds of directions. To commission episode two, the producer needs to know what kind of situation makes a story for this show. Write that down, then try six premises. If they all end with the same advice, you may have one film. ## Connect the format to its next story. Money Matters makes the financial lives of real people the material of the show. The published case describes the requirement behind a promising episode: a decision, a tension and a story the audience can recognise in its own life. That is a useful way to examine a format. “Personal finance” names the subject. A person facing a consequential financial decision supplies a situation a conversation can explore. The individual, numbers and circumstances can change while the programme remains recognisable. Rishwajeet worked across strategy, production and distribution for the first 58 episodes during his Warikoo engagement. The guest-selection system belonged to that wider operation: the format needed a continuing supply of people whose circumstances could sustain its promise. When developing your own series, put the format and the source of future stories beside each other. What makes someone or something suitable for an episode? Can you find more examples with different uncertainties? What access and preparation does each require? A recurring title is easy to write; a series needs a dependable way to find and make the next worthwhile story. [Read the public account of Money Matters. ](https://machinehouse.media/work/ankur-warikoo) The format bible below is a separate invented exercise; it is not the Money Matters production document. ## A compact format bible: The Extra Step. Original proposed teaching series. The scenarios below are invented development examples. No access, footage, client result or released programme is being claimed. | Decision | Proposed format | Audience | Small-business operators and people building tools for them. They want to understand why an apparently inefficient habit persists. | Promise | Each film finds the consequential decision hidden inside one everyday workaround. | Medium | A 6–8 minute authored observational film. This is a practical pilot constraint, open to revision. | Engine | Begin with an object and action; investigate competing explanations; follow the decision; try a small change; show what remains unresolved. | Host | An interested observer who asks for a demonstration and challenges their first explanation. The operator retains room to disagree. | Access | One actual process, its operator and someone affected by the decision. Clear permission for recordings and documents; use labelled reconstructions where needed. | Production boundary | One process per film, readable evidence, intelligible sound and a manageable location. No generic montage standing in for the action. The bible is a shared decision document. It helps a researcher reject an attractive story that cannot show its central decision, and helps an editor recognise when a scene is only atmosphere. The [BBC Maestro series-bible material ](https://assets.cdn.bbcmaestro.com/fb6bcb97c856a4db3429dc092eed5b05.pdf?version=1623169272) is a television reference; this smaller creator-format sheet is our adaptation. ## Six episodes that ask different questions. | Premise | Decision and evidence | The proof nobody can release | Two “final” PDFs and a green tick still lead to a phone call. Follow who can authorise printing which version. | The appointment with no spare part | A bicycle repair calendar has space, but the promised repair needs a part. Follow the difference between booking capacity and fulfilment. | The file called final-final | A designer fixes an export after approval. Trace who can replace the deliverable and how the recipient knows which file to use. | The handover nobody owns | A support case has detailed notes but no next owner. Follow it across a shift without exposing customer information. | The sample that changed in the room | A renovation finish looked different in a photograph. Compare the physical sample, lighting and who agreed to the final choice. | The binder that trained yesterday’s job | A process changed but its training instructions did not. Follow who may update the procedure and what the next trainee actually learns. These premises share a lens, but the uncertainty changes: authority, inventory, version control, ownership, representation and knowledge maintenance. Before commissioning them, confirm actual access. An elegant premise without observable evidence is still a research lead. ## Produce the pilot to resolve uncertainty. For the print-proof pilot, the uncertainty is whether a seemingly small action can sustain an intelligible film. Begin with the operator preparing the machine while a proof waits on screen. Do not reveal the answer in an abstract introduction about efficiency. Let the discrepancy create the question. The proposed progression is object and action, operator’s explanation, a competing explanation, the release decision, a modest test and the remaining limit. The [editable format bible ](https://machinehouse.media/resources/show-format/format-bible.txt) contains a beat map. Its times are planning estimates, not references to footage that exists. Before filming, secure permission to observe a real job and redact customer details. Ask what happens when the approver is absent or a new file arrives. Those exceptions test whether the initial explanation holds. Record what happens; if the planned “reveal” does not occur, revise the film rather than stage evidence for it. Review a rough cut with an unfamiliar viewer and an operator. Ask the viewer to describe the decision in their own words. Ask the operator which explanation is wrong or incomplete. Log confusing moments and missing evidence, along with the actual access, edit and review effort. Choose to commission, revise and retest, or stop. Set any numerical threshold for this project before the review; there is no universal pilot pass rate. ## Change the medium, change the work. An interview-led version would need the operator to reconstruct the scene aloud: what was on the screen, who called whom, what could go wrong. The host would challenge the interpretation and ask for exceptions. Audio cannot depend on a PDF the listener cannot see. An authored film can inspect the object and sequence directly, but needs location access, clearance and visual explanation. A short social series might make one decision legible in under a minute; it would surrender much of the investigation. Choose the form your evidence and operating capacity can support. After the pilot, update the format bible where reality changed it: recurring beats, host role, access burden, feasible length and sources of future stories. Keep those decisions visible before commissioning the next episode. [Download the completed example and blank format worksheet ](https://machinehouse.media/resources/show-format/format-bible.txt). ## Questions about the method. How many episodes should a new format plan include? + Enough distinct premises to expose repetition. Six is a useful exercise in this example, not a requirement for every series. Each premise needs a different uncertainty and plausible access to evidence. Is format development only for podcasts? + No. It applies to authored films, interviews, documentaries, social series and other repeatable editorial forms. The medium changes the host, evidence and production requirements. ## Keep exploring. [Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Content production ↗ ](https://machinehouse.media/services/content-production)[Podcast production ↗ ](https://machinehouse.media/services/podcast-production)[Build the editorial slate ↗ ](https://machinehouse.media/guides/editorial-slate) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/podcast-trailer-editing [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Podcast trailer editing # Give the listener a reason to stay. Listen to two trailers cut from the same teaching conversation. Compare the openings, check every cut against the full recording and see why three clips were left out. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Name the job before choosing clips. A show trailer introduces the recurring promise of the programme. An episode trailer gives someone a reason to choose this conversation. A cold open sits before the episode’s main opening and should hand the listener into what follows. They can share material, but they do not have the same job. [Spotify’s show-trailer guidance ](https://creators.spotify.com/resources/create/why-your-podcast-needs-a-trailer) concerns that first, show-level invitation. For an episode trailer, write what a new listener needs to understand and the answer they should want next. Utility offers a useful decision; identity offers recognition; stance offers an argument worth examining. Choose the strongest reason in the actual conversation. A succession of dramatic statements does not create progression by itself. ## The conversation behind these edits. Original fictional teaching example with synthetic voices. A host interviews an invented software founder about print-proof approval. The shop, people and pilot are fictional. Daniel and Samantha, the macOS voices, read the original script; neither represents a real guest or Rishwajeet. These are audio teaching edits, not finished video trailers. ### What the source actually says D02, Founder: “There were two PDFs in the chat, both called final. Someone replied, looks good. The operator had the machine ready, but he was still making a phone call.” D04, Founder: “The green tick told us someone had clicked. It did not tell us whether they could release the job.” D06, Founder: “If one person owns all three, the process should stay simple.” The full source goes on to explain a version-specific release decision, an absent approver and the limits of a small pilot. Read those answers before selecting the lines above. The claim is about clarifying responsibility; it is not evidence of time saved or a software product’s performance. [Read the complete timecoded transcript ](https://machinehouse.media/resources/podcast-trailer/source-transcript.txt) or [download the source audio ](https://machinehouse.media/resources/podcast-trailer/conversation-source.wav). ## Two different ways into the episode. Scene first · 00:58.849. Begin at the print shop. The action establishes a puzzle, the green-tick distinction explains it, and the host’s objection prevents the argument from becoming “everything needs more process.” Finish by asking what changed. The full episode supplies that answer. Scene-first episode trailer · synthetic demonstration Question first · 00:54.418. Open with the founder’s retrospective question, “Who can say yes, and yes to what?” Put the host’s objection next, then explain the green tick. This version leads with a useful challenge for someone already dealing with approvals. It gives up the print-shop scene’s immediate visual specificity. Question-first episode trailer · synthetic demonstration The second edit rearranges complete turns. D05 originally followed D04; here it challenges D14’s approval question. Check that relationship explicitly: the objection still addresses unnecessary approval complexity, and its qualifying answer remains attached. Neither version changes a speaker’s words or manufactures an outcome. Prefer the scene version for a new audience; consider the question version where the audience already recognises the problem. That is an editorial hypothesis to test, not a predicted winner. Using the scene edit as a cold open: its final question, D07, can hand directly into the full D08 answer after a brief show identification, rather than replaying the conversation from D01. The source transcript supplies that answer. Check the transition in the complete episode timeline and remove any duplicated question or introduction. These supplied MP3s stop at the question; the finished episode assembly is not part of the download. ## The source-to-edit sheet. | Source clip and range | Scene edit range | Purpose | D02 · 00:04.148–00:22.726 | 00:00.000–00:18.578 | Show the puzzling action | D03 · 00:23.226–00:25.330 | 00:19.078–00:21.182 | Ask the immediate question | D04 · 00:25.830–00:38.114 | 00:21.682–00:33.966 | State the distinction | D05 · 00:38.614–00:44.548 | 00:34.466–00:40.401 | Test an objection | D06 · 00:45.048–01:00.070 | 00:40.901–00:55.922 | Keep the qualification | D07 · 01:00.570–01:02.997 | 00:56.422–00:58.849 | Open the next answer Source timings locate the conversation recording; edit timings locate the assembled trailer. They are different clocks. Both versions use complete spoken turns with short pauses between them. The [editable JSON selection sheet ](https://machinehouse.media/resources/podcast-trailer/selection-sheet.json) contains both cut orders and the exact ranges. The [text version ](https://machinehouse.media/resources/podcast-trailer/source-transcript.txt) combines transcript and timings. ## Three good clips to leave out. | Clip | Why it is tempting | Why it is absent | D08: the version and release-owner fix | A complete useful answer. | It resolves the question the trailer is inviting the episode to answer. Keep it in the episode, or make a separate standalone advice piece. | D10: silence cannot mean approval | A concise, forceful principle. | Without the absent-approver question and agreed-substitute context, it introduces another branch of the story. | D12: the pilot proves little | An honest qualification. | The chosen trailer makes no performance claim that needs this specific correction. Preserve it in the full episode; include it if a future edit implies proven operational improvement. A rejected clip is not necessarily weak. It may belong in a different asset. Keep a reject log so the next editor knows whether the reason was meaning, repetition, length or a deliberate decision about the unanswered question. ## Finish the argument before decorating it. Build intelligible dialogue and transitions first. Add picture that helps identify speakers, objects and consequences. For this proposed film treatment, a clearly labelled reconstruction could show the two proof versions; generic laptop footage would add little. No such footage is supplied with this audio example. Use music only when it serves the intended feeling without making a measured answer sound like an accusation. Leave enough pause for the objection and qualification to register. Prepare accurate captions and a transcript for the actual destination. Do not cut the negation, narrow the subject or join words into a sentence the guest never said. Review the cut against the complete recording, then ask an unfamiliar person what they believe the guest claimed and what answer they expect next. Repair a misleading promise even if the isolated opening sounds stronger. [Download the blank selection and review sheet ](https://machinehouse.media/resources/podcast-trailer/blank-selection-sheet.txt). ## Questions about the method. Should every podcast trailer be 45–90 seconds? + No. These examples use that range to demonstrate a complete short progression. The destination, audience and material should determine the length. Can I make a trailer from a transcript alone? + You can identify candidate passages, but the finished cut needs checking against the recording for exact words, tone, speaker identity, pauses and surrounding meaning. ## Keep exploring. [Brief a video editor ↗ ](https://machinehouse.media/guides/video-editor-brief)[Podcast production ↗ ](https://machinehouse.media/services/podcast-production)[Guest research ↗ ](https://machinehouse.media/services/podcast-guest-research) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-approval-workflow [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Content approval workflow # Find out where the review got stuck. Use a weekly delays report, an approval map and three worked review loops to decide who gives feedback, who resolves disagreements and what happens after a late change. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 27, 2026 ## Find out where the time is going. A script can sit with a reviewer for a day and receive ten minutes of attention. Log when it was ready, who needed to respond and when the next instruction arrived. That tells you whether to change the writing process, the review deadline or who is allowed to decide. Separate elapsed time from labour. Two people working at once do not create twice the elapsed duration. Record a few comparable pieces before setting targets; one unusually smooth release is not proof of a better system. ## What a weekly delays report makes visible. In November 2024, Rishwajeet described a weekly report used in Ankur Warikoo’s content operation. Every Tuesday, the team could see the delayed tasks. Work lived in Asana, and the operation used buffers to absorb some of the slippage. The week in the account contained 348 tasks and 20 delays. The useful detail is the visibility. A delayed task became something the team could examine: a recurring difficulty, a process that needed changing, or a place where someone needed support. The report gave that conversation a shared starting point. For an approval workflow, take the same approach one step further. Record when the draft became ready, who had the next decision, when they responded and whether that response reopened something already settled. That lets you distinguish a writer missing the handoff from a finished draft waiting for review. Those two delays call for different repairs. The published account does not contain a before-and-after approval experiment. Its 348 tasks and 20 delays describe one reported week, not a measured reduction caused by the report. Use the example to design a visible process, then measure what happens in your own operation. [Read the original Tuesday report. ](https://machinehouse.media/writing/the-tuesday-report) ## Give each stage a decision and an owner. | Stage | Decision | Owner and advisers | Premise / outline | Is this the audience problem and argument we want to publish? | Founder or commissioning lead decides; editor tests the story and audience fit. | Script / evidence | Can we stand behind these words, examples and claims? | Editorial lead resolves the draft; subject expert checks accuracy. Escalate material disagreement to the commissioning owner. | Rough cut | Does the sequence explain the intended story? | Editorial lead consolidates comments from the producer and relevant stakeholders. | Finish / release | Is this exact version ready for the agreed destination? | Named release owner checks completion of required corrections, captions, credits and delivery requirements. One person may hold several roles. A short founder post may only need a draft review and release decision. A complex film may need more stages. [Frame.io’s production approval guidance ](https://blog.frame.io/2022/04/18/10-steps-for-client-approval/) provides a fuller film workflow; use only the stages your work requires. Advisers can flag risks and suggest alternatives. The feedback resolver decides what instruction reaches the maker, or escalates a decision outside their authority. This prevents the editor being left to infer which stakeholder outranks another. ## Three review loops, with the conflict resolved. Simulated teaching timeline. The following launch-video example is invented. Its times illustrate categories of delay; they do not demonstrate time saved. ### Loop one: the premise Monday 10:00, script v01: marketing says “Open with our three features.” The founder says “Open with the customer problem.” Both comments reach the writer without a decision. Tuesday 10:00, the editorial lead resolves the conflict: open with the print operator’s unresolved release decision; show one relevant feature after the audience understands it. The founder approves that premise. The 24-hour interval is elapsed queue and review time; writing labour has not been measured. ### Loop two: the rough cut Wednesday 12:00, rough cut v02: sales wants the dashboard at 00:08. The producer says the first twenty seconds never show what the operator is waiting for. Wednesday 15:00, the editorial lead gives the editor one instruction: show the two proof versions at 00:08, label the scene as a reconstruction, then introduce the dashboard once the release decision is clear. The approved premise stays. The replacement cut still needs review; this decision records what to change in v02. Three hours have elapsed between the review request and that instruction; editing labour remains unmeasured. ### Loop three: the late claim Thursday 16:00, finish v03: the founder requests a new customer result that is absent from the source. Adding it would require evidence and renewed script approval. Friday 10:00, the release owner chooses to publish v03 without the unsupported result and investigate a later piece. The approved release record names v03; a future claim will need its own review. Eighteen elapsed hours are visible, but there is no baseline from which to claim improvement. ## Translate reactions into usable feedback. “This feels flat” tells the maker that something is wrong, but leaves the diagnosis open. Try: “At 00:08 I still cannot tell what the operator is waiting for. Can we show the two proof versions before introducing the dashboard?” That identifies the moment, the viewer’s problem and a possible repair. Separate observation from prescription. “I lost track of who can approve the print run” is an observation. “Add a red arrow” is a proposed solution. The editor may have a clearer way to restore meaning. Ask the resolver to preserve the underlying problem when consolidating comments, including useful objections from people whose proposed fixes were rejected. Use a review request that states the decision, version, deadline and open questions. “Please review” can invite a premise debate during colour finishing. “Review v03 for completion of the four approved corrections; a new premise changes the release schedule” is more precise. ## Record the version and the tradeoff. An approval record should identify the file or link, version, decision owner, time and conditions. “Approved with corrections” needs the correction list and someone responsible for checking it. A replacement upload must be distinguishable from the approved file. For late changes, name the reason, affected decision, required work and options. An urgent factual correction may justify moving the release. A new preference may wait for the next piece. The owner chooses with the cost and deadline visible; the editor should not silently absorb a different assignment. When a reviewer is unavailable, use an agreed substitute or move the decision. A deadline does not convert silence into consent. For genuinely urgent work, settle the substitute’s authority in advance and reduce the number of optional deliverables. [Download the editable review workflow, time log and simulated example ](https://machinehouse.media/resources/content-approvals/review-workflow.txt). ## Questions about the method. Does every content team need a five-stage approval process? + No. Combine stages for small work and separate them where a decision is expensive to reopen. The essentials are the decision, exact version, owner and next action. How should contradictory feedback reach the editor? + A named resolver should consolidate it into a coherent instruction, explain the decision and escalate conflicts outside their authority. Do not ask the maker to guess which stakeholder has the final say. Does the Tuesday report prove that approvals became faster? + No. The November 2024 account describes 348 tasks and 20 delays in one week of a broader content operation. It is an example of visible work and process review, not a before-and-after approval experiment. ## Keep exploring. [Write an editor brief ↗ ](https://machinehouse.media/guides/video-editor-brief)[Scope production costs ↗ ](https://machinehouse.media/guides/content-production-costs)[Content production ↗ ](https://machinehouse.media/services/content-production)[A knowledge system for a content team ↗ ](https://machinehouse.media/guides/content-team-knowledge-system)[Content capacity planner ↗ ](https://machinehouse.media/labs/content-capacity-planner) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/founder-thought-leadership [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Founder thought leadership # Find the point of view inside the decision. Read Rishwajeet’s 40-hours essay alongside the decisions behind its argument. Use the interview worksheet to find equally specific material in a founder’s own experience. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Start with an event the founder can reconstruct. Ask the founder to reconstruct a decision: a feature they removed, a customer they turned down or a hire that changed the team. Who was there? What did they think would happen? What did they do? Those details give you material to write with. Ask what they thought at the time, what they noticed and what changed. An interview should leave room for surprise. [StoryCorps’ conversation guidance ](https://storycorps.org/participate/tips-for-a-great-conversation/) is a useful reminder to listen and follow specific stories; the commercial writing process here is our adaptation. Before drafting, separate firsthand observation, interpretation and externally testable claims. A founder can describe one customer conversation without having established what the entire market believes. Keep private details out unless their publication is authorised. ## The event contains the argument. Rishwajeet’s January 2026 essay begins with a decision that was difficult to justify on an hourly-rate calculation: he spent 40 hours directing and editing a video for a strategy client. The source contains more than the number of hours. He could picture the pacing and the cuts, but the team did not yet have a process that could translate that picture into the finished work. A conventional brief would have left too much of the judgment unexplained. He made the video himself, then used the finished piece as something he could teach from. That sequence gives the writer an event, a constraint, a choice and a change in understanding. It is much stronger material than asking the founder for a general opinion about craftsmanship. ## Three possible arguments; one that earns the ending. | Possible argument | What the source supports | Editorial judgment | Founders should personally edit their videos. | Rishwajeet did so on this project. | Too broad. One worthwhile intervention does not establish how every founder should spend their time. | More production effort produces more views. | The essay reports 40 hours of work and 500K views. | Unsupported as a causal claim. The account does not isolate the effect of those hours. | A finished example can teach creative judgment that an abstract brief leaves unclear. | He describes a specific execution gap, makes the piece and then uses it to explain the decisions. | The strongest argument: the lesson follows from what happened, and a reader can apply it to a comparable handoff problem. The important turn is what became possible once the video existed. Keep the interpretation tied to that sequence. Inventing client feedback or promising the same outcome on another project would make the argument less trustworthy. ## Read the finished piece as a craft example. The [original essay ](https://machinehouse.media/writing/forty-hours-below-my-rate) shows the whole progression. Its opening gives us the apparently uneconomic choice. The middle explains what Rishwajeet could see and why a normal handoff would lose it. The final movement connects making the piece to teaching the team. The production details carry the voice: pacing, where a cut lands, how long a shot holds, sitting with the finished film. Replacing those with a phrase such as “a commitment to excellence” would remove the evidence that makes the argument his. When interviewing another founder, use this structure to find their own material. Ask what they could see that the existing process could not express, what they chose to do about it and what the finished work made possible. Follow their answer; the aim is to uncover their decision, not to give everyone Rishwajeet’s story. ## Preserve the person while improving the structure. Keep a small voice record from the authorised source: characteristic words, how the founder qualifies a claim, and whether they naturally explain through scenes, contrasts or technical detail. In the 40-hours essay, the language of cuts, pacing and how long a shot holds makes the creative judgment concrete. A writer can remove false starts, arrange the argument and explain missing context. They should not manufacture firsthand experience, intensify certainty or make a cautious founder sound contemptuous. Read the draft aloud. Ask the founder which sentence they would never say and which detail changes the meaning. An owned essay can develop the counterargument. A LinkedIn post may enter through a sharper incident and close with a useful question. A spoken film needs words that make sense on first hearing and images that do explanatory work. The [distribution guide ](https://machinehouse.media/guides/content-distribution-plan) explains how to choose each treatment’s audience job and destination. ## Review the claim before the commas. Ask whether the experience is accurate, whether the interpretation is theirs, what is missing and where the advice stops applying. Resolve those questions before fine-tuning punctuation. Record approval of the exact version and destination. A founder’s approval of an interview transcript is not automatically approval of every future argument drawn from it. Choose a cadence that leaves enough time for worthwhile conversations and review. Judge early responses by relevance: did the intended reader recognise a decision, offer a useful counterexample or start an appropriate inquiry? Keep those observations separate from claims about sales impact. [Download the interview, voice and review worksheet ](https://machinehouse.media/resources/founder-writing/interview-and-review.txt). For a separate practice task, use the [complete fictional source-to-argument exercise (TXT) ](https://machinehouse.media/resources/founder-writing/fictional-worked-example.txt). It includes three possible theses and a finished illustrative passage, with links to the transcript and synthetic recording. ## Questions about the method. Can a writer help develop a founder’s point of view? + Yes. The writer can identify competing arguments, test their evidence and arrange the piece. The founder must recognise and stand behind the experience and claim attributed to them. Does thought leadership require a daily LinkedIn post? + No. Choose a cadence that supports useful source conversations, careful writing and actual review. Written posts, owned essays, newsletters and spoken films can serve different audience situations. What does the worked example show? + It examines Rishwajeet’s published account of spending 40 hours directing and editing for a strategy client, compares three possible arguments and explains how the source supports the finished essay. A separate fictional exercise and blank worksheet are also available. ## Keep exploring. [Founder content partnership ↗ ](https://machinehouse.media/services/founder-brand-content)[About Rishwajeet ↗ ](https://machinehouse.media/rishwajeet)[Read the essays ↗ ](https://machinehouse.media/writing)[Plan how the argument reaches readers ↗ ](https://machinehouse.media/guides/content-distribution-plan) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-distribution-plan [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Content distribution planning # Plan what happens after the main release. A worked release schedule with three companion pieces, their audiences, destinations and owners. Use it to plan what must be ready before each piece goes out. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Choose the audience and the next useful step. The example below takes one recorded conversation and plans an episode release, a trailer and an original short. Each needs its own audience, destination, owner and deadline. Decide those before asking the editor for three more assets. A full episode, a trailer and an original short can come from the same conversation while making different promises. The episode offers the complete conversation. The trailer creates interest in an answer inside it. A standalone short should deliver one useful idea on its own. Plan and assess each against its own job. The worked plan uses an [original fictional approval conversation ](https://machinehouse.media/resources/podcast-trailer/source-transcript.txt). Its insight is that a manual workaround may carry a decision the software has not understood. It does not establish measured savings or a rule that every manual step should remain. Choose channels from audience behaviour and your capacity to make useful native work; you do not need every channel for every idea. ## Give each release a destination and owner. | Asset | Encounter and audience job | Route onward | Release dependency | Owned essay | A reader wants the fuller explanation and counterexample. | Relevant workflow resource or enquiry if the problem fits. | Web editor checks the actual page, labels, links and readable resource. | LinkedIn post | A feed reader recognises the decision problem. | The essay where a fuller explanation would help. | Founder reviews the exact post and working destination; the fictional label stays visible. | Original short | A viewer understands the difference between comment and release authority through a visual example. | The essay through the channel’s supported route. | Producer completes the film, captions, asset checks and destination setup. | Optional newsletter | An existing subscriber receives a useful explanation or follow-up. | The essay or a reply to a specific question. | Mailing permission, reviewed email and tested links; assign an actual send date. These are planned routes for a teaching example. They do not represent posts sent, audiences reached or enquiries generated. Use the relative schedule below only after assigning people and checking the dependencies. ### Set the sequence and the review window. | When and owner | Release and audience job | Observe | T0 · web editor | Publish the owned essay for readers seeking the fuller explanation; verify the actual destination. | Relevant reading and resource use within the available analytics; record what is not observable. | T+1 · founder | Publish the LinkedIn treatment for feed discovery, linking to the essay where appropriate. | Substantive replies from the intended readers and tagged visits over a defined seven-day window. | T+3 · video producer | Publish the native short on the chosen channel for visual explanation; use the supported path to the essay. | Engaged viewing and relevant follow-through, using that platform’s definitions. | T+8 · editor | Review what each treatment did and decide the next change. | Sample sizes, useful objections, source of inquiries and unresolved attribution. These are relative planning dates, not scheduled posts. Replace role names with people and dates before release. Owned, organic, partner and paid distribution have different dependencies. A partner share requires their agreement; paid distribution requires its own authorised budget and measurement plan. Neither is assumed here. For a podcast release, make the full episode the primary destination. Lock its title, description and playable page before publishing the episode trailer. Select a clip that resolves one useful question, link it to the full conversation through the destination’s supported route, and give the show notes the source or worksheet discussed on air. Assign an episode publisher, trailer editor and response owner explicitly. A show-level trailer introduces the programme and need not be remade for every guest. For the fictional approval conversation, the scene-first trailer invites the listener into D08’s answer about versions and release authority. A separate standalone short can explain that answer completely, as the complete companion script does. Those two assets have different promises: wanting the episode’s next answer and understanding one idea immediately. Review the first for relevant episode starts and the second for comprehension and appropriate follow-through; do not add incompatible platform counts into a claimed unique audience. ## Adapt the encounter, then make the asset. The same source can support different encounters. The excerpts below show how the treatment changes; all refer to the fictional print-shop example. [Read the complete essay, LinkedIn post and visual/script treatment (TXT) ](https://machinehouse.media/resources/content-distribution/complete-treatments.txt). ### Owned essay: explain the decision. The essay develops the responsibility problem and its counterexample for someone willing to read the argument. Its opening gives the reader a question to investigate: A useful workaround looks inefficient from the outside. Consider an invented print-shop scenario. An operator receives a customer’s “looks good” message, then calls for confirmation before printing. The complete treatment follows the two files and the release decision, then asks what a small next test could establish. ### LinkedIn: make the problem recognisable. The proposed post compresses the scene for a feed reader. It has not been posted: Two files called final. A green tick. And an operator still making a phone call before printing. In this fictional product-design example, the call looks like an unnecessary step. The rest of the post explains what the call settles and points interested readers towards the fuller argument. An optional newsletter would address an existing subscriber, explain why the subject matters now and offer a specific reply question. Use the actual mailing permissions, review the email and test its links. ### Short video: let the object explain it. The proposed film keeps two recreated files visible while separating wording, version and permission to release. The spoken explanation begins: Two files called final. Someone says, looks good. Why is the print operator still making a phone call? The complete visual/script treatment answers that question and ends with a useful decision the viewer can apply. No film has been produced. Keep the fictional label in the film itself, use readable original mockups and accurate captions, and clear the audio. This is an original standalone script; a clip taken from an interview needs its own source and cut check. ## Track the path without inventing a result. For external campaign links, use a consistent naming convention for source, medium and campaign, then verify that the destination retains the parameters. [Google Analytics documents manual campaign tagging ](https://support.google.com/analytics/answer/10917952?hl=en), including case-sensitive values. Use additional fields where relevant to your reporting. Keep acquisition tags off internal navigation. A tagged visit shows a recorded path. It does not prove that the post caused a later sale, and an untagged inquiry may still have encountered the work. Ask how people found you, keep the answer alongside the analytics and report uncertainty. Compare each asset to its own purpose before combining numbers with incompatible definitions. The next action should follow the evidence: clarify an explanation readers misunderstood, develop a useful counterargument or stop making a treatment that has no clear job. [Download the release plan and measurement worksheet ](https://machinehouse.media/resources/content-distribution/release-plan.txt). ## Questions about the method. How many assets should come from one long-form piece? + Choose the number from distinct audience jobs and production capacity. An asset should offer a useful encounter and an appropriate destination, not exist only to increase the output count. Is distribution the same as repurposing? + Repurposing creates another treatment from source material. Distribution decides who should encounter it, where, when, through which route and how its contribution will be assessed. Where are the complete content treatments? + The guide links to an editable text companion containing the full illustrative essay, LinkedIn post and visual/script treatment. They use a fictional source and are not evidence of a published campaign. ## Keep exploring. [Long-form to original shorts ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Founder content ↗ ](https://machinehouse.media/services/founder-brand-content)[Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Follow the inquiry into the business ↗ ](https://machinehouse.media/guides/content-business-impact) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/ai-content-workflows [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ AI content workflows # Choose the step worth automating. Examine Money Matters’ guest-selection process, then compare manual work, fixed automation and a model step. Try the evaluation exercise with an answer that sounds convincing but outruns its sources. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Start with the work, not the agent. Pick a task your team already does: collect sources, prepare a guest brief or check references in a script. Give the proposed automation the same inputs and compare its output with work your team would accept. Record where someone still has to check, correct or finish it. The manual baseline is not an obstacle to automation. It reveals which decisions experienced people make without writing them down. If nobody can explain why a brief is accepted, a model’s fluent output can hide disagreement rather than resolve it. Measure a few real examples before making a time-saving claim. ## Choose among manual, fixed and agentic work. | Approach | Good candidate | Boundary | Manual editorial work | Choosing a premise, assessing a sensitive claim, resolving stakeholder disagreement. | Human work still needs evidence and a recorded decision. Familiarity is not a substitute for review. | Fixed automation | Rename known exports, validate required fields, generate links from a manifest, render an approved page. | Useful when the steps and rules are clear. Unexpected inputs should fail visibly rather than be guessed through. | Bounded model step | Draft an evidence-linked summary or propose alternative arguments from supplied material. | Treat the text as a candidate. Check factual meaning, missing qualifications and attribution. | Agentic workflow | A genuinely variable research task where the next permitted step depends on what has been found. | Needs limited tools, a stop rule, cost bounds and review of both outputs and actions. More autonomy increases the evaluation burden. [Anthropic’s engineering discussion ](https://www.anthropic.com/engineering/building-effective-agents) distinguishes predefined workflows from agents that dynamically direct their process. The distinction is useful even as products change. Begin with a simpler workflow when it can perform the task reliably; do not introduce dynamic tool use just to rename a sequence of fixed steps. ## Start with an actual selection bottleneck. Money Matters had a concrete selection problem: thousands of applications, each requiring enough attention to judge whether it could become a useful episode. During his work with Ankur Warikoo, Rishwajeet built a workflow with two stages. The first checked demographic and structural eligibility. The second scored editorial potential using criteria informed by earlier episodes. A strategist reviewed the shortlist and made the final selection. The separation matters. A submission can fit the show’s requirements without containing a strong conversation. Combining eligibility and story judgment into one unexplained score would make it harder to see why a candidate was rejected or what a reviewer should examine. | Stage | Job in the published case | Question to answer when adapting the method | Eligibility | Check demographic and structural fit. | Are the criteria explicit, and what happens to an incomplete or ambiguous application? | Editorial scoring | Assess potential against patterns from earlier episodes. | Can the reviewer see the story evidence behind the assessment? | Strategist decision | Choose from the shortlist. | Which judgment remains with the person responsible for the programme? The public case reports that the initial workflow took four days to build and reduced weekly review from days to under 60 minutes. Those are historical figures from that operation, not a benchmark for another team. Before claiming an improvement in your own workflow, measure the complete accepted task, including review and correction. The case is evidence for a useful division of work. It does not require calling every step an agent. Start by identifying the bottleneck, separating decisions that need different treatment and keeping the consequential editorial choice with its owner. [Read the Money Matters case. ](https://machinehouse.media/work/ankur-warikoo) ## A complete research-to-brief workflow. Suppose the team needs an editor brief from an approved script and source collection. The final deliverable must let an editor find the right material and understand the intended sequence. - Intake, human: agree the audience, purpose, source permissions and who approves the final brief. Confirm that selected files exist and are accessible. - Manifest, fixed rules: assign stable source IDs and check required fields. A missing file enters an exception list; it is not replaced by a guessed reference. - Draft, bounded model: propose an outline and source-linked selects using only the supplied material. Mark uncertain claims and missing evidence. Do not invent timecodes. - Checks, mixed: code can validate fields and reference IDs. An editor must assess whether a quote supports the claim, whether the sequence is faithful and whether the suggested visual exists. - Approval, human: resolve exceptions and accept an exact version. Changes to meaning reopen the relevant decision. - Delivery, authorised: generate the approved handoff and verify access at the intended destination. External sending or publication requires its own authorisation. Keep failures recoverable. Store the source version, draft and review decision so a corrected input can be rerun without silently replacing the accepted brief. Give the exception queue an owner. A workflow that produces a draft but leaves inaccessible sources for the editor has moved work rather than completed it. At the manual level, the producer builds the manifest and an editor writes and checks the brief. At the fixed-workflow level, file checks and document assembly run automatically while the same editorial decisions remain explicit. Optional agentic research would only enter when an approved evidence gap requires choosing among permitted sources; it returns a source packet for review before the brief changes. It does not acquire authority to invent missing clips or publish the result. ## Catch an output that sounds more complete than the evidence. Original teaching fixture. The memo and candidate below are invented for evaluation, not output from a measured production system. ### Source memo Tuesday: two producers tried the draft brief template. One could locate the selected clips. One could not open the source folder. No editing-time measurements were collected. Before another trial, the producer must repair access and the editorial lead must approve the claim wording. ### Faulty candidate “The new AI workflow cut editing time by 40% across the team. Both producers completed the pilot successfully. Automatically send the new brief to the client.” | Problem | Required response | Invented 40% improvement | Reject. The source contains no editing-time measurements. | Two successful producers | Correct. One producer could not open the source folder. | Automatic client delivery | Block. Access repair and claim approval remain open; external sending is not authorised. Acceptable rewrite: Two producers tried the draft brief. One found the selected clips; the other could not open the source folder. Repair access before another trial. No editing-time improvement has been measured. The editorial lead still needs to approve claim wording. A second model could help flag discrepancies, but agreement between models is not evidence. The reviewer should be able to trace every material statement to the memo. [AP’s published generative-AI guidance ](https://www.ap.org/the-definitive-source/behind-the-news/standards-around-generative-ai/) treats output as unvetted source material. That is a useful verification principle; this guide is not a claim of AP endorsement or a reproduction of its newsroom rules. ## Evaluate failures you would actually care about. Build a small set of representative fixtures: a clean source, a missing file, contradictory accounts, an unsupported number, a changed version and a request outside the system’s authority. Record the expected decision before running the workflow. Keep some examples separate from prompt development so the evaluation can reveal unfamiliar failures. Score concrete defects: unsupported claims, lost qualifications, incorrect source links and unauthorised actions. Also record human correction time, latency and cost. A schema-valid draft can still be substantively wrong. A polished average can hide a rare failure that makes the workflow unsuitable for its intended use. Set release criteria from the task’s consequence and your actual baseline. Recheck affected fixtures when the prompt, model, tools or source structure changes. Do not convert one successful demonstration into a claim that the operation is autonomous. ## Build, buy or keep the step manual. Buy when an existing tool fits the task, permission model, exports and review flow closely enough to reduce maintenance. Build when a specific repeated gap justifies owning integrations, monitoring and recovery. Keep the work manual when volume is low or the judgment cannot yet be evaluated sensibly. Compare total work: setup, source preparation, review, corrections, subscriptions, integration upkeep and recovery from failure. Test with representative material before committing. The useful question is whether the complete accepted task gets easier at an acceptable error rate, not whether the first draft appears instantly. [Download the workflow, faulty-output fixture and evaluation fields ](https://machinehouse.media/resources/ai-workflows/workflow-and-evaluation.json). ## Questions about the method. When does a content operation need an AI agent? + When a useful task requires dynamically choosing among permitted next steps and that added autonomy can be evaluated and controlled. Many content tasks work well with fixed automation and a bounded model step. Can AI check its own factual accuracy? + It can propose checks and flag inconsistencies, but its agreement with itself or another model does not establish truth. Material claims still need verification against appropriate evidence. ## Keep exploring. [AI content systems ↗ ](https://machinehouse.media/services/ai-content-systems)[Content approval workflow ↗ ](https://machinehouse.media/guides/content-approval-workflow)[Research to script ↗ ](https://machinehouse.media/guides/research-to-script)[Evaluating AI content systems ↗ ](https://machinehouse.media/guides/evaluating-ai-content-systems)[A knowledge system for a content team ↗ ](https://machinehouse.media/guides/content-team-knowledge-system) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/podcast-guest-selection [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Podcast guest selection # Who can answer this episode’s question? Masoom’s guest-strategy example and a three-candidate exercise show how the audience goal changes a shortlist. The worksheet records the evidence, access and reason for each choice. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Choose the episode question first. For an episode about a difficult decision, look for someone who made it and can explain what happened. The shortlist should tell the host which experience each candidate brings, what evidence you have and whether you can get them into the room. Write the premise before researching names: who is watching, what they already believe, what the conversation should help them understand and why this person is needed. If the premise changes during research, update it openly. Do not keep the original title merely because it made the first invitation easier. ## Let the audience goal change the shortlist. The Masoom Minawala Show provided a useful challenge to a familiar booking assumption: a more famous guest will necessarily do more to build the show. In a December 2025 account, Rishwajeet described tagging episodes by guest type, topic, mood and fame, then comparing the response. The Shloka Ambani episode had about 300K views in that account. The Mitesh Rajani episode had about 200K, with more subscribers gained and more comments reported despite its smaller view total. Those observations change the booking discussion. If the next episode needs to introduce the programme to a wider audience, reach matters. If it needs to give the intended viewer a useful answer and a reason to return, access to a practitioner’s specific knowledge may be more valuable. The show’s purpose—serving what its audience of women cares about—helps decide which job comes first. Write that job into the shortlist. For each candidate, state the question they can answer, the experience that lets them answer it and the audience response the episode is meant to encourage. A recognisable name still needs an episode promise. A specialist still needs a conversation that the audience wants. This is a historical comparison reported by the strategist, not a controlled test of fame. The published essay does not provide exact subscriber conversion rates, and the guests, topics and packages differed. The useful lesson is to examine more than reach before choosing the next guest. [Read the original guest-strategy account. ](https://machinehouse.media/writing/celebrity-vs-expert) ## Three candidate roles, one specific decision. Original proposed teaching episode. The question is: “What does the red patch on a city heat map actually tell us?” The candidates below are hypothetical roles, not researched real people or booked guests. The supporting public-source brief distinguishes the measured surface from the air temperature in a weather report. | Candidate role | What they could contribute | What must be checked | Urban-temperature researcher | Explain measurement methods, uncertainty and what a particular map can support. | Direct relevant work, a primary publication or project, communication sample and access to an appropriate example. | Landscape practitioner | Describe installation, maintenance and practical decisions around trees and shade. | Firsthand responsibility and documented work. Do not assume expertise in remote-sensing methods. | Well-known general commentator | Offer a familiar voice and accessible framing for a wider audience. | Whether they can substantiate this specific explanation; prominence alone does not answer the question. Research the measurement-specialist role first. That is a selection of expertise to investigate, not permission to invent a biography or send an invitation. If the premise becomes “Who keeps a street tree alive?”, the practitioner may become the stronger lead. The guest choice follows the episode’s job. ## Check the evidence behind each candidate. For each real candidate, preserve an unambiguous identity link, relevant firsthand work, the source that establishes it and a communication sample. A similarly named person or a general job title is insufficient. Read or watch enough of the actual work to understand what the candidate did, rather than treating a search snippet as a credential. Separate verified facts from promising questions. “Published a study using this method” needs the study and contribution. “Might explain why this map misleads people” is a hypothesis to test in a pre-interview. The shortlist should make that difference obvious to the person approving it. Look for something the conversation can reveal beyond a biography already available online: a consequential decision, a disagreement within the field, a surprising exception or a process the guest can demonstrate. Keep relevant interests visible when they could shape the explanation. A commercial relationship does not automatically disqualify a guest, but it can change the questions and disclosure needed. ## Test the conversation, not a personality score. A pre-interview can test whether the guest can reconstruct a specific moment, explain a difficult distinction and acknowledge a limit. Ask one real question from the intended episode. Listen for the answer and the follow-up it creates. Do not judge only whether the guest is polished; some people become much clearer when shown an object or asked about work they performed. For the heat-map premise, ask the candidate to identify the variable, unit and time on one suitable example, then explain a conclusion it would not support. A fluent answer without a source does not close the research gap. A careful answer that corrects the premise may improve the episode. Consider host fit and production requirements together. A guest may need an interpreter, a remote recording, a visual aid or time to clear an example. Those are planning inputs, not reasons to make undocumented assumptions about their ability. ## Record why this guest, and why now. Use a short decision record: invite, research further, hold for another premise or pass. Include the reason, unresolved evidence, availability constraints and who owns the next step. Avoid a weighted total that makes weak source credibility disappear beneath a large reach score. The current slate also matters. Two strong candidates may make nearly the same episode. A less familiar practitioner may add a missing perspective if the premise and evidence justify it. Review the sequence of conversations the audience will encounter, not just the individual booking. Before outreach, confirm the exact person, sender, purpose and proposed ask. Booking is a separate action from research. After acceptance, turn the shortlist into the fuller [interview research brief ](https://machinehouse.media/resources/editorial-decisions/interview-research-brief.txt) rather than giving the host a list of profile links. [Download the completed comparison and blank selection worksheet ](https://machinehouse.media/resources/editorial-decisions/guest-selection.txt). ## Questions about the method. Should I choose the guest with the largest following? + A larger following can be relevant to reach, but the guest still needs to serve the episode premise. Verify the contribution, evidence, communication fit and overlap with the rest of the slate before deciding. What is the difference between guest selection and guest research? + Selection decides who should take part and why. Research prepares the specific conversation, sources, questions, counterarguments and production needs once a candidate or premise is being developed. ## Keep exploring. [Guest research service ↗ ](https://machinehouse.media/services/podcast-guest-research)[Build the host brief ↗ ](https://machinehouse.media/guides/interview-research-brief)[Develop the show format ↗ ](https://machinehouse.media/guides/show-format-development)[Define the outcome before forecasting ↗ ](https://machinehouse.media/guides/content-performance-prediction) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/interview-research-brief [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Interview research brief # Prepare the host for the follow-up question. A completed interview brief with six questions, sources, follow-ups and gaps to resolve. See how Masoom’s research process informs preparation, then adapt the teaching example to your guest. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## A research brief is a route through a conversation. Put the episode’s question at the top of the brief, followed by the few things the host needs to remember. Keep sources and longer notes behind that. During the interview, the host needs to find the next question without searching through a biography. ## What should the research change? For The Masoom Minawala Show, the research task includes understanding the conversations a guest has already had in public. Repeated answers can show where an interview risks becoming familiar. Contradictions, recent work and less-explored experiences can suggest questions worth pursuing. Rishwajeet’s April 2026 account describes preparing research for 12 guests across more than 1,000 sources in six hours, with an 80-page brief per guest reported in the public case. The useful editorial question comes after that collection: what should Masoom now ask that she would not have known to ask before? Judge the brief at that point. A source should establish something, challenge an assumption or open a useful line of inquiry. A proposed question should have a purpose in the conversation. The host card should make the few important decisions accessible without requiring the host to perform the whole dossier aloud. The reported scale describes one preparation example. It is not a promise that every source on every guest was found, or that more pages automatically create a better episode. The completed teaching brief below makes the source, question purpose, follow-up and unresolved gap inspectable in a smaller example. [Read the original research account. ](https://machinehouse.media/writing/twelve-guests-six-hours) ## A completed teaching brief. Proposed episode, not a booked interview. “What the red patch on a heat map actually tells you” is an original teaching premise for curious city residents. No local temperature has been measured and no guest biography has been invented. | Brief field | The decision | Audience starting point | The viewer has seen a dramatic heat map and may assume its colours describe the air temperature they would experience on the street. | Episode question | What does this kind of map measure, and what else would we need to understand a particular location? | Guest requirement | Someone with verified direct experience of the relevant measurement methods; a specific identity and contribution remain to be researched. | Audience promise | Distinguish the measurement, its context and a claim about one street before interpreting the image. | Must survive the edit | The difference between surface and air measurements; the limits of an unverified local claim. | Do not imply | That an unbooked guest measured this location, or that the proposed film has produced a result. The [structured brief ](https://machinehouse.media/resources/editorial-decisions/interview-research-brief.json) contains the complete source map, question purposes, follow-ups, gaps and handoff responsibilities. The [text version ](https://machinehouse.media/resources/editorial-decisions/interview-research-brief.txt) includes a blank template. ## Put the limit beside the source. | Source | Useful support | What it does not establish | S1 · [NASA’s land-surface temperature explanation ](https://science.nasa.gov/earth/earth-observatory/global-maps/land-surface-temperature/) | A land-surface temperature map concerns the surface visible to the sensor; it differs from the air temperature in a weather report. | The temperature difference on a particular street, or a direct measure of a person’s experience. | S2 · [EPA’s account of trees and vegetation ](https://www.epa.gov/heatislands/benefits-trees-and-vegetation) | Shade and evapotranspiration are cooling mechanisms associated with vegetation. | One guaranteed effect size for every tree, place and time. Both sources were checked on 7 September 2026. They establish a starting distinction, not a completed local investigation. Before using a specific map, identify its dataset, acquisition time, unit and resolution. Clear the actual asset for publication. A source URL alone does not tell the editor which image or claim was approved. ## Six questions, each with a job. | Question | Follow-up | When someone points to the reddest patch and says this is the air temperature, what would you check first? | Can you point to the variable and unit in one cleared example? | What surface is the sensor actually seeing here? | Would the interpretation change if it sees a roof or a canopy? | How would you investigate the difference between two nearby streets? | What would you hold comparable, and what would still vary? | How do shade and water movement change the explanation? | Which claim would this evidence still not let us make? | What is the strongest reason this attractive map could mislead us? | What qualification must survive the trailer and thumbnail? | What should a viewer check before sharing the next map? | Give us the shortest accurate version, then tell us its limit. This is a proposed arc, not a script the guest has agreed to perform. Begin by making the measurement concrete, then examine the missing evidence and end with a usable action. If a guest corrects the premise, follow that correction. Repeating the next prepared question can discard the most valuable turn in the conversation. The worksheet records each question’s purpose and source status. A follow-up should respond to the answer: ask for the example, the decision, the counterexample or the evidence that would change their view. Avoid writing all follow-ups as variations of “tell us more.” ## Make uncertainty actionable. A useful gap list names what is missing and who will resolve it. Here, the producer needs a real guest and access; the researcher needs the selected dataset and evidence; the visual producer needs cleared assets; the host needs the essential qualification. None of those tasks is completed by adding more general articles to the folder. Distinguish an unresolved fact from an open editorial question. “Which sensor produced this image?” needs an answer before presenting it. “Which example makes the distinction easiest to understand?” is a creative choice that can be tested. Keep both visible, but do not let the second hide the first. Before recording, check the brief’s identity, source links, question arc and production dependencies. After recording, attach verified source timecodes and update claims against what was actually said. The pre-interview packet does not become an accurate transcript simply because the conversation followed a similar outline. ## End with a host card and a usable handoff. The host card for this episode has four reminders: ask for the measurement behind the colour; separate established facts from proposed questions; do not rescue an unsupported local number with confident narration; preserve the distinction in the title, trailer and graphics. That is enough to hold the editorial intention while the source packet supplies detail. An editor should receive the approved claim wording, exact source references, clear asset status and unresolved items with owners. Link this to the [shoot plan ](https://machinehouse.media/resources/editorial-decisions/shoot-plan.txt) so the crew captures what the argument needs. [Download the complete brief and blank template ](https://machinehouse.media/resources/editorial-decisions/interview-research-brief.txt). ## Questions about the method. What belongs in a podcast research brief? + The audience and episode question, verified guest relevance, source-backed claims and limits, a question arc with follow-ups, counterarguments, open gaps, production requirements and a concise host card. Can AI prepare the brief? + It can help organise supplied material or propose questions, but identity, source meaning, claims and missing evidence still need verification. A generated biography or invented quote must not enter the host’s factual brief. ## Keep exploring. [Select the guest ↗ ](https://machinehouse.media/guides/podcast-guest-selection)[Commission guest research ↗ ](https://machinehouse.media/services/podcast-guest-research)[Fact-check the script ↗ ](https://machinehouse.media/guides/fact-checking-content) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/editorial-slate [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Editorial slate planning # Choose the next argument before filling the calendar. See how Roshan’s comments changed the next essay, then work through a six-premise practice slate. Each idea has a question, evidence to find and a production route. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Decide what to commission. After Roshan’s IT video broke out, the comments gave us a subject for the next essay: graduates struggling to find work. The decision came before the calendar slot. The exercise below asks you to do the same for six possible pieces. ## The hit that changed the next commission. After Roshan Kotla’s Indian IT essay broke out, the obvious follow-up was another video about IT. Rishwajeet’s July 2026 account describes resisting that sequel and reading all 443 comments instead. The comments suggested a wider concern: people had followed the expected route through education and a supposedly safe degree, yet could no longer rely on the job at the end. Roshan had experience in software and hiring that gave him something specific to contribute to that concern. That changed the next commission. The follow-up addressed graduates struggling to find work. A third essay approached the middle-class career promise. The recurring audience concern remained; the subjects and arguments changed. | Decision | What the case shows | Tempting next topic | A sequel about the same industry as the breakout video. | Audience evidence | A reading of 443 comments, interpreted as anxiety about the education-to-job promise. | Creator’s contribution | Roshan’s own experience of software, running a business and hiring. | Editorial choice | Find another subject where that concern met his experience. | Commission | An essay about graduates unable to find work, followed by the middle-class career promise. For your own slate, look for the point where a meaningful audience question meets something the creator can explain from experience. Then test whether each proposed piece has a distinct argument. Changing the nouns in the last successful title is not enough. The three releases show how the direction developed. Their published cumulative view totals were recorded at different ages; they do not isolate why the follow-ups worked or establish a guaranteed formula. Comments are also a self-selected part of the audience. Treat the interpretation as a reasoned editorial choice that subsequent work can challenge. [Read the case and dated release figures. ](https://machinehouse.media/work/roshan-kotla) [Read Rishwajeet’s original account. ](https://machinehouse.media/writing/did-it-twice-more) ## Six leads through one editorial lens. Original proposed teaching slate. A city-explainer channel helps residents examine what everyday urban observations actually mean. These are research leads, not verified findings or scheduled releases. | Proposed premise | Distinct audience job | Dependency | What the red heat-map patch tells you | Interpret a measurement before sharing a dramatic claim. | A suitable dataset, relevant expert and cleared visual example. | Where the shade moves | Notice how an observation changes with time. | Location access and a feasible repeatable visual plan. | Who keeps a street tree alive | Understand responsibility after an installation. | A practitioner with firsthand work and supporting records. | Why this puddle returns | Investigate a recurring local problem without assuming its cause. | Appropriate expertise and actual site evidence. | What a bus-stop map leaves out | Understand a transfer or route information problem. | Current transit information and a cleared observation plan. | The sign everyone walks past | Examine a wayfinding decision. | Permission and real observation; no invented user behaviour. The topics share an editorial lens, but each asks a different question. The slate is not six variations of “cities should be better.” Its usefulness depends on the evidence and production route behind each premise. ## Choose what is ready to become a commission. For each lead, write the proposed argument, strongest counterargument, source route, visual or conversational engine and production dependency. A researcher should be able to say what would make the premise fail. “We will find some interesting facts” is not a sufficient research plan. The heat-map piece is the first research candidate because it already has a clear measurement distinction and a primary-source starting point. It still needs a suitable actual asset and verified guest. The shade observation may offer a stronger visual encounter, but its feasibility depends on access and timing. Keep it in development until that is resolved. Do not confuse confidence with importance. A worthwhile question with weak evidence needs more research. An easy topic with little audience value may not deserve a slot. Use scores, if helpful, to expose a discussion; do not let a total replace the underlying judgment. ## A four-slot calendar with real dependencies. | Slot | Work and owner | Release condition | First | Heat-map explainer · researcher and editor | Dataset and guest fit verified; claim boundaries and visual rights resolved. | Second | Shade observation · producer | Access and a workable observation period confirmed; no predetermined result. | Third | Tree-maintenance interview · booking owner | Contributor and evidence support the promised conversation. | Fourth | Response to a meaningful audience question · editor | Reserve capacity; select from an actual response after release rather than inventing the audience’s question. These are relative slots, not dates already booked. Convert them into a schedule only after checking who can do the work and when decisions are needed. Leave space for research failure, late changes and a useful response to the first releases. A blocked prerequisite should change the plan; a date does not make unsupported material publishable. ## Let results change a specific next choice. Record the hypothesis before publication. For the first piece, it might be that a concrete map-reading problem gives unfamiliar viewers a reason to watch an explanation of measurement. Then inspect the actual opening, viewing behaviour and relevant audience responses. A large view count alone does not say which part of that hypothesis held. Look for explanations that change the next slate: a misunderstood distinction, demand for a practical follow-up, or evidence that the packaging attracted a different audience. A few comments can suggest a question; they do not represent every viewer. Read them alongside the distribution context and the content itself. In the Roshan example above, the audience reading changed the subject of the next commission. Apply that level of specificity to a slate review: name the piece you would now make differently and the evidence behind the choice. ## Use a slate that can carry a decision. A usable row includes the audience question, proposed argument, evidence and counterargument, production needs, owner, next decision and expected response. Add the publication date when the piece is commissioned. Preserve the earlier decision so the postmortem can compare intention with what shipped. Review the slate at a rhythm that matches production. Change what new evidence warrants, while giving makers enough stability to finish. The aim is a stronger sequence of work, not a meeting that constantly reopens every idea. [Download the completed slate, relative calendar and blank worksheet ](https://machinehouse.media/resources/editorial-decisions/editorial-slate.txt). ## Questions about the method. How do I choose YouTube video topics? + Start with a specific audience question and the channel’s editorial promise. Compare possible arguments, evidence, visual or conversational potential, overlap with recent work and production feasibility before assigning a date. How is an editorial calendar different from editorial strategy? + Strategy establishes audience, purpose and the recurring editorial lens. The slate selects particular pieces. The calendar coordinates the owners, dependencies and dates needed to produce them. ## Keep exploring. [Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Develop a repeatable format ↗ ](https://machinehouse.media/guides/show-format-development)[Review what a release taught you ↗ ](https://machinehouse.media/guides/content-postmortem) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/fact-checking-content [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Fact-checking a video script # Check the words. Check the picture too. Work through six fact-checking decisions: unsupported figures, misleading comparisons, an invented quote and graphics that claim too much. Download the register with corrections and source notes. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Check exact statements, including the visual ones. A temperature map can show surface heat while the narration describes air temperature. Both measurements may be accurate, but putting them together can mislead the viewer. This guide follows six such claims through the decision to keep, change, remove or investigate them. Mark numbers, dates, quotations, identities, comparisons and causal statements. Also mark visual assertions: a recreated chart presented as observed data, a stock shot standing in for a location, or a caption assigning a statement to the wrong speaker. The source list is the starting point; the job is to determine what each source actually supports. ## Repair an attractive but unsupported explanation. Original teaching example. The proposed episode concerns a city heat map. No local temperature data or guest interview has been collected. Its scientific starting points are [NASA’s land-surface temperature explanation ](https://science.nasa.gov/earth/earth-observatory/global-maps/land-surface-temperature/) and [EPA’s account of trees and vegetation ](https://www.epa.gov/heatislands/benefits-trees-and-vegetation). | Draft claim or treatment | Check and decision | “The red patch proves the air here is 12 degrees hotter.” | Reject. No selected dataset, comparison, unit context or local air measurement supports this number. | “The satellite map measures the same thing as the weather report.” | Correct the measurement distinction using the NASA source. | “Trees can cool through shade and evapotranspiration.” | Retain with appropriate scope; EPA supports the mechanisms, not a guaranteed local amount. | “Every tree cools every street by the same amount.” | Reject the universal claim. The supplied evidence does not establish it. | “Our guest measured this exact street.” | Remove until a real guest’s work and the location are verified. No guest is booked in this example. | Paint a red region on an invented map and present it as data. | Label it as an illustration or replace it with cleared actual data and sufficient metadata. ### A corrected passage A land-surface temperature map tells us about the surface seen by the sensor. It is not the same measurement as the air temperature in a weather report. Trees can influence temperatures through shade and evapotranspiration, but these sources do not establish a particular temperature difference on this street. We would need the relevant data and measurement context to make that claim. ## Trace the claim to the strongest usable evidence. Prefer the original dataset, study, record, recording or firsthand account when it can answer the question. A secondary article may be useful context, but check whether it has narrowed, rounded or expanded what the original says. Several pages repeating one claim do not become independent corroboration. Record the exact source location, the version or date checked, and the scope: place, population, period, unit and denominator. A percentage without its base, an old cumulative count presented as a recent result, or a comparison between different measurement windows can sound precise while being wrong. For quotations, compare the exact words with the recording or authorised original and retain the surrounding meaning. A paraphrase belongs outside quotation marks. An automated transcript helps locate a passage; it is not authority for words that the recording does not contain. The [Associated Press’s published news values ](https://www.ap.org/about/news-values-and-principles/) provide an accuracy and correction reference; the worksheet here is our practical production method. ## Choose retain, narrow, qualify, remove or investigate. Not every failed check means discarding the whole argument. Sometimes the claim needs a narrower subject, an explicit date or the correct unit. Sometimes the compelling number has no defensible source and must go. Preserve the useful question while removing certainty the evidence cannot carry. A qualification should sit where it changes the interpretation. Burying “results may vary” in a description will not fix a spoken promise of a universal result. Equally, do not overload a simple supported sentence with caveats that do not affect its meaning. The aim is accurate understanding. Give unresolved claims an owner and a deadline. A missing source does not become verified when the script is approved for tone. If the evidence cannot be obtained, decide whether the argument can stand without that claim before the shoot or finishing work makes it expensive to change. ## Carry the check through the edit and packaging. After picture lock, check the actual spoken words, on-screen text, chart labels, image credits, captions, thumbnail and description against the approved register. A safe script can become unsafe to trust after a punchier title or a shortened clip. Log changed claims rather than assuming the earlier review still applies. If a mistake is found after release, identify the affected asset and the correction needed. Preserve an internal record of what changed and when; use an appropriate visible correction where the audience needs it to understand the work. Do not silently treat an amended page as proof that the original publication was correct. The [editable fact-check register ](https://machinehouse.media/resources/editorial-decisions/fact-check-register.txt) includes the six decisions above, corrected wording and fields for the actual source, reviewer, date, affected assets and correction status. Use it alongside the [research brief ](https://machinehouse.media/guides/interview-research-brief) and [approval map ](https://machinehouse.media/guides/content-approval-workflow). ## Questions about the method. Can a fact checker verify a script from its source links alone? + The links must be read against the exact claims. Check source scope, dates, units, quotations, comparisons and what the source does not establish. The finished edit and graphics also need review. Should I remove every uncertain statement? + No. Some questions and interpretations are useful when clearly identified. Narrow or qualify what the evidence partially supports, and remove or investigate claims that cannot be defended. ## Keep exploring. [Turn research into an argument ↗ ](https://machinehouse.media/guides/research-to-script)[Prepare the interview brief ↗ ](https://machinehouse.media/guides/interview-research-brief)[Content production ↗ ](https://machinehouse.media/services/content-production)[Evaluating AI content systems ↗ ](https://machinehouse.media/guides/evaluating-ai-content-systems)[Quote checker ↗ ](https://machinehouse.media/labs/quote-checker) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-shoot-plan [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Content shoot planning # What must you bring back from the shoot? A proposed half-day shoot for a heat-map explanation, with a schedule, shot list, crew responsibilities, fallbacks and editor handoff. Use the worksheet to plan your own recording. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Start with the evidence the finished piece needs. For the heat-map film in this example, recording an interview will not be enough. The editor also needs a map with its date and measurement details, a clear explanation of surface and air temperatures, and the qualification on what the map can tell us about one street. ## A proposed half-day plan. Original teaching schedule. No guest, venue or crew has been booked, and no film has been produced. Times are illustrative allocations to be replaced with the actual production plan. | Time | Work | What must be true before moving on | 08:30 | Set up and confirm access. | Location and contributor permissions are clear; private material is controlled; responsibilities are assigned. | 09:00 | Camera, sound and sync check. | Play back a recorded sample. Check actual picture and speech, not only meters or a red record light. | 09:20 | Interview around the selected map. | Host covers the variable, unit, time, interpretation and strongest limitation; researcher logs missing answers. | 10:20 | Break and media check. | Verify the files exist and can be played; identify pickups while the contributor remains available. | 10:40 | Capture explanatory pickups. | Record the cleared legend and metadata, a readable demonstration and any missing qualification. | 11:20 | Clarification and sound. | Capture the essential follow-up and appropriate room tone; no staged local measurements. | 11:40 | Verify copies and handoff. | Complete the manifest, verify agreed copies and confirm the editor can access the approved material. The sequence includes buffers and decision points. Actual setup time depends on the location, equipment, people and requirements. If a recording test fails, repair it before consuming the interview slot. The schedule serves the capture; it should not pressure the crew to ignore a known failure. ## Make each shot carry part of the explanation. | Material | Editorial purpose | Dependency | Interview coverage | Follow a clear explanation and the host’s challenge. | A framing and sound plan suited to the actual room and contributors. | Map legend and metadata | Show what is measured, in which units and when. | A cleared actual source, readable scale and accurate crop. | Original conceptual illustration | Make surface versus air measurement legible. | Researched wording and an explicit illustration label. | Follow-up answer | Keep the limitation inside the story. | Host and researcher identify the missing qualification before wrap. A beautiful tree montage cannot establish a temperature difference at a location. Use atmosphere where it belongs, but keep it separate from evidence. Record the source and usage status of every external visual so an editor does not have to guess whether an attractive asset is cleared. If the piece requires essential visual information to be accessible, plan how the script, description or transcript will convey it. [W3C’s media-planning guidance ](https://www.w3.org/WAI/media/av/planning/) explains why accessibility belongs in pre-production. Captions and a transcript should be planned as deliverables, with owners and review time. ## Name the decisions behind the logistics. The producer owns access, scheduling and the go/no-go decision. The host owns the conversation. The researcher owns source metadata and claim checks. Camera and sound owners verify capture. The editor receives the actual material and approved story scope; a named release owner accepts the finished version. A small team can combine roles when that is practical. Make the combination explicit so nobody assumes a check was someone else’s job. Keep personal contact details and logistical information in the private production sheet, not a public example or webpage. Confirm the recording format, frame rate, audio configuration, orientation and delivery needs for the actual project. There is no universal camera setting that replaces the destination, edit and accessibility brief. A sample recorded and opened in the intended workflow is more useful than an equipment list alone. ## Decide what happens when the plan fails. If the guest is unavailable, reschedule or approve a separately authored explainer; do not present a substitute narration as their interview. If the map cannot be cleared, use an original labelled conceptual illustration and remove claims that depend on the missing local data. If outdoor conditions make observation unsuitable, drop the optional location work or choose another date. Agree who can stop the shoot for unusable sound, lost access or another material problem. A fallback should preserve the audience promise where possible and make a scope change visible where it cannot. “We will fix it in post” is not a plan for evidence that was never captured. ## Leave the editor with files they can use. Create a manifest with file IDs, exact names, recording source, takes, transcript references, permissions and notes. Confirm that copies exist on the agreed separate storage, verify integrity where hashes are part of the workflow and play back representative material. Then check the editor’s actual access. A successful upload to a folder is not proof that the recipient can open it. The brief should identify what is usable, what is missing and who will resolve each gap. Connect it to the [editor handoff ](https://machinehouse.media/guides/video-editor-brief) rather than sending a folder with “everything is in here.” [Download the completed shoot plan and blank worksheet ](https://machinehouse.media/resources/editorial-decisions/shoot-plan.txt). ## Questions about the method. Is a shoot plan the same as a call sheet? + A call sheet coordinates people, place and time. The shoot plan also connects the story to required evidence, questions, shots, capture checks, fallbacks and the editor handoff. What should be checked before wrapping a content shoot? + Confirm essential answers and visuals were captured, files exist and play correctly, agreed copies are verified, permissions and source notes are recorded, and missing material has an owner. ## Keep exploring. [Prepare the interview ↗ ](https://machinehouse.media/guides/interview-research-brief)[Prepare the editor handoff ↗ ](https://machinehouse.media/guides/video-editor-brief)[Content production ↗ ](https://machinehouse.media/services/content-production)[Accessible content production ↗ ](https://machinehouse.media/guides/accessible-content-production) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/audience-retention-analysis [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Audience retention analysis # Read a retention graph alongside the video. Investigate three moments on an invented retention curve: an opening drop, a dip and a spike. Use the video’s content and audience context to choose an explanation to test. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Retention locates a question in the video. When the graph drops, watch that part of the video. What was the viewer waiting for? Did the explanation get confusing, or did the audience get the answer and leave? Write down the change you can see before deciding which explanation to investigate. In [YouTube’s audience-retention documentation ](https://support.google.com/youtube/answer/9314415?hl=en), dips can reflect skipping or abandonment, while spikes can involve rewatching or sharing. A spike may also indicate that something was difficult to understand. The documentation also describes the first-thirty-second introduction measure and comparisons with similar-length videos. Those are useful report definitions, not universal creative targets. ## A worked curve with three questions. Entirely synthetic teaching data. This curve is not a YouTube export and does not report Machine House or client performance. The six-minute example makes three moments easy to inspect. Illustrative segment retention. Exact teaching points are in the downloadable data; no causal explanation is encoded in the line. | Moment | Observation | What to inspect | 00:00–00:30 | The line goes from 100% to 77%. | Does the opening fulfil the title’s promise? Who arrived, and from which sources? | 02:00–02:30 | The value falls from 64% to 50%. | The teaching sequence repeats an earlier definition here. Is it repetition, a confusing transition or a change in the audience? | 03:00–03:30 | The value rises from 53% to 66%. | A comparison diagram appears. Did viewers return because it was useful, unclear or shared? The content descriptions are part of the invented example. They are candidate explanations to examine, not conclusions produced by the numbers. The [downloadable dataset ](https://machinehouse.media/resources/editorial-decisions/measurement-example.json) contains all thirteen points and explicitly separates this sparse curve from the aggregate video examples. ## Compare the right viewing situations. Before calling retention good or bad, record video duration, format, age at measurement and how the audience arrived. A search-led answer, a returning-audience interview and a broadly recommended film can create different viewing situations. Compare relevant segments where the platform makes them available and retain the definitions used in the report. A percentage viewed and an average viewing duration answer different questions. In a six-minute example, 60% corresponds to 3:36; in a twenty-minute film it corresponds to twelve minutes. That arithmetic does not make the longer film automatically better. The intended job and the viewer’s experience still matter. Use a suitable comparison cohort and enough observations to avoid turning a noisy early result into a rule. An average can conceal a strong opening followed by an unearned middle, or a weak introduction followed by a valuable explanation. Read the shape and the material together. September 2026 measurement note: a public playback-start count does not establish how many people stayed. Keep the retention report and its stated basis with the curve. Do not reconstruct average viewing or subscriber conversion using a different public counter. YouTube’s API defines engaged views as viewing beyond the first frame or a click or tap to play; the name does not guarantee a substantial watch. [Metric definitions ](https://developers.google.com/youtube/analytics/metrics) · [Why the distinction matters. ](https://machinehouse.media/writing/youtube-changed-what-a-view-means) ## Write the observation separately from the explanation. ### A review note that can lead to a test Observation: in the synthetic example, retention is fourteen percentage points lower at 02:30 than at 02:00. The segment repeats a definition already established. Hypothesis: the repeated explanation delays the next useful step. Alternative: the transition is unclear, or this audience is seeking an answer elsewhere in the video. Next test: in a comparable future piece, move directly from the definition into a worked application. Preserve the title’s promise and record the actual treatment. This note says what to change and why, while leaving the conclusion open. “People have short attention spans” does neither. Do not cut every pause simply because the line slopes down. Some pauses let a difficult point land; some fast sequences lose the audience because the relationship between ideas is missing. ## A popular moment may need explanation, not repetition. For the spike at 03:30, inspect whether the diagram is readable, whether it introduces new terms and whether the narration gives the viewer time to understand it. Look for relevant audience questions without treating a few comments as a representative survey. If the graphic was unclear, copying it into every future video would repeat the problem. If the moment works as a standalone insight, it may suggest another piece. Build that decision from the complete context and the audience’s next question. A retention spike alone does not determine that a clip will perform well on another platform or that the same packaging promise will remain accurate. ## Turn the analysis into one editorial decision. End the review with the observation, competing explanations, selected change, owner and evidence to check next time. Keep the original hypothesis and actual release treatment so the next review can assess whether the intended change happened. Do not rewrite the explanation after seeing a result and pretend it was the original prediction. Equally, do not preserve a favourite theory when a later comparison contradicts it. The [postmortem guide ](https://machinehouse.media/guides/content-postmortem) shows how to connect retention to equal-age performance and production context. [Download the review worksheet ](https://machinehouse.media/resources/editorial-decisions/postmortem-template.txt). ## Questions about the method. Does a retention spike prove viewers loved that moment? + No. It can reflect rewatching, sharing or difficulty understanding the material. Inspect the actual sequence and relevant context before choosing an explanation. What is a good YouTube retention percentage? + There is no single percentage that fits every duration, format and audience situation. Use relevant comparisons and inspect the actual moments, traffic context and intended job of the video. ## Keep exploring. [Commission a channel audit ↗ ](https://machinehouse.media/services/youtube-channel-audit)[Compare releases fairly ↗ ](https://machinehouse.media/guides/content-postmortem)[Keep an honest prediction record ↗ ](https://machinehouse.media/guides/content-performance-prediction) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-postmortem [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Content postmortems # Compare the releases on equal terms. Start with Masoom’s catalogue totals, then compare three invented videos at the same age. Work through the arithmetic, possible explanations and a decision for the next release. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Return to the decision made before release. Twenty-seven per cent of the views in Masoom’s 2026 comparison came from older episodes. If we credited that whole total to new releases, we would be reviewing the wrong body of work. Start by identifying what the result contains, then return to the decision you made before publication. Separate three questions: what happened, what might explain it and what to do next. Numbers answer parts of the first. The content, audience and production context help with the second. The third requires an editorial choice whose uncertainty remains visible. ## Define what the total contains. The September 2026 Masoom case compares the show’s long-form catalogue from 1 January to 3 September in 2025 and 2026. The 2026 period recorded 4.66 million views. Of those, 1.26 million—27%—came from older episodes. The 35 episodes released in the current-year period contributed 3.40 million. That split changes the interpretation. The total describes the audience’s use of the catalogue during the period. It cannot all be assigned to the new releases. Older work is still part of what the show is delivering, and a review of only the latest episodes would miss a substantial share of the viewing. Before interpreting a release report, write down which question you are answering. How much viewing did the catalogue generate this year? How did newly published work contribute? How did two episodes perform in their first seven days? Each can be useful, but the comparison needs the corresponding set of videos and measurement window. The Masoom comparison includes sponsored episodes and older long-form releases, and excludes personal uploads and Shorts. It records show outcomes from the combined work of Masoom and the team. It is not an equal-age comparison of individual videos and does not isolate the effect of a particular editorial change. The exercise below addresses the separate equal-age question with invented data whose arithmetic can be followed in full. [Read the case and its measurement scope. ](https://machinehouse.media/work/masoom-minawala) ## An unequal-age comparison gives the wrong starting point. Entirely invented teaching dataset. Video V01 has 80,000 views at thirty days. V02 has 18,000 at seven days. Calling V01 the stronger first-week release from those two totals would be wrong. Use the same measurement window before drawing that comparison. | Release | Duration | First 7 days: views | First 7 days: average viewed | V01 · What the red patch measures | 6:00 | 12,000 | 3:48 · 63.3% | V02 · Where the shade moves | 6:00 | 18,000 | 3:18 · 55.0% | V03 · Who maintains the tree | 8:00 | 9,000 | 4:48 · 60.0% In this equal-age view, V02 has more views than V01, but a lower average viewing duration. V03 has fewer views and a longer average viewing duration; its longer runtime matters to that interpretation. The correct comparison is more informative than a single winner label. ## Check the arithmetic and the context. | Release | Views × average viewing seconds ÷ 3,600 | Watch hours | V01 | 12,000 × 228 ÷ 3,600 | 760 | V02 | 18,000 × 198 ÷ 3,600 | 990 | V03 | 9,000 × 288 ÷ 3,600 | 720 These aggregate teaching numbers reconcile exactly by construction. In real exports, rounding, different filters or different report definitions may create small differences; investigate material mismatches before analysing them. Do not merge snapshots with incompatible timezones or cutoffs. V02’s synthetic external-traffic share is 22%, compared with 8% for V01. That difference may be relevant to who watched, but it does not establish that external traffic caused the retention difference. The topics also differ. This is an observational comparison, not a controlled test of one creative change. Measurement update · 9 September 2026. YouTube changed public view counting on 24 August to count playback starts across formats. Before comparing periods across that date, record the exact report, metric definition and export date. Reconcile a comparable series where possible; otherwise label the break. The field name “views” alone does not establish comparability. [YouTube’s explanation ](https://support.google.com/youtube/answer/2991785?hl=en) and our [analysis of what the change means ](https://machinehouse.media/writing/youtube-changed-what-a-view-means) explain the distinction. ## A complete postmortem decision. ### The review record Original intention: test whether a concrete observation can give unfamiliar viewers an accessible entry into a city-systems explanation. What happened: V02 reached more viewers in the first seven days and generated more total watch time than V01, while its average viewing duration was lower. Its traffic mix differed. What remains unknown: whether the difference came from topic, packaging, opening, audience mix or another factor. The totals cannot isolate those effects. Decision: develop a comparable next explainer with a concrete opening, then inspect the transition into the main explanation. Keep the promise accurate and log the actual title, opening and distribution treatment. Owner and next evidence: the editor records the change before publication; the analyst reviews the same first-seven-day window and relevant audience segments after data are available. ## Include the work needed to make the result. A release can perform well while depending on unsustainable access, overtime or revision. If those costs were measured, include them alongside the content result. If they were not, record the gap rather than inventing an efficiency improvement from a smoother-looking process. Capture what the team learned about the brief, research, recording, edit and approval sequence. A missed answer may explain an awkward middle better than a stylistic critique. A late premise change may have consumed the time intended for packaging. Give each repair an owner and connect it to the next production plan. Keep the conversation about decisions and conditions the team can change. A postmortem that ends with “try harder” has not produced a useful operating lesson. ## Keep one record that the next release can inherit. Save the release ID, exact treatment, publication time, data cutoff, cohort, observations, competing explanations and chosen action. Link the actual source data. Do not turn every review into a new disconnected report that nobody consults before the next commission. The [Roshan case ](https://machinehouse.media/work/roshan-kotla) is a real account of an audience-informed follow-up; its published cumulative snapshots are not the equal-age dataset used here. The [synthetic data ](https://machinehouse.media/resources/editorial-decisions/measurement-example.json) and [blank postmortem worksheet ](https://machinehouse.media/resources/editorial-decisions/postmortem-template.txt) make the comparison method inspectable without inventing a client experiment. ## Questions about the method. Can I compare two videos by their current total views? + You can describe those totals with their ages, but they do not establish equal-age performance. Use the same time since publication, consistent filters and relevant context for that comparison. What should a content postmortem produce? + A verified account of what shipped and happened, plausible explanations with limits, and a concrete next decision with an owner and evidence to review. It should not end at a dashboard summary. ## Keep exploring. [Read the retention curve ↗ ](https://machinehouse.media/guides/audience-retention-analysis)[Choose the next slate ↗ ](https://machinehouse.media/guides/editorial-slate)[Channel diagnosis ↗ ](https://machinehouse.media/services/youtube-channel-audit)[Content catalogue review ↗ ](https://machinehouse.media/guides/content-catalogue-review) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-business-impact [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Measuring content business impact # Follow the inquiry past the view count. Follow six example records through deduplication, qualification, signed work and payment. See what the record can tell you about content’s contribution and where attribution remains uncertain. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · Updated September 9, 2026 ## Define the business job before collecting more numbers. The example starts with six records. After removing a duplicate, there are five inquiries, four qualified opportunities and two wins. The signed value and money received are different totals. Keeping those steps separate lets you discuss what content contributed without counting every form submission as a sale. For a service business, a qualified inquiry might mean a real organisation with a relevant problem, an appropriate scope and a plausible buying process. Write that definition before scoring the leads. A form submission, a call booking, a proposal, a signed engagement and received money are separate events. ## Reconcile six records into five actual inquiries. Entirely invented teaching records and amounts. No Machine House or client commercial result is claimed. The original contact data are included in the [downloadable example ](https://machinehouse.media/resources/editorial-decisions/measurement-example.json). | Contact | Observed or reported path | Status at cutoff | L01 · P01 | Paid search → guide → direct inquiry; reports search. | Qualified; no won engagement. | L02 · P02 | LinkedIn → guide → newsletter → inquiry; reports founder post. | Qualified and won; ₹100,000 booked, ₹60,000 received. | L03 · P03 | Organic search → guide → inquiry. | Not qualified under the chosen definition. | L04 · P04 | No tracked path; reports hearing the podcast. | Qualified; no won engagement. | L05 · P05 | Referral → guide → inquiry; reports a colleague. | Qualified and won; ₹80,000 booked and received. | L06 · P02 | A second record for the same P02 inquiry. | Duplicate of L02; do not count another inquiry or sale. The six records become five unique inquiries, of which four are qualified and two are won. Booked value is ₹180,000; received money is ₹140,000 at the cutoff. The outstanding ₹40,000 is not cash received. A reporting system that adds every form event or treats a booking as revenue will overstate the result. ## Keep observed paths and reported influence together. Use stable content IDs and sensible campaign names, then connect the records you are authorised and able to connect. Ask a concise discovery question during the inquiry: how did they first hear about the business, and which piece helped them understand the offer? Preserve the answer rather than forcing it into the channel your analytics happened to record. L04’s podcast answer is self-reported evidence with no tracked path. It should not be erased because a browser report cannot see it, or upgraded into a fully observed multi-step journey. L02 has a recorded newsletter visit and reports a founder post; those describe different parts of the encounter. Tracking can miss devices, private sharing and visits outside the observable system. Identity matching and consent also constrain what is available. Report those limits. More detailed tracking does not automatically produce a complete explanation of why a person chose to buy. ## Separate assisted influence from assigned credit. Define “content-assisted” explicitly. For this example, it means a won contact encountered the guide before the inquiry. Both won contacts meet that definition, so ₹180,000 is guide-assisted booked value. That does not make the guide the exclusive cause of either engagement. The same sale can have several meaningful touches, so assisted amounts across assets must not be added as unique revenue. An attribution model assigns credit using a rule or a model. [Google Analytics documents its attribution approach ](https://support.google.com/analytics/answer/10596866?hl=en); record the model, scope and lookback window used in your own report. Changing the rule can change credited values without changing what the customer did. The raw example does not contain an analytics model export, so it does not claim an official attributed revenue figure. It lets the reader inspect the path and the chosen assisted definition. Keep observed activity, self-report and model-assigned credit as separate fields. ## A useful touch is not a measured incremental effect. Incremental impact asks what would have happened without the content. A path containing a guide does not answer that counterfactual. The prospect may already have been ready to buy; the guide may have accelerated a decision; or another touch may have mattered more. The records alone cannot distinguish those explanations. Where feasible, a properly designed holdout or controlled experiment can help estimate an incremental effect. Define the intervention, outcome, comparison and measurement period before running it. Low-volume service businesses may not have enough observations for a useful estimate; a before-and-after change with many moving parts should remain an observational finding. You can still make a practical decision with incomplete causal evidence. Repeated qualified inquiries describing the same useful piece may justify further work on that subject, while you continue to record the uncertainty. Avoid turning that judgment into a percentage lift the design cannot support. ## Report value and cost without mislabelling ROI. The synthetic example assigns ₹90,000 to content cost. Dividing ₹180,000 of assisted booked value by that cost produces 2.0. It is not an ROI calculation: some money is uncollected, delivery costs are absent, the assisted definition is not exclusive and incremental profit is unknown. For an actual investment review, state which production, media, software, management and review costs are included. Distinguish booked value, cash receipts and contribution after delivering the work. Use a consistent period and explain whether a long sales cycle leaves the cohort incomplete. The decision record should say what the evidence justifies now: continue a useful explanation, improve the qualification route, investigate a recurring objection or stop an activity whose job is unclear. [Download the impact worksheet and reconciliation ](https://machinehouse.media/resources/editorial-decisions/business-impact.txt). Define the denominator before comparing cost per view. A fixed fee divided by a more inclusive view counter produces an apparently lower unit cost even when the underlying viewing is unchanged. Label the platform, format, viewing event and period beside the price calculation. Keep this descriptive ratio separate from sales attribution or incremental return. [See the worked denominator example. ](https://machinehouse.media/writing/youtube-changed-what-a-view-means) ## Questions about the method. How do I measure content-assisted conversions? + Define the meaningful content touch and conversion event, deduplicate people or opportunities, preserve observed and self-reported paths, and state the attribution window and rule. Assisted involvement is not exclusive or necessarily incremental credit. Is a call booking revenue? + No. A booking, qualified opportunity, won engagement, invoiced amount and received payment are different states. Report each with its actual status and cutoff date. ## Keep exploring. [Plan distribution and measurement ↗ ](https://machinehouse.media/guides/content-distribution-plan)[Review the release ↗ ](https://machinehouse.media/guides/content-postmortem)[Content strategy ↗ ](https://machinehouse.media/services/content-strategy)[Content catalogue review ↗ ](https://machinehouse.media/guides/content-catalogue-review) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/youtube-changed-what-a-view-means [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 9 SEP 2026 · 4 MIN READ # YouTube changed what a view means A 95% growth claim on our own homepage is a good place to start. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media A Machine House essay The “+95% views” on our homepage came from a fairly careful comparison. We took The Masoom Minawala Show’s long-form catalogue, excluded Shorts and personal uploads, and compared 1 January to 3 September in 2025 and 2026. Views went from 2.39 million to 4.66 million. The calculation gives 94.8% growth. We need to check what that growth measures. On 24 August, near the end of our comparison period, [YouTube changed its public view counter ](https://support.google.com/youtube/answer/2991785?hl=en). A view now counts when a video starts playing, across long-form videos and live streams as well as Shorts. The earlier measure remains available as engaged views. Our export has a column called “views.” That was enough to do the arithmetic. It isn't enough to settle whether we're comparing the same measurement across both years. YouTube's own [API documentation ](https://developers.google.com/youtube/analytics/metrics) says that field's meaning depends on the report. I work with these numbers for a living, and we still have a claim to go back and check. ## YouTube mentions the sponsors The [announcement ](https://support.google.com/youtube/thread/433409976?hl=en&msgid=460983850) says the new counter will help creators “express their true scale and value to brand partners and sponsors.” It also says the change won't affect creators' earnings or eligibility for the Partner Program. Those are two quite different conversations. A creator can put the new public count in a pitch to a brand, while YouTube continues to use separate qualifying measures for its own programme. There is a reasonable case for counting starts. Creators should be able to see how often their videos begin playing, and using a consistent rule across formats makes that easier to follow. I would want access to the number too. But a sponsor may be trying to buy something further into the video: a product explanation, a recommendation, or enough time with the host to understand why they use it. The first frame doesn't tell them whether any of that happened. Suppose a video has 100,000 play starts, 40,000 engaged views and 1,000 likes. These are invented numbers for one video, not a before-and-after YouTube result. Divide the likes by engaged views and you get 2.5%; divide them by starts and you get 1%. Put the second rate next to the first in a report and the video appears to have become much less engaging. Now use the same two counts to report a ₹1 lakh sponsorship. The fee works out to ₹2,500 per thousand engaged views or ₹1,000 per thousand starts. You could make a fairly cheerful slide about that saving. The sponsor paid the same amount for the same video. ## The pitch can change faster than the work The public counter is convenient because the creator and the brand can both see it. Explaining what happened after the video started usually takes more information from the creator. So there's a reason for the simplest available number to become the number people negotiate around, even when both sides know it leaves things out. That puts an awkward choice in front of a creator. Lead with engaged views and a competitor's larger public count may look better. Lead with starts and you may be setting an expectation you'll later have to explain. The brand can also decide that a view is now easier to get and ask for more of them at the same fee. A higher counter doesn't settle who benefits from it. We've had versions of this problem for years. [Netflix calculates views by dividing hours watched by runtime ](https://about.netflix.com/en/news/what-we-watched-a-netflix-engagement-report). For a one-hour film, one person watching the whole thing and two people watching half each contribute the same amount. That's useful for comparing viewing volume. It would be a strange number to put beside YouTube playback starts and treat as a comparison of people reached. But both can appear in a pitch under “views.” The definition tends to get less space than the number. For our own claim, the next job is quite specific: check a comparable view series for both periods, rather than assume the field name guarantees it. “Engaged” needs care too; the API includes a click or tap to play in that definition, so it doesn't promise a substantial watch. The revised homepage leads with the separately recorded watch-hour result: about 103% growth. The two view totals stay in the case study, with the unresolved comparison beside them. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/services/video-scriptwriting [Home ](https://machinehouse.media/)/ [Services ](https://machinehouse.media/services)/ YouTube & video scriptwriting # An idea worth watching. A script that delivers it. Machine House develops research and scripts for YouTube videos, expert-led essays and brand films for creators and brands in India. Work with Rishwajeet on the premise, the argument and the words your host will actually say, with source references and direction for the edit. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Service · September 18, 2026 [Talk about your content ↗ ](https://calendly.com/rishwajeet/intro-call) ## A small channel. Three consecutive hits. Roshan Kotla’s channel had a median of around 100 views per video before its breakout. We helped turn his experience in software and hiring into ‘The Indian IT Dream is Dead’, then developed two further essays about graduate jobs and the middle-class career promise. By 7 September 2026, the three films had reached 791K, 93K and 91K views. Machine House developed the research, thesis, outlines, scripts and packaging. Roshan brought his experience, stories and delivery. Those are results of the finished films, with all those contributions involved. [Watch the three releases and read the full Roshan Kotla case. ](https://machinehouse.media/work/roshan-kotla) ## Bring the material you already have. You may have a subject and no clear angle, an expert with years of experience, or a draft that keeps losing its thread. Each needs a different starting point. - A topic: we research the subject and develop a premise specific enough to sustain a video. We agree what the viewer should understand by the end before expanding the outline. - Expert notes or a conversation: we find the examples and claims the speaker can stand behind, then build a sequence a new viewer can follow. - An existing script: we review the opening, information order, evidence and ending. The brief can cover a focused rewrite or a new structure. ## The script needs to work on the shoot. A sentence can read well and still be difficult to say. A statistic can be correct and still need a chart to make sense. We separate the spoken script from instructions for the host and editor so each person can find what they need. - Premise and outline: the audience, central question, argument and order of the evidence. - Source-linked draft: narration or host lines, examples and references supporting factual claims. Gaps that need the expert’s answer stay visible during review. - Visual and production notes: suggested demonstrations, graphics, B-roll and screen text, with essential visual evidence identified. - Review and handoff: revisions against the agreed direction and a script ready for the next production stage. We set the review rounds and approver when scoping the project. For an inspectable example, follow our [research-to-script guide ](https://machinehouse.media/guides/research-to-script) into the [editor brief and playable edit ](https://machinehouse.media/guides/video-editor-brief). It is a labelled teaching example with downloadable source material. ## Technology, careers and health. For [Scaler ](https://machinehouse.media/work/scaler), our work connected source research, a central argument, scripts, visual and production briefs, and packaging across 17 releases. The subjects included software careers and the AI industry. On [Rocket Health’s protein film ](https://machinehouse.media/work/rocket-health), we developed the research, script, visual direction and briefs for the host, crew and editor, working with Studio 73 on production. For a new subject, we agree who checks specialist claims. The expert reviews accuracy and their own position; we work on how the explanation unfolds for the viewer. ## One film, or an ongoing slate. Commission the writing while your team handles the shoot and edit, or discuss a wider [production engagement ](https://machinehouse.media/services/content-production). Rishwajeet leads full-service work with his team at The Breakout Company, which he co-founded. The proposal identifies the team and business delivering the work. Research depth, access to the expert, intended length, visual complexity and the review process determine the scope. For recurring work, we agree how topics are commissioned and when the host needs each script. Bring a channel link, the subject you want to cover, a reference video and any existing notes. If there is a shoot date, include it so we can plan backwards from the recording. ## Questions about the work. Can I hire Machine House just for YouTube scriptwriting? + Yes. We can develop the research, outline and script while your existing team records and edits. Visual notes, packaging and wider production can be included in the agreed scope. Do I need to provide a finished brief? + No. Bring the topic, audience and material you have. We can help develop the premise before drafting. Existing research or a recorded expert conversation gives us a more specific starting point. How do you keep the script in the creator’s voice? + We use the creator’s existing videos, examples and review feedback to shape the language. Their experiences and opinions need to come from them. The creator reviews the draft before recording. How much do video scriptwriting services cost? + We scope the work around research depth, length, format, access to the expert and review rounds. Share the subject and production deadline so we can define the deliverables and quote for the work. Can you also develop titles, thumbnails and Shorts? + Yes. Packaging and short-form adaptations are available alongside the main script. We agree those outputs separately so their research, writing and review time are included. Will a new script guarantee more views? + No. The premise, packaging, delivery, edit and audience response all affect a release. Our case studies document particular projects; they are not a promise of the same result for another channel. ## Keep exploring. [Titles and thumbnail strategy ↗ ](https://machinehouse.media/services/youtube-packaging)[Original Shorts from long-form material ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Content strategy and editorial direction ↗ ](https://machinehouse.media/services/content-strategy)[Checking the claims in a script ↗ ](https://machinehouse.media/guides/fact-checking-content)[A worked Hinglish script adaptation ↗ ](https://machinehouse.media/guides/hinglish-scriptwriting) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/clips-vs-original-shorts [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Clips versus original Shorts # Keep the moment. Or rebuild the idea. A good passage can become a confusing clip when its setup disappears. Compare a recorded excerpt with a newly written short from the same source, then use the decision brief on your own episode. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · September 18, 2026 ## The cut starts one thought too late. “For the classroom comparison, choose a projection that preserves relative area.” That sentence is useful halfway through our map-projection script. Cut it out on its own and a new viewer has to guess what the class is comparing. The missing information is small: students need to compare how much land different countries occupy. The production decision is whether the recording contains a clean way to restore that setup, or whether someone needs to record new words. This guide uses an [original teaching script ](https://machinehouse.media/guides/research-to-script). The audio is synthesized demonstration narration. These examples have not been released as client Shorts or tested for audience performance. ## This passage can stand on its own. The Mercator passage names the projection, explains its useful property and keeps the qualification about shortest routes. Its opening does not depend on the classroom setup. Recorded excerpt: 00:33.342–00:55.918 of the [full source audio ](https://machinehouse.media/resources/editor-brief/map-source.wav). The cut is 22.576 seconds, with no added or reordered words. Download the [audio excerpt ](https://machinehouse.media/resources/clips-vs-original-shorts/mercator-excerpt.mp3). Exact transcript · recorded excerpt On a standard Mercator map, a straight line can represent a constant compass bearing. That makes it useful for navigation, although the line is not generally the shortest route. Mercator also preserves angles locally. But as you look towards the poles, the areas become increasingly exaggerated. A map with a useful navigation property can give a misleading impression of how large a country is. A package such as “Why a navigation map changes country sizes” fits this explanation. Keep a standard Mercator map visible and label it. A more dramatic title claiming the map is useless would contradict the passage. ## A different question needs a new opening. Suppose the short is for a teacher choosing a classroom map. The useful opening is now the decision they face. Here is a new script built from the complete source: New spoken script · recording required You need a world map for a classroom. The students will compare how much land different countries occupy. A standard Mercator map makes that harder: areas become increasingly exaggerated towards the poles. Yet the same projection has a useful property for navigation. A straight line can represent a constant compass bearing, though it is generally not the shortest route. For the classroom, choose an equal-area projection. Country shapes may look unfamiliar, but their relative areas will support the comparison. The map you choose depends on what you need people to judge. This version needs a new recording. It changes the information order and adds a direct setup; it cannot be presented as an exact excerpt from the existing audio. The factual ideas come from the source’s opening, Mercator explanation, classroom recommendation and ending. ## Choose the edit before commissioning the output. - Use the recorded cut when the speaker, context and complete thought survive together. Preserve a distinctive performance when that is what makes the moment worth watching. - Record a pickup when a brief missing explanation can reconnect an otherwise useful passage. Check that the new introduction does not change the speaker’s claim. - Write a new short when the viewer question or information order needs substantial work. Link each claim back to the complete source and give the creator a script to approve. - Drop the candidate when the required qualification cannot fit naturally, the source does not support the proposed claim, or the piece offers nothing distinct from the others. Set the output count after reviewing the material. Five cuts of the same answer can repeat one idea while a quieter passage offers a different idea. ## Leave the editor a decision they can follow. The [editable decision brief ](https://machinehouse.media/resources/clips-vs-original-shorts/decision-brief.txt) contains both treatments, the source locations, visual requirements and a blank commissioning form. It makes new recording needs visible before an editor is assigned the job. On review, watch the piece as a new viewer: identify every “this”, “they” and “that comparison”, and check what each refers to. Then compare the final claim with the complete source. A cleaner opening is only useful if the meaning survives. For the map facts, [Esri’s Mercator documentation ](https://doc.esri.com/en/arcgis-pro/latest/help/mapping/properties/mercator.html) explains the constant-bearing property, increasing area distortion and the distinction from Web Mercator. The original [research packet ](https://machinehouse.media/resources/research-to-script/research-packet.txt) retains the broader sources. ## Questions about the method. Is a rewritten Short the same as a clip? + No. A clip uses an existing recording. A rewritten Short develops new spoken text from source ideas and needs a new recording or a separately agreed production treatment. How many Shorts should come from one episode? + Choose the count from the distinct, well-supported ideas in the source and the production time available. A fixed quota can force repeated ideas or cuts that lose necessary context. Can a title or caption fix missing context? + Sometimes a name or location can be supplied visually. If understanding the argument depends on an omitted explanation, restore that explanation in the piece rather than relying on the viewer to reconstruct it. Does rewriting improve performance? + These examples demonstrate editorial choices, not measured performance. Compare released pieces using their actual audience, format and measurement windows before claiming an improvement. ## Keep exploring. [Commission clips and original Shorts ↗ ](https://machinehouse.media/services/long-form-to-shorts)[Turn the decision into an editor brief ↗ ](https://machinehouse.media/guides/video-editor-brief)[Plan where each piece belongs ↗ ](https://machinehouse.media/guides/content-distribution-plan)[Research and write the next video ↗ ](https://machinehouse.media/services/video-scriptwriting) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/hinglish-scriptwriting [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Hinglish scriptwriting # Write the way the host would say it. Compare an English explanation, a literal adaptation and a spoken Hinglish script. The task is to preserve the idea while choosing words, examples and a rhythm the host can use. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · September 18, 2026 ## Keep the claim before changing the language. Start with a small explanation: a menu price and a checkout total can differ. Here is the English source for this exercise: English source · fictional example The menu price is only one part of what you pay. In this example, the food costs 199 rupees. A delivery charge of 35 rupees and a handling charge of 10 rupees bring the total to 244 rupees. If you are comparing two orders, compare the final checkout totals rather than the first prices you see. The figures are fictional. They describe no particular app or pricing policy. The claim to preserve is the comparison between final totals, including the stated charges. ## Correct words can still sound wrong for the speaker. Literal adaptation · formal register Menu ki keemat aapke dwara bhugtaan ki jaane waali raashi ka keval ek hissa hai. Is udaaharan mein bhojan ki keemat 199 rupaye hai. 35 rupaye ka delivery shulk aur 10 rupaye ka handling shulk kul raashi ko 244 rupaye tak le jaata hai. This version uses a more formal Hindi register. That may suit a particular presenter. For a host who normally says “menu”, “checkout” and “delivery charge”, it changes the way they speak and slows a simple explanation down. Choose the register from the host’s actual recordings and the audience’s familiarity with the subject. There is no useful universal quota of Hindi words and English words. ## Give the viewer a bill to follow. Spoken Hinglish · new script Menu pe 199 dikha. Checkout pe 244. Beech mein kya jud gaya? Is example mein, 35 rupaye delivery ke aur 10 rupaye handling ke. Order karne se pehle yahi total dekhna hai. Ab maan lo doosri jagah wahi order 219 ka hai, aur koi extra charge nahi. Menu pe woh mehenga lag raha tha. Checkout pe pehle order se 25 rupaye sasta hai. Dono jagah final amount compare karo. The script opens on the two amounts, then explains what changed. It adds a second fictional order so the viewer can make the comparison: 199 + 35 + 10 = 244; the other order totals 219. The difference is 25 rupees. That second order is new writing. It is not a literal translation of the source. Adding an example is useful only if it keeps the original claim true and its invented details are clear. ## Choose each switch for a reason. - Keep familiar task words. In this version, “menu” and “checkout” are words the imagined viewer already uses. For a different audience, explain the action in Hindi before using the label. - Keep the arithmetic audible. The spoken amounts and the receipt must agree. If the host reads a number differently, update the final captions to match the recording. - Leave the host room to choose. “Menu pe” and “menu par” can both preserve the meaning. The host’s own usage should decide; a style rule need not settle it for everyone. - Explain unfamiliar terms. If “handling” needs explanation, find out what the actual charge covers before defining it. This fictional example only identifies a ten-rupee line item. ## Finish with the spoken version. Have the host read the script at their normal pace. Mark awkward phrases, change them together and check that any qualifier or comparison survived. Keep those changes in the script the editor receives. For this exercise, the editor needs two clearly labelled sample receipts. The final comparison is 244 against 219. Showing the earlier menu prices at that point would undo the explanation. Agree whether captions use Roman Hindi or Devanagari, then review them against the recording. The [editable adaptation sheet ](https://machinehouse.media/resources/hinglish-scriptwriting/adaptation-sheet.txt) includes all three drafts, the arithmetic, visual notes and a blank host-review form. ## Questions about the method. How much Hindi and English should a Hinglish script use? + Choose the language mix from the host and audience. Keep words they use naturally, explain unfamiliar terms and judge the spoken passage as a whole rather than setting a percentage. Should a Hinglish script be written in Roman letters or Devanagari? + Use the writing system the host can read comfortably. Decide the caption writing system separately for the audience, and check both against the final recording. Can I translate an English video script directly? + A literal draft can help check meaning, but it may retain English sentence order or introduce a formal register the host does not use. Revise for speech while preserving factual claims and qualifications. Is the checkout example based on a client video? + No. It is an original fictional teaching exercise. The numbers are chosen to demonstrate the comparison and the adaptation process. ## Keep exploring. [Commission research and video scripts ↗ ](https://machinehouse.media/services/video-scriptwriting)[Build the argument before drafting ↗ ](https://machinehouse.media/guides/research-to-script)[Rebuild a long-form idea as a Short ↗ ](https://machinehouse.media/guides/clips-vs-original-shorts)[Check what changed in the rewrite ↗ ](https://machinehouse.media/guides/fact-checking-content) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/accessible-content-production [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Accessible content production # Plan for the audience before the edit is locked. A caption file cannot explain a graphic the host never describes. Use this worked production plan to assign captions, transcripts, visual description and the final player checks before a release reaches the publisher. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · September 18, 2026 ## What does “this one” refer to? A video shows two sample receipts. The host says, “This one is cheaper.” A viewer listening without the picture cannot know which order the host means. Captions reproduce the words, but they do not repair that missing explanation. Original fictional teaching example “The first order totals 244 rupees. The second totals 219, so the second order is 25 rupees cheaper.” That narration gives the listener the comparison. The graphic can still help someone follow the arithmetic. The [Hinglish example ](https://machinehouse.media/guides/hinglish-scriptwriting) shows how the same fictional prices can be written for another spoken register. ## Name the versions the audience needs. Begin with the media type: a recorded audio interview, a video with essential graphics and a live conversation need different production arrangements. Record the languages, publication platforms and audience requirements in the brief. [W3C’s media guidance ](https://www.w3.org/WAI/media/av/) distinguishes captions, transcripts, description of visual information, sign language and accessible media players. Captions carry speech and meaningful non-speech audio. A descriptive transcript can also explain important visual content. These are related deliverables with different jobs. ## Put essential information in the recording plan. Give the writer the intended graphics before the shoot. In the receipt example, they need both totals and the 25-rupee difference. The host can say them naturally while the editor shows the comparison. Other visuals may need a separate description treatment. Decide that with the actual content and audience in view. [W3C’s planning resource ](https://www.w3.org/WAI/media/av/planning/) provides checklists by media type and explains how requirements relate to WCAG levels. Use those checklists to set the project’s requirements. ## Assign the final files to people. - Writer and producer: identify essential visual information and agree how it will be communicated. - Caption editor: correct words, speaker identification and meaningful sounds against the final recording, then check timing and readability. - Transcript editor: work from the released version and include the visual descriptions agreed for the descriptive transcript. - Publisher: check the actual uploaded tracks, transcript links and player controls on each destination. If an edit changes 244 to 249, the correction has several homes. Assign a reviewer to reconcile the spoken amount, graphic, captions and transcript instead of treating each export as finished in isolation. ## Review the experience that will be published. The [editable production plan ](https://machinehouse.media/resources/accessible-content-production/production-plan.txt) includes the worked receipt example, owners, acceptance checks and a blank sign-off record. For the receipt example, listen without the picture: can you identify both totals and the difference? Then inspect the captioned version without sound. Test keyboard operation and transcript access on the player people will actually use. Those checks find practical defects; they do not certify a whole site or film. Record unresolved requirements and bring in specialist review where the team needs it. A later recut or different platform export can require another check. ## Questions about the method. Are captions and a transcript the same deliverable? + No. Captions are synchronized with the media. A transcript is a separate text version; a descriptive transcript also includes relevant visual information. Scope each against the actual audience and media. Can automatic captions be the final version? + Treat automatic output as a draft to check against the actual recording. Review names, numbers, speaker identification, meaningful sounds and timing before publication. Does this worksheet establish WCAG compliance? + No. It helps commission and review production work. Use the applicable W3C requirements and an appropriate accessibility assessment for the actual media and publishing environment. ## Keep exploring. [Brief the editor with clear obligations ↗ ](https://machinehouse.media/guides/video-editor-brief)[Plan the recording ↗ ](https://machinehouse.media/guides/content-shoot-plan)[Commission production and finishing ↗ ](https://machinehouse.media/services/content-production) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-catalogue-review [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Content catalogue review # Your next useful release may already be in the catalogue. An old recording can still answer a current question. Review recent demand, outdated advice and missing follow-through before deciding what to keep, update or make again. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · September 18, 2026 ## Six existing videos. Six different decisions. This is an original fictional catalogue, reviewed on 18 September 2026. Every item was published before the recent reporting window: 21 August–17 September, inclusive. The counts below are illustrative views within those same 28 days. | Video | Recent views | What the review finds | Next action | Start your first interview | 800 | Advice still current | Keep; link to the research guide | Our recording setup | 6,200 | Unavailable equipment and broken links | Correct the recommendations | Questions beyond the bio | 4,100 | Useful answer, generic promise | Investigate packaging | Summer workshop | 70 | Expired booking call to action | Close the sales path; assess archive value | Approving a cut | 2,700 | Viewers ask for a checklist | Add the missing resource | A guest cancels | 1,200 | Overlaps another current guide | Compare answers before consolidating The [full review sheet ](https://machinehouse.media/resources/content-catalogue-review/review-sheet.txt) also includes publication dates, lifetime counts and an action record. “Start your first interview” has the largest lifetime total. “Our recording setup” is the urgent job because people are still reaching advice that needs correction. ## Choose a window that answers your question. For maintenance, ask what the existing library is doing now. Record the same recent period for each item, then inspect the traffic sources and any unusual promotion. A new item that existed for only part of the window needs a separate comparison. For a question about launch performance, compare equal ages after publication instead. A three-year lifetime total and a three-week total describe different opportunities to accumulate views. Keep the metric definition and export source beside the numbers, particularly when comparing platforms or formats. ## Separate the repairs from the experiments. Broken links and expired advice have a direct acceptance check: the corrected destination works and the recommendation is current. Packaging changes need a hypothesis. In the third example, first inspect impressions, click-through rate and traffic mix before deciding whether the promise is the problem. If the content itself no longer answers the question, a new title will not repair it. Give each item one primary action and a reason. A useful video with a missing checklist needs a resource. An obsolete event page may need an archive notice. Two overlapping pieces need a comparison of their actual answers before either is removed from navigation. ## Write the action so someone can finish it. For the recording-setup video, assign a producer to verify replacement recommendations and inspect links. Record where the correction can be made, whether an on-page note is adequate, and whether the central advice needs a new recording. Check references and embeds before changing availability. The worksheet has separate fields for the owner, exact version, approval and follow-up. Keep a record of what changed. Later traffic can inform the next review, but a before-and-after increase alone does not establish what caused it. ## Questions about the method. How often should we audit old content? + Set a review interval around how quickly the advice changes and how much the catalogue is used. Time-sensitive recommendations and expired offers need earlier checks than stable teaching material. Assign dates and owners rather than relying on an occasional cleanup. Should we delete videos with few views? + A low count alone does not settle usefulness. Check the audience need, accuracy, incoming references and whether a small number of qualified viewers matters. Archive, unlisting and removal are separate decisions with different effects. Can this template be used for a podcast or article library? + Yes. Keep the decision fields and use the relevant medium’s metrics, links and update options. Do not treat a download, a video view and an article visit as the same measure. ## Keep exploring. [Review a completed release ↗ ](https://machinehouse.media/guides/content-postmortem)[Connect content to business outcomes ↗ ](https://machinehouse.media/guides/content-business-impact)[Commission a channel audit ↗ ](https://machinehouse.media/services/youtube-channel-audit) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/evaluating-ai-content-systems [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Evaluating AI content systems # Before you trust the workflow, give it work it can fail. A polished brief can contain a false quote, miss a required question or hide a source conflict. Evaluate the whole job against explicit acceptance checks, then count the human work needed to finish it. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · September 18, 2026 ## A good sentence can still fail the assignment. Original constructed test, not a model result Source: We cancelled the Monday recording because the guest was ill. The replacement session happened on Thursday. Task: Give one exact quote explaining the cancellation. Failing answer: “Production delays forced us to reschedule Monday’s recording.” The answer invents a reason and puts words in quotation marks that the source never said. Smooth prose does not offset either error. A passing response could quote: “We cancelled the Monday recording because the guest was ill.” An accurate paraphrase would still fail this particular task if presented as a verbatim quote. Grade against the requested job as well as the underlying facts. ## Build a small set with different failure modes. The [downloadable test cases ](https://machinehouse.media/resources/evaluating-ai-content-systems/test-cases.json) contain five original exercises: exact quotation, receipt arithmetic, conflicting schedule records, a required source gap, and an instruction embedded inside research material. Each includes the input and explicit checks. These are teaching fixtures, not reported benchmark results. Run the cases through the same steps your team will actually use. If the system retrieves documents, test retrieval as well as the final draft. If a producer has already selected every source and written the answer into the prompt, record that preparation as part of the workflow. ## Use a check that fits the failure. Exact amounts and required fields can often be checked directly. Editorial usefulness needs judgment: does the answer help the next person make the intended decision? A quote checker can find matching text, but a reviewer still needs to assess attribution and context. [Anthropic’s evaluation guide ](https://www.anthropic.com/engineering/demystifying-evals-for-ai-agents) distinguishes tasks, repeated trials and graders, and explains the roles of code, model and human review. For these content exercises, record factual checks separately from the editorial assessment. Calibrate any model-based judge against examples a human has reviewed. Decide which errors block use before looking at results. A fabricated quotation should not disappear into an average score because the structure and tone were good. ## Keep a record you can compare later. Save the source bundle, prompt, system version, raw output and review decision for each attempt. Repeat tasks to see whether failures recur. Use the same inputs and review criteria when comparing a change, while recording model or retrieval changes that prevent a clean comparison. Keep some real tasks aside from prompt development. Once the team has tuned against an example, it is useful as a regression check but weaker evidence of performance on unfamiliar work. The five public fixtures are a starting point; passing them alone does not establish readiness for your client material. ## Count the work after the first answer. The [review sheet ](https://machinehouse.media/resources/evaluating-ai-content-systems/review-sheet.txt) separates input preparation, human checking, repairs, machine elapsed time and direct tool costs. A two-minute generation followed by an hour of correction is a different operating result from a usable brief in two minutes. Finish the evaluation with a specific decision: usable for this task with named review, revise and retest, or unsuitable for the current use. Record the remaining failure and who owns it. Widen the scope only when the additional kinds of work have their own evidence. ## Questions about the method. How many examples do we need? + Start with actual jobs and the failures that would make their output unusable. These five exercises teach the method; they are not a statistically sufficient sample for every workflow. Expand coverage as you find new task types and failure modes. Can another AI grade the writing? + It can assist with a clear rubric, but inspect its agreement with human judgments and preserve the source evidence. Keep mechanical checks for facts that can be checked directly, and do not let a style score cancel a factual failure. Does this evaluate a model or a whole system? + The intended unit is the workflow your team uses: inputs, retrieval, instructions, model, tools and human review. Record the components so a change can be traced and the result can be interpreted. ## Keep exploring. [Design the underlying workflow ↗ ](https://machinehouse.media/guides/ai-content-workflows)[Check claims before publication ↗ ](https://machinehouse.media/guides/fact-checking-content)[Build an AI-assisted content system ↗ ](https://machinehouse.media/services/ai-content-systems)[Quote checker ↗ ](https://machinehouse.media/labs/quote-checker) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/content-team-knowledge-system [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ A knowledge system for a content team # Leave the next person more than a folder. A team needs to know which decision is current, why it was made and who can change it. Use a worked decision record to carry that context from research into scripts, graphics and the next edit. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · September 18, 2026 ## Two drafts. One missing decision. A writer finds “the first order is cheaper at 199 rupees” in an old draft. The editor has a newer script saying the second order is cheaper at 219. Both files sit in the same folder. A search result that finds both has located the disagreement without resolving it. In this fictional teaching example, the first draft left out delivery and handling. The approved comparison uses final totals: 244 and 219 rupees, a difference of 25. The team needs that reason as well as the latest sentence. ## Make the decision retrievable with its evidence. Original fictional decision record · RECEIPT-01 Decision: Compare full order totals. Evidence: First order: 199 + 35 + 10 = 244. Second order: 219 all-in. All prices are invented for the exercise. Reason: Comparing the first item price with the second final total omits charges from only one side. Status: Approved teaching version 2. Version 1 is superseded. Owner: Producer. Recheck whenever the receipts or comparison premise change. The [complete editable record ](https://machinehouse.media/resources/content-team-knowledge-system/decision-record.txt) includes source location, approved wording, the rejected alternative and affected deliverables. In a live project, replace role placeholders with named owners and link the actual approval record. ## Give different kinds of material different jobs. Keep raw sources available for checking. Put approved decisions in a clear home and link them from the working brief. Keep reusable methods separate from facts that apply to only one project. A good interview question can travel further than a confidential guest detail. A record should identify the source version or retrieval date, the person responsible and the next review trigger. A folder called “final” cannot explain whether advice is still current. Neither can a recent file timestamp establish that somebody approved a change. ## Carry a correction to the files that use it. If the receipt totals change, updating the decision record is only the first step. The script, graphic, captions, descriptive transcript and short adaptation may all use those numbers. The record names those dependencies so the producer can assign the affected revisions. Retain the earlier decision as history, mark it superseded and point it to the current version. If the team uses AI search, test that a question about current guidance returns the approved record with its source and status. Include conflicting and superseded records in that test. If sources disagree and the owner has not resolved them, mark the decision unresolved. Do not let the newest filename silently become the approved answer. ## Test the handoff with someone new to the job. Give a new editor the record and ask which order is cheaper, why the old line fails and which files need review after a correction. If they still need the original writer to explain the choice, add the missing context to the record. Set access and reuse rules before inviting people or connecting retrieval tools. A shared search interface should preserve the underlying permission boundaries. The downloadable example is public teaching material; a real client brief may require a much narrower audience. ## Questions about the method. Do we need a new knowledge-management app? + Not necessarily. A shared document or existing workspace can hold the record if the team can find its canonical location, identify the approved version and manage access. Choose software after identifying what the handoff is missing. What should be kept from a finished episode? + Keep the source evidence, consequential editorial decisions, approved reusable methods and relevant outcome notes. Preserve the permission and date context. Storing every draft without a current-version rule can make the next handoff harder. Can AI search replace maintaining the records? + Search can help locate material, but someone still owns approval, corrections and access. Test retrieval with conflicting versions and verify that the answer identifies the current record and any unresolved gap. ## Keep exploring. [Assign approval responsibilities ↗ ](https://machinehouse.media/guides/content-approval-workflow)[Test the retrieval and drafting workflow ↗ ](https://machinehouse.media/guides/evaluating-ai-content-systems)[Build connected content systems ↗ ](https://machinehouse.media/services/ai-content-systems) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/labs/content-capacity-planner [Home ](https://machinehouse.media/)/ [Labs ](https://machinehouse.media/labs)/ Content capacity planner # How much can your content team finish? Enter each person’s available hours and their total work per release. Find the weekly capacity, the person limiting it and the time left after a buffer. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Tool · September 18, 2026 ## Plan one repeatable release. Start with one unit: an episode, an article or a defined release package. Use one row per person and combine all their tasks. The starting values are an illustrative team, not a recommended staffing ratio. Reserve as buffer (%) Enter hours available to this project after other commitments. The buffer is reserved from those hours. Add person Calculate capacity This calculator needs JavaScript. The method and worked calculation below can also be used manually. The estimate appears here when the calculator runs. Inputs stay in this page and are not saved by the tool. Copy your assumptions into the project plan before leaving. ## Calculate the capacity for each person. For each person, multiply available weekly hours by the share left after the buffer. Divide that by their hours per release. The smallest result limits the team; round it down to complete releases. In the starting example, the editor has 24 hours and a 20% buffer, leaving 19.2 hours. At eight hours per release, that is 2.4 releases. The writer can support 3.2 and the reviewer four, so the estimate is two complete releases per week . The editor has 3.2 hours left after those two releases. A person doing both research and writing belongs in one row with the combined time. Listing them twice as two independent roles would overstate the hours available. ## Check the assumptions before making a promise. This estimates steady weekly capacity. It does not schedule dependencies, waiting for approval, holidays or the elapsed time to finish the first release. Those belong in the actual production plan. An estimate of zero complete releases means the entered hours do not finish a whole release within one week; work may carry across weeks. Use measured working time where you have it. Include the usual review and revision effort in hours per release; leave the buffer for uncertainty. A mixed slate needs a defined package or a separate resource plan, because a podcast and a short do not necessarily consume the same work. The [assumptions worksheet ](https://machinehouse.media/resources/content-capacity-planner/assumptions.txt) preserves the release definition, people, estimate and unresolved dependencies for the next planning conversation. ## Questions about the work. What does the buffer do? + It reserves a percentage of each person’s available project hours before calculating capacity. The 20% starting value is an editable example, not a benchmark for every team. Why can spare hours remain when the team is at capacity? + Work is constrained by the limiting person and rounded to complete releases. Spare writing time cannot automatically replace editing time. Use the result to discuss task allocation, scope or the next release. Can I plan several formats at once? + Define a repeatable package, such as one episode plus two original shorts, and enter each person’s combined hours for that package. For an irregular mix, use a resource schedule that assigns hours to each actual job. ## Keep exploring. [Account for review and approval ↗ ](https://machinehouse.media/guides/content-approval-workflow)[Plan the actual recording ↗ ](https://machinehouse.media/guides/content-shoot-plan)[Scope production costs ↗ ](https://machinehouse.media/guides/content-production-costs) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/labs/quote-checker [Home ](https://machinehouse.media/)/ [Labs ](https://machinehouse.media/labs)/ Quote checker # Does the quote appear in your source? Paste a transcript or document and the quotation you want to check. Find the matching words and read them in context before using them in a draft. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Tool · September 18, 2026 ## Compare the quote with the supplied text. Source text Quote to check Allow whitespace, curly quotes and equivalent Unicode characters to match Exact matching is tried first. Normalization keeps capitalization, words, numbers and other punctuation unchanged. Ellipses are treated literally. Check quote Clear text This checker needs JavaScript. You can also use your document’s Find command and compare the surrounding passage manually. Paste source text and a quote to begin. The tool processes your text in this browser. It does not send the pasted text for analysis or save it. Up to 200,000 source characters and 20,000 quote characters. ## A plausible quote can change the reason. Original constructed example Source: We cancelled the Monday recording because the guest was ill. Draft: Production delays forced us to reschedule Monday’s recording. The draft sentence is absent from the source and substitutes a different reason. Try both versions in the checker. A matching passage establishes that those characters occur in the supplied text; it does not establish that the text itself is accurate. ## Read the result with the source still open. Exact text match means the same character sequence occurs in the text you supplied. Typography-normalized match means a match appeared only after the stated normalization. No matching text found means neither permitted comparison located it. Check the speaker, surrounding qualification and completeness of the excerpt. A phrase inside a denial, a question or someone else’s quoted words can match perfectly and still be misleading when attributed to the speaker. A missing match can also come from an incomplete transcript, a correction or an excerpt assembled with omissions. The tool does not infer those changes. Check the original recording or document before treating the draft as wrong or the quotation as usable. Use the [quote review record ](https://machinehouse.media/resources/quote-checker/review-record.txt) to save the source version, timecode or passage location, attribution check and approved excerpt. ## Questions about the work. Is this a fact checker? + No. It locates matching text in the material you paste. It does not verify the source, identify the speaker, assess whether a claim is true or decide whether the excerpt preserves context. Does it search the web or listen to audio? + No. Supply the relevant text yourself. If a transcript may be wrong, check the original recording. Nothing is retrieved from a URL or external database by this tool. Can it match a quotation with an ellipsis? + Only if those exact characters appear in the source. Check each continuous excerpt separately, then review the omitted passage and attribution manually. ## Keep exploring. [Review claims and attribution ↗ ](https://machinehouse.media/guides/fact-checking-content)[Test the full AI workflow ↗ ](https://machinehouse.media/guides/evaluating-ai-content-systems)[Prepare a sourced interview brief ↗ ](https://machinehouse.media/guides/interview-research-brief) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/writing/before-the-proposal [Home ](https://machinehouse.media/)/ [Writing ](https://machinehouse.media/writing) WRITING · 12 JAN 2026 · 2 MIN READ # Before the proposal What a client should be able to see before deciding to work together. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet) Founder, Machine House Media [Originally on LinkedIn → ](https://www.linkedin.com/posts/rishwajeet-singh_i-cant-ask-for-money-until-ive-already-activity-7416411126248153088-DXoQ) I can't ask for money until I've already solved the problem in my head. Apparently that's weird. Most consultants figure out what to do after you've paid them. That's not a criticism. That's just how it works. You see their credentials, their case studies, their client logos. You hop on a call. They ask about your goals, your challenges, your budget. They sound smart. You sign. Then they start thinking about your actual problem. The real work happens after you've committed. ## The absurd thing Mentioned my process to a more experienced content strategist recently. He's good — been doing this longer than me, bigger clients, knows the game. I told him I build direction before sending proposals. Before calls, sometimes. Before anyone's paid anything. His response: "Why would you do that? I don't do any real work until they've signed." He's not wrong. That's the standard approach. But I genuinely didn't know you could do it that way. ## How I think I can't ask someone for money unless I already know what I'd do with it. Not vague confidence. Actual clarity. For a YouTube client, I've already looked at their channel — titles, thumbnails, scripts, the niche, where they sit in it. All public data. I've figured out where it needs to go. I can see the direction. Only then can I quote a price. Some consultants sell on credentials and figure it out once they're inside. I can't. My brain doesn't let me ask for commitment without having vision first. ## Why it works for me When I've already done the thinking, the conversation changes. I'm not pitching credentials. I'm showing direction — here's where you are, here's where you need to go, here's why. We're talking about vision, not whether I'm qualified to find one. I want you to be confident about what we'll do together. ## Know what you're buying My friend's way is legitimate. I've hired people on credentials myself. Track record matters, of course. But if you're ever hiring a consultant — or any expert — it helps to know what you're actually getting. Some sell on credentials. You're trusting their past work predicts what they'll do for you. Some sell on vision. They've already thought about your specific problem. You can see their thinking before you commit. Neither is wrong. They're different purchases. Credentials say "I've solved problems like this before." Vision says "I've already started solving yours." Both can work. Just know which one you're buying. This is how we think at Machine House. If you want it applied to your content — [Book a call ](https://calendly.com/rishwajeet/intro-call) [Email instead ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/choosing-youtube-growth-agency [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Choosing a YouTube growth agency # How to choose a YouTube growth agency. Examine the diagnosis, check the claimed results and compare what each team will own. A worked proposal comparison and an editable brief for founders and marketing teams. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · October 13, 2026 ## Who chooses what the four videos are about? The proposal promises four videos a month, eight Shorts and a report. Your editor can already make those. The difficult part is deciding what the founder should say next Tuesday. Start there. Write down the decision or task your current team cannot handle. If the subjects are strong but the edits keep arriving late, you may need production capacity. If nobody can explain who the videos are for, adding another editor leaves the editorial problem with you. “YouTube growth agency” can describe quite different engagements. Ask which of these responsibilities the team will take: - Strategy: choosing an audience, developing subjects and deciding what the next releases should test. - Production: researching, writing, recording and editing the agreed videos. - Channel management: coordinating the release calendar, packaging, uploads and the review that informs the next slate. - Paid promotion: running an advertising campaign with an agreed budget and reporting its results separately. Give the missing responsibility a named owner in the proposal. ## Ask them to explain one of your videos. Before buying an ongoing retainer, ask how the agency would investigate the channel. A public review can examine the videos and their packaging. A deeper diagnosis may need authorised Analytics access and a paid audit. Ask to discuss one video before commissioning the deeper review. Choose a recent video. Ask what the title promises, where the opening delivers that promise, and which next decision they would change. A useful answer should point to something you can inspect. Invented review example The title promises the cost of opening a café. The first two minutes describe the founder’s childhood. The agency proposes starting with the lease deposit and fit-out bill, then bringing in the backstory when it explains a spending decision. That is a specific editorial suggestion. To claim the opening caused viewers to leave, the reviewer would also need the retention curve and viewing context. The next cut can test the idea; the public view count alone cannot settle the cause. Watch for advice that survives every possible channel: post more often, improve thumbnails, make the hook stronger. For the cost breakdown, would the new opening preserve enough context to understand the numbers? ## Follow one claimed result back to the work. A channel grew while an agency worked on it. What did the agency actually do? Ask for one relevant case, including a weaker release, and walk through the decisions. A team that wrote the scripts should be able to show how a subject became a script; a team responsible for packaging should explain the title and thumbnail choices. - Starting point and dates: what period is being compared, and how old were the videos when measured? - The set of releases: is the claim about the full engagement or its biggest hit? What happened to the other videos? - Distribution: how much came through paid promotion, and which formats or traffic sources are being combined? - Contribution: who chose the topics, appeared on camera and supplied the production team? What changed because of this agency? Redacted working documents or a client reference can help where the team cannot share private Analytics. Treat a public screenshot as evidence of the number it shows, with the cause still to be established. Be careful with a promised click-through rate. YouTube says CTR varies with the audience and where impressions appear. A percentage from a small familiar audience is not a universal target for a video reaching new viewers. [YouTube’s CTR guidance explains the comparison limits. ](https://support.google.com/youtube/answer/7628154?hl=en) Paid views also need their own explanation. YouTube describes Promote as advertising and says promotions do not affect how videos are recommended organically. Ask for the media budget and paid results separately from the agency fee and organic performance. Paid distribution may serve a useful campaign objective; it does not by itself demonstrate stronger organic discovery. [Read YouTube’s Promote guidance. ](https://support.google.com/youtube/answer/12199105?hl=en) ## Two proposals for the same founder channel. This is a fictional buying exercise. A founder can record on one day each month. The marketing manager handles product launches and can consolidate feedback, but cannot research and write four episodes. Both proposals below offer four long videos and eight Shorts. Prices are omitted so you can first see what the team would be buying. | Decision | Proposal A: production support | Proposal B: editorial and production | Topics | Client supplies four approved topics. | Agency proposes a slate with audience and source rationale; founder selects it. | Before recording | Client supplies scripts and arranges the shoot. | Named writer researches and drafts; agency prepares a recording plan for the agreed shoot setup. | Edit and packaging | Edits supplied footage and designs one thumbnail from the client’s direction. | Develops the cut, title directions and thumbnail concepts; production and revision limits are specified. | Eight Shorts | Client selects eight excerpts for vertical edits. | Agency selects suitable excerpts; any newly written or recorded Shorts are separately specified. | Monthly review | Reports publishing and performance figures. | Reviews the releases and brings a proposed change to the next slate. | Work left with the client | Editorial planning, scripts, recording and clip selection. | Founder expertise, factual approval, recording availability and consolidated feedback. For this founder, Proposal A leaves the missing writing capacity unresolved. It could still work if the business appoints an editorial lead or buys writing separately. Compare that combined cost and the manager’s coordination time with Proposal B. Proposal B needs checking too. Meet the person doing the writing, inspect a relevant script and ask how many other accounts that person handles. Establish who attends the recording and what happens if the founder’s day moves. “Full service” is only useful once the people and commitments behind it are clear. Then compare the commercial terms on equal scope: filming and travel, media spend, revisions, extra formats, taxes where applicable and the handover files. The [production-cost guide ](https://machinehouse.media/guides/content-production-costs) explains those cost drivers. For an India-based shoot, state the city, working language and travel assumptions; a Hindi interview and an English scripted explainer may need different writers and review arrangements. ## Decide what the first review must answer. Agree an initial body of work and a review date before signing. For the fictional four-video scope, the first review could check whether the team delivered the agreed research and scripts, whether recording fitted the founder’s available day, and which editorial decisions should change for the next batch. Choose the performance comparisons in advance too. Review long videos separately from Shorts, use comparable publication ages, and keep paid exposure visible. If qualified enquiries are the business goal, record those alongside viewing behaviour. Four releases may leave the growth question unresolved. A team can still explain what it observed, what it thinks happened and what it proposes to try next. Ask what evidence would make it change direction. The answer should leave room for a weak idea to be dropped. At the review, make a concrete commissioning decision: keep the format, revise the opening, change the subject, reduce the workload or investigate further. The [content postmortem guide ](https://machinehouse.media/guides/content-postmortem) has a worked example of comparing releases. ## Put the working arrangement in the brief. Give each shortlisted agency the same channel background, audience, business objective and practical constraints. Include the people you already have and the decisions you want the agency to own. Our [editable YouTube agency brief ](https://machinehouse.media/resources/youtube-agency-brief/agency-brief.txt) includes space for their diagnosis, evidence and proposed first review. Agree access according to the job. YouTube’s channel permissions allow access without sharing your Google Account credentials. Viewer roles support reviewing channel details; publishing needs a role with the appropriate editing permissions. Some roles can see revenue data, while limited roles cannot. Check the current [YouTube permission table ](https://support.google.com/youtube/answer/9481328?hl=en) before granting access. Also specify what you receive when the engagement ends: final masters, editable project files where agreed, thumbnails, approved scripts, source and licensing records, and the current release plan. Name who transfers the files and who removes channel access. Machine House offers [focused channel audits ](https://machinehouse.media/services/youtube-channel-audit) and ongoing [YouTube strategy and channel work ](https://machinehouse.media/services/youtube-growth). Full-service engagements run through The Breakout Company, co-founded by Rishwajeet. You can bring the same brief to a conversation with us. Before accepting the proposal, fill in the blank beside “Who brings the next four episode ideas?” ## Questions about the method. What should a YouTube growth agency include? + The scope should identify which team owns audience strategy, topics, research, scripts, recording, editing, packaging, publishing and performance review. Some agencies cover selected stages. Compare those responsibilities with the work your own team can supply. Should I hire a YouTube agency or a video editor? + An editor can be the right hire when you have strong subjects, usable footage and someone who can give a clear editorial brief. If the missing work is choosing topics or writing episodes, agree who will do that before increasing editing capacity. Can an agency guarantee YouTube growth in 90 days? + A proposal can commit to work delivered, a release schedule and review decisions. A fixed view or subscriber outcome needs scrutiny: ask what assumptions support it, whether paid promotion is involved and what happens if the target is missed. An initial review period is not a guarantee of audience response. Do I need to share my YouTube password with an agency? + Use YouTube channel permissions to grant the access the work needs without sharing your Google Account credentials. Review each role's capabilities and remove access when the engagement ends. ## Keep exploring. [YouTube growth and channel management ↗ ](https://machinehouse.media/services/youtube-growth)[Begin with a channel audit ↗ ](https://machinehouse.media/services/youtube-channel-audit)[Compare content production costs ↗ ](https://machinehouse.media/guides/content-production-costs)[Measure enquiries beyond the view count ↗ ](https://machinehouse.media/guides/content-business-impact)[Rishwajeet: before the proposal ↗ ](https://machinehouse.media/writing/before-the-proposal) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media) --- Source: https://machinehouse.media/guides/video-podcast-launch-plan [Home ](https://machinehouse.media/)/ [Guides ](https://machinehouse.media/guides)/ Video podcast launch plan # How to launch a video podcast. Work backwards from the first release: test the show, finish a usable episode buffer and give every handoff an owner. A worked schedule and an editable launch plan. [Rishwajeet Singh ](https://machinehouse.media/rishwajeet)Guide · October 15, 2026 ## Three recordings. One finished episode. To launch a video podcast, define what the audience comes for, test it in a pilot, finish the first releases and set up the places people will watch or listen. Choose the public launch date after you know how long those steps take. Imagine the first shoot has gone well. Three guests recorded. The first episode is approved, the second needs an edit and the third contains a number the guest still has to check. Announce a weekly show now and you have one publishable episode, with the next two dependent on unfinished work. The plan below follows a fictional interview show through that problem. It uses a fortnightly schedule because that is what this small team can support. Adapt the decisions to your own recording access, edit capacity and approval process. [Download the editable launch plan ](https://machinehouse.media/resources/podcast-launch-plan/launch-plan.txt) for the completed example and a blank worksheet. ## Give the first three episodes different work to do. The fictional show is called After Closing . Its audience is people running small shops. Each interview reconstructs one decision: what the owner tried, what it cost and what they changed afterwards. Fictional first slate - The bakery: why the owner stopped accepting delivery orders during the morning rush. - The hardware shop: how the owner decides which rarely requested spare parts deserve shelf space. - The salon: what happened when a weekday discount filled appointments but left less money at closing. These are proposed conversations, not stories already secured. The producer must establish whether each guest can reconstruct the decision and discuss the relevant details. A famous shop owner who only wants to give general business advice may be wrong for this show. The host can use the same basic sequence across all three: the old arrangement, the incident that exposed a problem, the decision and its consequences. The [show-format guide ](https://machinehouse.media/guides/show-format-development) goes deeper into building a repeatable premise. For this launch, the test is whether three different guests can deliver it. ## Take the pilot all the way through the edit. Record a pilot before booking a full batch. It can remain unpublished. Use a willing participant with a real decision to discuss, in the room and recording setup you intend to use. For the private pilot, the host talks to a shop owner who stopped opening on Sundays. Asked what happened to sales, the participant answers, “It worked out better.” The producer needs a follow-up: better in which way, over what period, and compared with what? The pilot has exposed a research gap. The next guest brief asks for a specific before-and-after account, including what the guest cannot substantiate. The host also practises asking for the missing detail during the conversation. Then finish the pilot edit. Listen to each microphone, check that the cameras stay in sync and watch the exported file. Listen once without looking at the screen: if a gesture or chart carries the explanation, add the spoken context or adapt the audio edition. Record the actual edit time and how long feedback took. Use those observations to revise the scope. If the pilot consumes the editor’s whole week, three interviews plus a trailer and twelve clips will need more capacity, fewer deliverables or a slower release pace. The [production-cost guide ](https://machinehouse.media/guides/content-production-costs) helps specify that tradeoff before commissioning the batch. ## Build the calendar around finished releases. For After Closing , the host handles the conversations and editorial approval. The producer researches guests, coordinates feedback and publishes. The editor prepares the video and audio editions. These are responsibilities; one person may hold more than one if their calendar allows it. The following dates are a worked planning example, not a promised production turnaround. Day 0 is the first full episode. The team chooses a fortnightly release and requires two approved episode packages before announcing it. The pilot is a separate exercise and does not count towards those releases. | When | Work and owner | What must be ready | Before Day −21 | Host, producer and editor finish the pilot and revise the format. | Usable recording setup, tested handoff and a realistic edit allowance. | Day −21 | Producer runs the first shoot: bakery, hardware shop and salon. | Research complete, permissions agreed and copied media checked before the team disperses. | Day −14 to −7 | Editor finishes the first two episodes; host approves; producer sets up publishing. | Bakery and hardware packages approved by Day −7. Feed submitted early with a short trailer; platform status checked. | Day 0 | Producer releases the bakery episode and checks the published versions. | Hardware episode ready for Day 14; salon still awaiting a factual check. | Days 1–21 | Producer releases hardware on Day 14, resolves the salon query and runs the next shoot on Day 21. | Salon package approved by Day 21 for Day 28. Next guests and recording time already confirmed. | Day 28 | Producer releases salon; editor continues the next batch. | The following episode has an owner, review deadline and release slot. Suppose the salon’s factual check arrives on Day 18. A weekly plan using these three episodes would have missed its third release on Day 14. The fortnightly plan leaves time to correct and approve the cut before Day 28. The recording count was identical; the publishing commitment was different. If the first two packages are not ready on Day −7, move the announcement. If a later episode slips after the schedule is public, use an already approved substitute or communicate the revised date. Decide who can make that call before the deadline arrives. ## Define what “ready” includes. An episode package for this show contains the final video, an audio edition, checked captions, title, thumbnail, description and the approval record. The producer also holds the source files and permission records. A finished timeline on the editor’s computer is one step earlier. Agree the review sequence before recording. The editor sends a first cut to the producer. The producer consolidates comments, obtains any agreed guest fact-check and sends a clear revision list. The host approves the final editorial version. The producer uploads that exact export and verifies the result. Put deadlines and review scope in the guest arrangement. A guest fact-check and full editorial approval are different commitments; decide which you are offering. If a response is late, follow the agreed process rather than treating silence as consent. Keep the launch extras small enough to finish. In this example, each release gets one excerpt and a short written introduction. Those can point to a real episode once its destination is live. Expand the package when the team can reliably deliver it. Use the [distribution-plan guide ](https://machinehouse.media/guides/content-distribution-plan) to assign each companion asset a job. ## Check video and audio destinations separately. The example chooses full video on YouTube and Spotify, with an audio RSS feed for Apple Podcasts. An RSS feed carries episode information and media links to podcast apps. The producer keeps a record of the hosting account, feed URL and each public show URL. Platform guidance checked October 15, 2026. Choose your hosting arrangement before following a particular upload workflow. - YouTube: create a podcast playlist for the full episodes and keep clips in separate playlists. Upload the video episodes and check their ordering. YouTube describes a podcast as a playlist whose episodes are videos; designation does not guarantee access to every podcast feature. [YouTube’s podcast setup instructions. ](https://support.google.com/youtube/answer/12751636?hl=en) - Spotify-hosted route: publishing through Spotify makes the show available there, but the producer must submit its RSS feed to other services. Spotify says video from this route stays on Spotify while other destinations receive audio. Check those other listings yourself. [Spotify’s distribution guidance. ](https://support.spotify.com/us/creators/article/distributing-your-show-to-other-platforms/) - Apple Podcasts: submit the chosen feed and allow for technical validation and review. Check the show’s availability settings and time zone. This example supplies audio; Apple also supports video through supported delivery arrangements, so check your host’s capabilities if you want video there. [Apple’s submission process ](https://podcasters.apple.com/support/897-submit-a-show) and [video guidance. ](https://podcasters.apple.com/support/3684-video-podcasts) A short feed trailer can help you set up the listings before the first full episode. Apple recognises trailers as an episode type and requires at least one episode for a show to become available. [Apple’s episode guidance. ](https://podcasters.apple.com/support/825-how-to-create-an-episode) Choose what the trailer promises carefully; it should not commit you to a date the team has not approved. On release day, open each destination as a listener. Play the start and a later section, check that the right version is present, and test the links. If a listing is still under review, leave its link out of the launch message and say where the episode is available. ## Book the next recording before the buffer runs out. In this example, the producer confirms the next guests before launch for the Day 21 shoot. Waiting until all three episodes are public would leave only fourteen days to research, book, record, edit and approve the next one. Hold the first production review after the third release has had a week in public. Compare planned and actual work: where did feedback stall, what did the edit require, and can the team sustain the cadence? For audience response, compare equivalent publication windows and keep each platform’s metrics separate. Use the comments and viewing patterns to improve the next interviews; this first batch cannot establish a reliable growth trajectory. The [editable plan ](https://machinehouse.media/resources/podcast-launch-plan/launch-plan.txt) includes the fictional schedule, an episode-status worksheet and space for your own owners and deadlines. Machine House’s [podcast production work ](https://machinehouse.media/services/podcast-production) can cover format development, research and the recording-to-release process; agree which stages you need. Before you announce the first episode, put the next recording date in the calendar and name the person who will get the guests there. ## Questions about the method. How many episodes should I record before launching a podcast? + Choose the number from your release cadence and the time needed to produce the next batch. Count approved, publishable episodes separately from recordings. The worked example records three and requires two finished packages before announcing a fortnightly launch; that is a planning choice, not a universal minimum. Should I launch a podcast weekly or fortnightly? + Use the pilot to measure research, recording, editing and review time. Choose a cadence the available team can sustain, including delays and the next recording cycle. A weekly calendar needs episodes to become publishable at that pace. Can I publish the same video podcast on YouTube, Spotify and Apple? + You can plan for all three, but check the delivery route for each. YouTube uses video episodes in a podcast playlist. A Spotify-hosted show distributes audio through its external RSS route, while Apple also supports video through suitable hosting and delivery arrangements. Verify the published format on each destination. Does my podcast pilot have to be published? + No. A pilot can be a private production exercise that tests the conversation, recording setup, edit and approval process. Decide whether it earns a public release after reviewing the result. ## Keep exploring. [Develop the show format ↗ ](https://machinehouse.media/guides/show-format-development)[Prepare the interview research ↗ ](https://machinehouse.media/guides/interview-research-brief)[Plan the recording day ↗ ](https://machinehouse.media/guides/content-shoot-plan)[Set the review and approval process ↗ ](https://machinehouse.media/guides/content-approval-workflow)[Podcast production scope ↗ ](https://machinehouse.media/services/podcast-production) ## Tell us what you want to make. Tell us about the channel or idea, the team you have and where you need help. We’ll discuss the scope on a 30-minute call. [Book a 30-minute call ↗ ](https://calendly.com/rishwajeet/intro-call)[rishwajeet@machinehouse.media ](mailto:rishwajeet@machinehouse.media)