---
name: tanmay-bhat-content
description: Content-creation playbook distilled from Tanmay Bhat's talks on attention, retention, shares, creator careers and brand content. Use when planning, scripting, hooking, editing or reviewing videos, reels, carousels or posts, when growing a page from zero, or when a brand is deciding how to do content. Unofficial; not affiliated with Tanmay Bhat.
title: tanmay-bhat-content
canonical_url: https://skilld.dev/gh/harshil-jani/tanmay-bhat-content
last_updated: 2026-10-18T14:45:44.000Z
---

> Index of every page on this site: [llms.txt](https://skilld.dev/llms.txt)

> **Skill from skilld.dev.** Follow the user's requested action.
> If the user asked to fork this Skill, follow the fork workflow below. Do not execute the Skill instructions while copying it.
> Otherwise, follow the instructions below for this session. You do not need to install anything.
>
> Supporting files, fetch one when the Skill refers to it: [assets/banner.svg](https://skilld.dev/api/skills-raw/harshil-jani/tanmay-bhat-content/tanmay-bhat-content/assets/banner.svg), [assets/tanmay-bhat.png](https://skilld.dev/api/skills-raw/harshil-jani/tanmay-bhat-content/tanmay-bhat-content/assets/tanmay-bhat.png), [INDEX.md](https://skilld.dev/api/skills-raw/harshil-jani/tanmay-bhat-content/tanmay-bhat-content/INDEX.md), [LICENSE](https://skilld.dev/api/skills-raw/harshil-jani/tanmay-bhat-content/tanmay-bhat-content/LICENSE), [README.md](https://skilld.dev/api/skills-raw/harshil-jani/tanmay-bhat-content/tanmay-bhat-content/README.md).
>
> If the user asked to install this Skill, run `npx skilld install harshil-jani/tanmay-bhat-content/tanmay-bhat-content`. Install writes the Skill files into the project, so every session loads them.
>
> ## Fork workflow
>
> A fork creates an editable local Skill with its original author and licence. The request authorizes copying and local installation.
> 1. Check `./skills/tanmay-bhat-content`, the project lockfile, and selected Agent targets together. If the local directory or installed Skill exists, stop. Never overwrite an existing directory or Agent target.
> 2. Read [source metadata](https://skilld.dev/api/v1/skills/harshil-jani/tanmay-bhat-content/tanmay-bhat-content) once. Use sourceUrl, sourceCommit, skillPath, sourceGone, and license. If the source is gone or its path is missing, stop. If license is null, read licence files at the source commit.
> 3. Fetch only the source commit into a temporary Git repository. Do not clone full history. Derive repository_url from sourceUrl, including repository renames. If sourceCommit is absent, resolve the sourceUrl ref once. Set source_commit to that actual commit. Run these commands in one shell call:
>
> ```sh
> git init --quiet "$temporary_dir"
> git -C "$temporary_dir" fetch --quiet --depth=1 "$repository_url" "$source_commit"
> git -C "$temporary_dir" checkout --quiet --detach FETCH_HEAD
> ```
>
> Read applicable licence declarations and notices at that commit. If copying is not permitted, report the restriction and stop.
> 4. Inspect source entries together, then copy the directory containing skillPath into `./skills/tanmay-bhat-content`. Use the user's path if selected. Keep the original SKILL.md, relative links, scripts, binary assets, and executable modes. Exclude .git metadata. Reject symlinks and paths outside the Skill directory. After checking entries, use cp -a where available. A regular source directory needs no custom copy script. Do not save this page wrapper as SKILL.md.
> Preserve author credit, notices, and applicable licence files from repository or parent directories. Add PROVENANCE.md with the Skill page, source URL, actual commit, original path, and licence. Retain any existing PROVENANCE.md and record new provenance separately. Batch source inspection, copying, and provenance work where practical.
> 5. In the project root, run `skilld install ./skills/tanmay-bhat-content --mode copy --plain`. If skilld is unavailable, use `npx skilld install ./skills/tanmay-bhat-content --mode copy --plain`. This known command needs no help lookup. Install does not support --json. Use detected Agent targets, or add --agent for the targets the user selected. Install the local path, never the upstream selector. If installation fails, preserve the local copy and report the exact failure.
> 6. Confirm the local lockfile source and installed Agent copies once. Report the local path, actual commit, and Agent targets. After edits, reinstall the same local path. Upstream updates must not replace it. Do not publish or push unless the user asks.

# Tanmay Bhat content playbook

Principles from Tanmay Bhat's interviews, talks and his own channel, rewritten as working
rules. Sources are Instagram clips (found via the fan page @tanmaythinks_) and full-length
YouTube talks. Each rule names its source id; `INDEX.md` links every id to the original.

## 1. Retention is the whole game

- Every format, long or short, lives or dies on watch time. Write for "will they reach the last
  second", not "is this good". (`2026-09-25_DdtfzTTNy2I_1`)
- Promise the payoff early, deliver it late: line one should make the viewer need the ending.
  Trends and countdowns withhold the answer until the close. In a demo, state the result up
  front, show the steps, reveal the finished thing last: the "oven moment". (`75tkthAzZog`)
- Predictability rewards, surprise rewards more. Write every piece as "X expected, Y happens":
  name the expectation and flip before scripting, and set up context fast; the sooner the
  viewer settles, the harder the flip lands. (`2026-09-28_Dd0_iCtBlAq_1`, `B_Uh5V4xD4k`)
- Use dramatic irony: let the viewer see the outcome coming before the people on screen do, and
  build toward it. The wait holds them. (`IMTetzcc6AE`)
- Whatever cures boredom today bores people tomorrow. Keep reinventing length, pacing and
  format before the audience tires of them. (`eQ-ljb2VFdU`, `B_Uh5V4xD4k`)
- Prefer premises everyone can enjoy: a polarising genre (horror reactions) can win a cult
  following but caps view duration. Do not build a series on a repetitive process (a
  weight-loss journey, someone painting); show only the moments where something changes.
  (`YRmo9aNAC24`)
- Optimise completion rate, not watch time, which longer videos inflate, hiding weak retention.
  His benchmark: about 70% average, around 60% weak, low-to-mid 70s strong. (`hKGe6WbHpWY`)
- Keep the screen alive: every few words a cut, slow zoom, music hit, sound effect or inserted
  meme, as a standing rule in the editor's brief. (`4Any9xEl-qA`)
- Cut to the creator's own laugh after a punchline: like a laugh track, it makes viewers feel
  they are watching with friends. (`hKGe6WbHpWY`)
- Treat each scene of a long video as its own short, with its own hook and closing beat. A few
  great moments carrying filler no longer hold an audience. (`4Any9xEl-qA`)
- Teaching? Package it like a short: quick points, music, changing on-screen text, something
  fun around it. Dry material gives the brain nothing to enjoy. (`EVu2Uxu_8iU`)
- Viewers can tell a video made only to keep a schedule; if you are phoning it in, change the
  format, do not fill the slot. Consistency helps only if each upload serves the reason people
  watch you: name it, check every piece against it. (`EVu2Uxu_8iU`, `q2_kutIbBxI`)
- Short-form views count everyone who scrolled past. For clicks or sales, lean on long-form:
  ten straight minutes build closeness and click-through; shorts get you found. (`JjDjDvNgkFo`)
- Followers matter less every year; platforms show you to people who clicked on you a few
  times. The real win is a viewer who opens your profile and watches ten more: make bingeable
  formats, package every video to earn the next click. (`75tkthAzZog`) One viral clip does not
  make paying fans: single-video viewers forget you; only those who watch many become a
  following that buys tickets. (`hKGe6WbHpWY`)

## 2. Design for the share button

- Every video earns one of two shares; pick one and write to it (`2026-10-02_Dd_dODBzpB6_1`):
  - **Send to a friend**: "this is so you / so us", built on shared identity. Sharpest form,
    "gifting an emotion": a near-universal situation (living with a flatmate) sent to whoever
    lived it with you. (`6XEK0I9HUk8`) Size the group on purpose: a shared emotion reaches
    almost everyone, a personal detail ("only tall guys get this") a few, very hard.
    (`B_Uh5V4xD4k`)
  - **Post to my profile**: the sharer uses it to look a certain way; give a take that sets
    them apart yet feels relatable. (`8DYIqJw5_bw`) A firm opening claim that splits people
    works: both sides share to show where they stand. (`RHi26oDdKQI`) This share usually wins
    with professional audiences: make the viewer look smart for passing it on.
- Be the "vehicle": say clearly what many feel but cannot put into words. They share it because
  it says what they meant. (`q2_kutIbBxI`)
- Learn from yourself: why you shared things is the template for what others share; why you
  clicked, finished a video or quit one are ideas or formats. (`q2_kutIbBxI`, `JjDjDvNgkFo`)

### Reliable share triggers (`6XEK0I9HUk8`, `8DYIqJw5_bw`, `B_Uh5V4xD4k`)

Pick one before scripting. With no clear trigger, default to warm and cheerful.

- **Topic of the day.** A ready-made opinion on what everyone is discussing gets forwarded as
  the viewer's own. Speed decides it: draft a post per likely outcome of a known event; publish
  as it lands. A never-ending sport (cricket) is a steady source: title a player's name on
  their big day. (`YRmo9aNAC24`) Ship the fast take where it takes minutes (a text post beats a
  reel for a live event); the video follows. (`SkU3J2vWUK8`) Topicality on autopilot: react to
  memes, which exist only about what is current; their subject is the title. (`YRmo9aNAC24`)
- **The opposite take.** Once a view is consensus, there is demand for a credible voice against
  it. A strong stance gets traction; fence-sitting does not.
- **Identity.** "Only people who X will get this" is close to a sure thing; identity plus
  positivity is the strongest pair.
- **Warmth.** Hyper-positive, cute (dogs, babies), nostalgic, or riding affection for someone
  loved. A break from a grim feed gets passed on.
- **Sensation.** Gossip or anything that breaks the everyday.

### Titles and thumbnails (`JjDjDvNgkFo`)

- Check that people care about the topic before reaching for title tricks; a clever paradox
  about something nobody cares about still gets nothing.

## 3. Starting a page from zero

1. Pick a narrow niche and a memorable page name. Test: for an obviously large topic (people
   management, which every office worker lives with), can you name its best creator? If not,
   there is room. (`RHi26oDdKQI`) Fill what is missing, with a personality on screen: a
   faceless, voiceless upload vanishes among thousands of identical ones. (`eQ-ljb2VFdU`)
2. Check topics against your market's culture before copying formats from abroad. India's
   dependable short-form themes were relationships, social inequality, local comedy and strong
   regional or national identity. (`SkU3J2vWUK8`)
3. Do not post yet. Cluster the top 10 videos of the niche's biggest channels into its 5 or 6
   recurring topics. Those titles are proven demand. (`2026-10-02_Dd_dODBzpB6_1`)
4. Start in short-form: it is cheap, so you get several attempts a day, not one a week, and an
   early win keeps you going. Move the audience to long-form once you have one. (`75tkthAzZog`)
5. For the first 1,000 views, ask friends with real followings to put a good video on their
   stories, then let the platform take over; who reshares beats how many. Working up close with
   a bigger creator shortens the climb a lot. (`75tkthAzZog`) Make early work for the specific
   people you want to notice you, posted where they already look, and put yourself in the
   building where the gatekeepers work; his college blog got him a magazine column that way.
   (`rHdheMfDnFg`)
6. Treat long-form as the anchor. A big YouTube audience tends to follow you elsewhere; a big
   Instagram audience rarely comes back. (`8DYIqJw5_bw`)
7. Run "spikes and waves": about weekly, a spike (a collab, or a special teased for days) pulls
   in new viewers; routine uploads keep them until the next. Measure fans who turn on
   notifications, not just views. A new channel cannot afford one big swing a month; that pace
   suits those with an audience. (`EVu2Uxu_8iU`) He plans it in YouTube's three tiers: "hero"
   (the rare big piece), "hub" (the special recurring format), "hygiene" (the dependable weekly
   upload). (`QLrCzQ3VRLc`)

## 4. Find your angle: "comedy plus X"

- Comedy (or any style) is the medium, not the content: the subject rides inside it. Compete on
  style plus real curiosity; pair side interests (a comic who is also a serious quizzer) and
  each world feeds the other. (`2026-09-23_DdooXuFSmLq_1`, `2026-10-04_DeE8mP3TQlu_1`,
  `RHi26oDdKQI`)
- In a crowded medium, perspective beats volume. Jokes are everywhere (memes, tweets); opinions
  and stories are scarce; the stand-up that broke through was story-driven, with a strong point
  of view. (`4Any9xEl-qA`, `hKGe6WbHpWY`)
- Once comedy is a medium, not an occasion, connection beats craft: people only laugh along
  with someone they like, which comes from being open about your flaws. (`SkU3J2vWUK8`)
- Know yourself to find what you share with millions, the common ground fans come for. His
  habit: an hour a night, phone away, journaling thoughts and feelings. (`q2_kutIbBxI`)
- Make what traditional media cannot or will not show. (`8DYIqJw5_bw`)
- Need your own distribution fast? Pick a "content hack": a low-prep format where your
  personality does the work (for him, gaming streams), so you can ship daily without writing
  each piece. (`2UDczkt_vaI`, `-GX9vu_gQ_Q`)
- Entering a field you are not known for (money, fitness, AI)? Come in as a learner, on camera
  beside someone who knows it, and the "what would he know" doubt goes away. Learning in public
  draws insiders to you; the mockery is the price. Go deep enough to spot parallels with content:
  that is how trends get seen early. (`4lhCsuwres4`, `QLrCzQ3VRLc`)
- A crowded category is still open if you break its conventions. Do the opposite of everyone:
  rude where the rest are polite, yesterday's stars where all cast today's. (`B_Uh5V4xD4k`)
  Break its format rules (long, film-like videos where all make the same short ones); a formula
  tired for one audience can feel new to another. (`EVu2Uxu_8iU`) The format alone can be the
  difference: his film reviews are fast jump cuts listing his thoughts while watching. Give the
  category your own repeatable wrapper. (`JjDjDvNgkFo`)
- Your audience follows you across formats (sketches, vlogs, reactions) while your core skill
  stays. On a bigger stage, do not become a stiffer version of yourself. (`EVu2Uxu_8iU`)
- Over ten years, judge an angle by how it scales and tires you: a skill for any subject
  (making anything funny) crosses genres; needing travel or special access wears you out.
  (`JjDjDvNgkFo`)
- React formats on already-viral clips have a quality floor: add something, do not be the
  boring part. (`IMTetzcc6AE`) Run a react panel as a game: the cast competes live for the best
  line, cast for contrast (a likeable small-town voice, a deadpan, a loud host who sets the
  energy), and viewers feel part of it when their joke is picked up. (`4lhCsuwres4`) Avoid formats where the audience can do your job better: in a
  react video the comments out-joke the panel every time. (`4Any9xEl-qA`)
- Every field has choices that impress fellow practitioners, not the audience. Drop the ones
  aimed at peer approval when they do nothing for the viewer. (`nWuk46Ifl9E`)

## 5. Reps beat talent

- Expect to throw most of it away: by his count, roughly a thousand pages a year for 50 to 100
  usable ones. The volume is how you get good. (`75tkthAzZog`)
- Shorten the loop (make, publish, read the response, adjust) and run it hundreds of times;
  judgment comes from loops, not years served. A feed posting several times a day learns more
  in a month than a twice-a-month channel in a year; past about the 150th attempt you can
  predict how a piece will do. Use such a channel to test idea combinations before a bigger
  production. (`75tkthAzZog`, `B_Uh5V4xD4k`, `9k8Gsb45o5Q`)
- Restarting? Book a fixed daily slot (he booked a room for a stand-up set every evening for 50
  days) and listen back to each set the same night. (`2UDczkt_vaI`)
- Take apart work you admire with a peer who knows how it is built: treat jokes, music and
  video as structure and rhythm, not magic. (`OjsNcrehxRg`)
- Take a wide range of work early (columns, TV, award shows, ads); it becomes the fluency that
  makes any format easy later. (`c6eC0gF6dHQ`)
- Learn how a piece is constructed, not one platform: construction carries across TV, ads,
  YouTube and reels, so you can move when the next medium arrives. Joke writing is the most
  portable; it improves a job post or a live stream too. (`75tkthAzZog`, `hKGe6WbHpWY`)
- In a slump, pick work with a fast feedback loop at your current skill level, such as
  live-streaming, not a distant goal like "impact". (`c6eC0gF6dHQ`)
- Let data pick the area and angle; the creative runs free in it. Data cannot write the idea.
  (`B_Uh5V4xD4k`)
- Run the channel like a product team: before each upload ask what this audience wants next,
  double down on what works, keep a few experiments beside it. Turn a hit into a series; give a
  series about three episodes, drop it if none nears the original. (`hKGe6WbHpWY`)

### Format succession (`IMTetzcc6AE`, `YRmo9aNAC24`)

- Run the channel like a TV network: each working format is a show in a slot; your standing job
  is scheduling the next. Start from something you find fun; the first shoot is the test: if
  the room reacts as you film, the audience probably will.
- Expect a winning format to wear out in about three months. Before then, run a string of cheap
  tests, expect most to flop, and have the next one live as the current one fades.
- Build the next format from the last one's mechanism, not a blank page: find why the winner
  works (say, a panel-by-panel build to a punchline), then apply it to a new subject.
- Prefer formats and topics that bring new viewers: a new guest or the day's news brings fresh
  ones each episode (`JjDjDvNgkFo`); content about a country or community pulls that group in
  as followers.
- A repeating data source (say, the week's most shared memes) can become a fixed weekly series,
  and a fixed series can grow into a property of its own.
- A format that depends on the people or places around you dies when your life changes; treat
  that as the cue to find the next one.
- Re-check format rules against how the platform works today: "every video must pass 10
  minutes" became "cut everything" within a decade. (`4Any9xEl-qA`)

### Numbers to watch (`hKGe6WbHpWY`)

- Read the channel monthly, not per video; the per-upload dashboard compares each video to the
  last ten like a stock ticker, built to make you anxious. Track a short list: monthly trend,
  comments per video, unique viewers.
- Read click-through with average view duration. A topic people are curious about but cannot
  sit through gets clicks, then a fast drop-off: failed, even if a few love it. (`YRmo9aNAC24`)
- Comments lead: on his channel, months with more comments brought more new subscribers. Script
  moments that make people want to reply.
- On a YouTube Premiere, the live concurrent count is an early read on click-through, and the
  chat shows the reaction. His bar: weaker videos stay under 30 to 40k concurrent, the best pass
  40k. (`4lhCsuwres4`)

## 6. Content is an asymmetric bet

- A post, like a book or a first date, can return far more than it cost. Volume of good
  attempts is the strategy. (`2026-09-26_DdwCFZwxnXv_1`)
- Risky ideas are how you go viral, and some will embarrass you in public; decide in advance to
  own those misses. Keep a proven format beside every risky one: after a flop, go straight back
  to what works and ship again; do not sit with it. (`9k8Gsb45o5Q`)
- Beware variable rewards. One breakout makes you chase it in every video after; judge a video
  on its own job. (`2026-10-07_DeMYoJMTdBW_1`)

## 7. Fit before money

- Enjoyment fit comes first, business fit follows. Creating only because it is the thing to do
  burns out. (`2026-09-24_DdrW_s0zCFP_1`)
- Keep the job while you start; quit only when content alone could carry you for about a year.
  His own switch took 3 to 4 years to match his old salary. (`75tkthAzZog`) Go where the money
  and attention are now (TV paid him five times what print did), and let a well-paid craft job
  fund the riskier art: his TV savings carried two years of unpaid stand-up. (`rHdheMfDnFg`)
- Keep a cash reserve of about three years of costs before investing, so a dry spell in deals
  cannot force your hand (his advice as retold by a creator he mentored, `0oKfNl14MEc`).
- Do not plan on short-form ad revenue. Money comes from long-form ads and brand deals, priced
  on views, genre (finance pays more) and how premium you seem; fewer deals a year raise the
  rate. (`75tkthAzZog`) Long-form ad income is bigger than most assume; the per-view rate rises
  with view length and with viewers in high-paying regions or on Premium. (`JjDjDvNgkFo`)
- Do not expect videos to pay: like a cricketer's endorsements or a singer's shows, money sits
  around the work. The asset is distribution and a brand people connect with. (`YRmo9aNAC24`)
- Education and finance creators have a product (courses); entertainers build one on their
  craft, as a comedy writer moving into advertising and advising startups. Co-found with past
  collaborators who run operations; your one job is growing the audience. (`JjDjDvNgkFo`)
- A service business is an easier first step than a consumer brand. In a niche, trust becomes
  money fastest through a fitting service (a nutrition plan, an interiors consultancy): you are
  the product and charge from day one. (`RHi26oDdKQI`) Partner with someone who knows the
  field; the audience feeds it only while you keep finding new viewers. (`75tkthAzZog`)
- Growing a page from 0 to 10,000 proves you can build distribution and sell; it beats a
  resume. (`2026-09-30_Dd6g0w2zUVp_1`, `2026-10-01_Dd9Rxv5TAzC_1`)

## 8. Team, partners and pitching

- Pitch by reading the work and the room. Concede the line you sense will not survive; giving
  ground buys trust for the rest. (`2026-09-29_Dd3ynUGSRSJ_1`) Never read to a cold room: spend
  the first minutes catching up and joking so they find you funny before the script, and never
  follow something sombre with light material. (`4lhCsuwres4`)
- When a client says no, ask exactly why. A precise reason becomes a guardrail for the next
  round; vague no's stall projects. (`EVu2Uxu_8iU`)
- Run small pods that each own one property (say, two editors and a producer); never one person
  juggling several shows. (`75tkthAzZog`)
- Know your partners: "sharpshooters" stay quiet, then give one strong idea; "workhorses" give
  range. Use each for their strength; keep the same collaborators through many rounds until you
  can tell an idea will land without the room laughing. (`9k8Gsb45o5Q`)
- Share upside: when a sponsored piece pays, everyone on it gets a cut; credit every team
  member by name and handle in each video. Many creators faded after their team left.
  (`2026-10-02_Dd_3MbUTgWz_1`, `JjDjDvNgkFo`)
- Generate wide, then narrow: collect the category's best work, pitch around 20 ideas,
  shortlist about 5, develop them, say "no winner yet" till one lands. (`B_Uh5V4xD4k`)
- Keep an "insight Bible": the team logs everyday observations by topic as they happen, so a
  new brief has raw material for ten ideas in a couple of hours. (AIB practice, `EVu2Uxu_8iU`)
- Send drafts constantly to a small circle with sharp taste, mainly to find what to cut; ignore
  critics performing cleverness. (`B_Uh5V4xD4k`) Test on a room that looks like the audience;
  aim for broad approval of the taste, not unanimity. (`c6eC0gF6dHQ`) After weeks inside a
  script, remember the viewer sees it once, fresh: do not cut bits that made the room laugh the
  first time just because you are tired of them. (`B_Uh5V4xD4k`)
- Hire for hunger, independence and adaptability, not hours. Ask "which of your opinions do most
  of your friends disagree with?"; be wary of applicants who mainly want to be near a famous
  name; a rigid mind is a deal breaker; repeat applications are a good sign. He likes "lazy
  geniuses" who find the shortest path to a strong idea. (`2UDczkt_vaI`, `-GX9vu_gQ_Q`)
- Tell each hire their role, what they can learn from you, and that they will leave one day,
  then back what they build next. His former social team's agency now runs his social media:
  alumni keep feeding the creator. (`2UDczkt_vaI`, `-GX9vu_gQ_Q`)
- Guard access to big names: never spend it on a request that wastes their time (a fan-style
  "can I come on", a selfie); approach as you would want to be approached. (`YRmo9aNAC24`)
- Keep makers always around: a room for collaborators who stay for days, a writers' room next
  to the set. Their constant company keeps ideas coming. (`YRmo9aNAC24`)

## 9. Content for brands

- Content pays back in ways a campaign sheet misses (it attracts hires, makes staff proud), so
  do not judge it on direct return alone. (`75tkthAzZog`) When pitching, name the knock-on
  value beyond views: easier hiring, proud employees, easier sales. The best clients
  collaborate, with a founder who has said a miss is fine. (`B_Uh5V4xD4k`)
- Separate campaigns that move a number now from brand building; do not judge one on every
  metric. (`EVu2Uxu_8iU`)
- Start with short video for fast feedback; run content always-on, not one-off. (`75tkthAzZog`)
  A creator mentality is going viral again and again for years, not one stunt. (`EVu2Uxu_8iU`)
- Staff it with creators: one in-house producer who owns the brand's taste, ideally someone who
  has built an audience; outsource the rest. (`75tkthAzZog`) Established brands should build an
  in-house creator team; startups can bring a creator in early. Hire marketers who make their
  own content and sometimes fail at it. (`EVu2Uxu_8iU`)
- Aim for great work, not for never making bad work. Test: could someone enjoy a year of your
  ads back to back, even for a dull product? (`9k8Gsb45o5Q`)
- Meeting expectations builds the brand; beating them makes the content. Keep one recognisable
  style but change execution often; audiences tire faster now. (`B_Uh5V4xD4k`) Do not hype the
  ad: a teaser promising an amazing main piece is a warm-up act promising the funniest comedian
  alive; keep expectations low and over-deliver. (`ooQA94UTk5k`)
- Work content-first: find what is interesting, then where the brand fits. For a weeks-long
  campaign, build a repeatable gag that makes people ask "who's next", bold enough to be talked
  about, catchy enough to survive heavy repeats. (`EVu2Uxu_8iU`)
- Treat a hard length cap as fixed; the first draft says too much. Cut at least two rounds
  before pitching; saying less leaves the audience room to fill in. (`B_Uh5V4xD4k`)
- Write to the assets you have: a concept built on a specific celebrity or prop dies without
  them, so confirm access first. (`B_Uh5V4xD4k`)
- Premium buyers pay to skip ads on every platform, so paid media barely reaches them; reach
  them with content they choose because it is useful or entertaining. (`RHi26oDdKQI`)
- A sketch and an ad are the same skill: decide the one thing the viewer should feel or do at
  the end; in an ad that end point is a product. (`QLrCzQ3VRLc`)
- An agency's weak ads stay anonymous; a creator-led shop's flops land on the creator's name.
  Judge yourself on the next piece out, and ask before every release whether this client will
  want to come back. (`QLrCzQ3VRLc`)
- Utility alone is forgettable; people get attached to useful and beautiful. Polish is
  loyalty-building, not decoration. (`nWuk46Ifl9E`)
- Celebrity endorsements work mostly through legitimacy: "this brand can afford them" reads as
  trustworthy and here to stay. Booking a big creator sends the same signal. (`JjDjDvNgkFo`)

## 10. A career that outlasts the algorithm

- Audience is step one, not the career. Plan for the decline while still growing, as if you
  must stop posting in two years; use today's reach to build what no platform can switch off.
  (`c6eC0gF6dHQ`)
- Viewers binge you for months, then move on; creators come in batches. Plan a steady inflow of
  new viewers; reinvent every few years or the next batch replaces you. (`JjDjDvNgkFo`)
- Do not live on brand money alone. Brand deals are "high rent": you are renting out your time.
  While growing, start streams two, three and four (affiliate, consulting, merch, a business);
  try all, since you cannot predict what your audience will pay for. (`JjDjDvNgkFo`) Earn
  directly from fans (a paid inner circle, where a few thousand monthly members is a real
  income) or from utility, with frictionless payment, not stitched from stories, chats and
  meeting links. (`EVu2Uxu_8iU`)
- Build income your opinions cannot touch before taking strong public stands: brand money
  leaves first when comments turn; financial security buys freedom to speak. (`JjDjDvNgkFo`)
- More followers is not the only route to brand money; creators giving real value earn more per
  fan than pure entertainers, whose pull fades; seek deals that help the audience.
  (`c6eC0gF6dHQ`)
- If growth stalls, a marketing job is a ready exit: brand marketing is mostly judging what
  will work, which you have practised more than most marketers. (`JjDjDvNgkFo`)
- Track views per subscriber as your relevance signal; it drifts down as channels age. Measure
  each push against your own best, not subscriber count. If old fans want the old style but the
  new one does better, follow the numbers: annoying some is the price of leaving a quietly
  dying format. (`IMTetzcc6AE`)
- Set the next concrete goal the day you hit one: first million-view video, then a sold-out
  show, then a format nobody in your market has tried. (`8DYIqJw5_bw`)
- Taking many projects from zero to one keeps a creator from boredom; scaling one from one to
  ten is repetition, dull but valuable. Know which stage you are in. (`YRmo9aNAC24`)
- Run the career as a "marathon of sprints": push hard on one bet, rest properly, start the
  next fresh, rather than grind at one pace until you burn out. (`IMTetzcc6AE`)
- Fans who already help each other in your comments are a community waiting for a home: give it
  topic channels, live voice rooms, good moderators; your first power users come from them.
  (`EVu2Uxu_8iU`) A closed space organises an audience but does not find one: grow on open
  platforms first. Decide before you open it whether you can moderate it; he shut a server of
  roughly half a million mostly young members rather than have its posts tied to his name.
  (`SkU3J2vWUK8`)
- Use live streams for closeness, not reach. Hours of live, unguarded time turn respect for
  your work into affection for you. (`SkU3J2vWUK8`) A stream holds people when someone dropping
  in halfway can read the situation at a glance, like a scoreline; open drop-in talk has no such
  state. Put stakes on the ending (a forfeit, a reveal) and go live at your market's peak hour
  (for him, about 9 pm India time). (`4lhCsuwres4`)
- Magazine lists, covers and prestige shows mostly serve the outlet. Ask what you get, accept
  only hosts you trust; a status appearance can still be worth it for the status.
  (`4lhCsuwres4`)
- Running a company as a creator? Hire a dedicated CEO. An artist's attention drifts back to the
  work, and a 40-person company needs someone whose only focus is the business. (`2UDczkt_vaI`)
- Do not let your self-worth rest on output or approval; he calls that dependence an
  occupational illness of creators. Build an identity outside the channel before it stops.
  (`2UDczkt_vaI`) Public work trains a harsher inner critic; judge the body of work on whether
  you made it honestly and your way, not on each verdict. (`OjsNcrehxRg`)
- Most pile-on hostility is envy or a bad day; the snarky reply under a big account is often
  "clout farming" for likes and says little about the work. (`SkU3J2vWUK8`) Still, read hostile
  comments as a preview of how critics will come at you once bigger, and prepare; the precise
  criticism that stings is the useful kind. (`c6eC0gF6dHQ`)

## Review checklist (run on every script)

- [ ] Does line one create a question only the ending answers?
- [ ] Is the best item or reveal last?
- [ ] Is there one twist that beats the obvious expectation?
- [ ] Does something change on screen every few seconds?
- [ ] Which share is this built for: send-to-friend or post-to-profile? Which trigger does it use?
- [ ] Is this a proven topic in the niche, or a guess?
- [ ] What is the angle: whose real curiosity or point of view is showing?
- [ ] Could the comment section do this better than we can?

## Rules of use

- Borrow the mechanics, not the man. Do not imitate his voice, reuse his lines or jokes, or
  write as if he said something he did not.
- Never imply Tanmay Bhat endorses, sponsors or reviewed the content being made.
- When attributing an idea to him, link the original from `INDEX.md`. If you have not checked
  the source, say "according to a talk by Tanmay Bhat", not a direct quote.

## Sources

- `INDEX.md`: every source with date or link. `2026-09-25_...` ids are Instagram clips;
  11-character ones (`75tkthAzZog`) are YouTube videos. "About him" clips carry no technique.
- `references/` (optional, local only): full transcripts if you built them yourself. Search
  with `grep -ril "<topic>" references/`; they are machine transcripts of Hinglish audio, so
  names and numbers can be wrong. They are not distributed with this skill.
