There’s a quote that reframed this whole series for me, from a strategist whose clients pull in roughly 10 billion views a year: “You don’t make videos that people watch. You make title-thumbnails that people click, and videos that people watch. It’s 50/50.”

Most creators spend 95% of their time on the video and 5% on the idea and the packaging. The people generating outsized results have that ratio closer to reversed. Once you understand the two-gate machine from Part 2, this stops being a hot take and starts being obvious: the idea sets your ceiling before you shoot a single frame, and the packaging is the entire input to Gate 2’s first prediction. Everything downstream is capped by what happens before the camera turns on.

The idea sets the ceiling

An idea that only appeals to 5,000 people can be executed perfectly and still only reach 5,000 people. No amount of editing fixes a ceiling that low. So the highest-leverage work happens before production, and it happens in a structured way, not by waiting for inspiration.

Three source buckets, roughly equal weight, refreshed monthly:

Internal : what has already worked on your own channel, or on your audience’s other platforms. This is your highest-confidence bucket, because you have direct evidence it works for exactly your people.

External : outlier videos across YouTube, defined precisely as videos doing 3 to 4 times a channel’s average. Not “big videos.” Videos that dramatically overperformed their own channel’s baseline, the same baseline concept from Part 2. The discipline here is extracting the pattern, not the video. A format like “5 minutes vs. 50 minutes vs. 500 minutes” isn’t owned by whatever niche it started in. It’s a structure you can port anywhere.

Original : genuine swings tied to something you can do that others can’t. Write this down literally: what is your competitive advantage? Access, experience, a unique story, resources, authority. Not every idea needs to come from this bucket, but every channel needs some.

Subscribe now

The adjacent-niche move. The best pattern-mining doesn’t happen inside your own niche, it happens one step sideways. Fishing and golf share almost nothing on the surface, but they share the same demographic, the same gear-heavy culture, the same skill progression. A format that’s exhausted in fitness content can feel completely fresh in personal finance, because it’s never been tried there. There’s a nice way to think about this: a magician doesn’t need new tricks, just audiences who haven’t seen the ones they have.

The filter that actually matters. Before an idea survives to production, it has to answer one question honestly: can this be packaged? Not “is this interesting to me,” not “do I have something to say here.” Can it become a title and thumbnail that earns a click. If the answer is no, the idea gets shelved, sometimes for years, until a packaging angle appears. This single filter eliminates most of what creators waste time producing.

The kill rate at the top of this funnel is brutal by design. For serious operations, somewhere around 99% of generated ideas never get made. If most of your ideas make it to production, you’re either not generating enough of them or not being honest about the filter.

^{If this is too much reading down below, we have a diagram for visual learners. Hope that helps :)}

Packaging: the 50% nobody trains for

If the idea sets the ceiling, packaging decides whether anyone finds out the ceiling exists. Two things carry almost all the weight: the title and the thumbnail.

Titles that work share a few patterns, and it’s worth having language for them so you can deliberately reach for one instead of hoping something clever occurs to you:

Superlatives: fastest, cheapest, first, biggest. Simple, visual, instantly legible.

Comparison: putting two things against each other taps something close to a compulsion. Humans can’t not resolve a comparison.

The three-step progression : a version of comparison with a built-in structure: $1 vs $10 vs $100. Ten minutes vs one hour vs twenty-four hours. This format does double duty because it also hands you your thumbnail composition for free.

Contradiction: the most powerful of the group, and the least used. State something, then undercut it. “I found the cheapest version of this car in the world, but nobody will buy it.” The second half detonates the first, and the viewer has no choice but to ask why. This is a curiosity gap manufactured on purpose rather than hoped for.

Exclusive access: showing viewers somewhere they can’t otherwise go.

A useful gut check: write the most boring, technically accurate version of your title next to the actual one. The gap between them is a rough measure of how much packaging is doing for you. It’s often enormous on videos that otherwise look identical in content.

Thumbnails are mostly psychology, not design. The design matters, but it’s a small fraction of what makes a thumbnail work. The bigger fraction is whether the image creates a question the viewer needs answered.

Subscribe now

A few ideas that show up constantly in how strong thumbnails are built:

The image should feel paused in the middle of something, like clicking will unpause it rather than start it. A door mid-open. A moment mid-speech. Something already in motion that the click completes.

Jeopardy beats resolution. A thumbnail where the outcome feels uncertain outperforms one where everyone’s smiling, and everything looks fine. Uncertainty is what pulls the click.

Text, if used at all, should be four or five words, maximum. The rest of the thumbnail should be readable in the amount of time someone spends deciding whether to keep scrolling, which is closer to a glance than a look.

And critically: build more than one option. Three genuinely different concepts at minimum, not the same image with different text. The habit of treating the first thumbnail idea as the final one is probably the single most expensive habit in the entire craft, because of everything from Part 2 and Part 3: packaging determines Gate 2’s first prediction, and small differences here compound nonlinearly through the expansion loop.

The 80% rule

Here’s the tension every growing channel eventually hits: staying consistent enough for the algorithm to understand you, versus staying interesting enough that you don’t get bored and the audience doesn’t either.

The resolution is a simple rule: aim for each video to interest roughly 80% of the people who watched your last few videos, and follow that discipline about 80% of the time. The remaining 20% is explicitly reserved for swings, experiments, and ideas you’re not sure about.

This maps directly onto the machine from Part 2. Consistency keeps your audience cluster coherent, which keeps you inside the candidate pool. The 20% margin keeps the channel from calcifying into something that only works because it never changes, and it gives you a controlled, low-risk space to test the original-bucket ideas that might become your next format.

The rule also answers a question a lot of creators ask themselves late at night: is my niche too small? Usually the honest answer is that the niche isn’t the problem, the packaging is. A narrow topic, packaged one level broader, can pull in adjacent audiences without losing the core. The content stays specific. The packaging gets wider.

What I’m doing with this

1. Time reallocation. I’m deliberately moving hours from production toward ideation and packaging, even when it feels like the “real work” is the editing. Based on everything in this series, it isn’t.

2. A written competitive-advantage list. What can I make that’s harder for someone else to replicate. If I can’t answer this in one sentence, the original-idea bucket stays empty.

3. Multiple thumbnails, always. Not as a nice-to-have. As the default, because the cost of one extra concept is small and the asymmetry in outcomes, from Part 3’s dashboard traps and Part 2’s expansion loop, is large.

Subscribe now

Let’s Learn From My Mistake

Recently, these thumbnails were made for my channel,

This thumbnail did not work; the reason was that I was copying ItalianBach and ArthurTV’s thumbnails. Their thumbnails work because they have familiarity bias; I do not have that yet. That being said, I need to make sure this mistake is not made again, so I will be redesigning my thumbnails to see what works. Frustrating? YES. But it is the name of the game.

Here is the diagram for those are visual learners.

The one-sentence version

The video is only half the product, the idea and its packaging are the other half, and almost everyone underinvests in the half that determines whether the machine from Parts 2 and 3 ever gets a chance to work in your favor.

Next in this series: retention engineering, the specific structural devices, the first ten seconds, the deliberate use of foreshadowing, that keep a viewer in the video once the click has already happened.

Talk Soon

Sid

Then I uploaded a video called “Dating By Outfits,” and the machinery turned around and graded my homework.

Here's the Video

I’m not going to bury this one. Here’s the scoreboard after four days:

• Impressions: 8.4K

• Impressions CTR: 1.0%

• Views: 100 (130 below what’s typical for this channel at this stage)

• Watch time: 1.4 hours, against a “typical performance” band that runs up to 20.8 hours

• Average view duration: 0:48, on a 13:49 video

That last number is the one that stung. Forty-eight seconds. On a video nearly fourteen minutes long. Let’s actually diagnose this instead of flinching at it.

First: the bouncer let it through

This is the useful part. Per Part 2, candidate generation only cares whether your channel has a coherent taste-space fingerprint it can slot new uploads into. 8.4K impressions in four days is the system saying yes, I know who to test this on. So whatever’s wrong with this video, it isn’t an invisibility problem. Nobody needs to hear “just be more consistent” here. The bouncer did its job.

Which means the damage is happening entirely at the next two gates.

Second: the packaging is getting rejected, not just under-rewarded

Here’s where I have to be honest with the number instead of explaining it away. Part 3 lays out the CTR trap: on a healthy video, CTR falls because impressions are expanding to colder, less-invested audiences. That’s what success looks like at scale.

1.0% CTR at only 8.4K impressions is not that story. This is still basically a warm test ring, the friendliest audience the video will ever get, and the click rate is still down at 1%. A healthy warm-audience CTR usually sits somewhere in the 4–10%+ range. There’s no charitable read of a 1% number this early. The title and thumbnail aren’t losing a fair fight against expansion. They’re just not earning the tap.

And per Part 2’s point about ignored impressions being spent, not free: every one of those roughly 8,300 people who scrolled past just taught the system, for me specifically, that this thumbnail isn’t worth showing again. That’s why the view curve is already flattening by day four instead of climbing into a second ring. The algorithm isn’t being cruel. It’s just done testing.

Third: the sixty seconds that mattered most, I spent on myself

Forty-eight seconds of average view duration on a 13:49 video is a textbook first-30-seconds cliff, not a gradual slope. Something in the opening broke the promise the thumbnail made.

The actual reason I’m writing this post: the hook itself was front-loaded correctly. The outfit reveal, the premise, all of it happens right at the top, exactly like the garlic bread lesson from Part 3 says it should. But I kept pausing the reveal to cut in with my own commentary over it.

Read that back next to the line I wrote two weeks ago: “every minute has to pay rent.” A pause to narrate isn’t neutral. It’s the video stalling at the exact moment it has the least trust banked. Someone clicked because the thumbnail promised motion, a bit playing out, a moment resolving. The first thing they got instead was me interrupting the moment to talk about the moment. That reads as “this isn’t going to move,” and gone they went, most of them before the one-minute mark.

It’s a genuinely easy trap to fall into as a solo creator, because commentary feels like the value-add. But sequencing is the value-add. Let the beat land first. Comment over the reaction, in a cutaway, after the payoff breathes for a second, not instead of the payoff.

Two separate jobs, not one

The thing I want to flag for future-me, and for anyone reading who’s staring at a similarly rough scoreboard: these are two unrelated failures wearing one video.

CTR lives outside the click. No amount of fixing the internal pacing moves a 1.0% number, because retention problems can only happen after someone has already decided to tap. The title and thumbnail need to be re-litigated entirely on their own merits, independent of anything I do to the edit.

Retention lives inside the click. Fixing the pause-and-narrate habit won’t rescue a thumbnail that isn’t earning attention in the first place, but it will matter enormously for whoever does click, both on this video’s eventual repackage and on the next ten uploads.

Conflating the two is how creators end up “fixing” the wrong thing after a flop.

What I’m doing about this one soon

Straight from my own rule in Part 2, no mourning, repackaging: I’m not eulogising this video.

1. New title and thumbnail, tested against what a scroller in this niche actually expects from a dating/outfit-reveal premise, not against what I personally find funny about the concept.

2. Re-cut the open so the reveal plays out uninterrupted, with commentary moved to sit over the reaction or in a quick cutaway after the beat lands, never frozen on top of it.

3. A note in my baseline sheet flagging this specific failure mode, pausing on the hook, so I catch it in the edit next time before it ever reaches an audience.

The one-sentence version

A coherent channel and a front-loaded hook can still get buried by a thumbnail that doesn’t earn the click and an edit that stalls the payoff it just promised, and the fix for each has nothing to do with the other.

YouTube Theory #1

But here’s what most creators never updated: that was fourteen years ago, and the platform has moved on again. Watch time is no longer the currency. Satisfied watch time is. And the difference between those two things explains a whole generation of channels that did everything “right” and still faded.

Let’s walk the metrics layer from top to bottom.

The funnel, in order

Every number in your dashboard is a verdict on something different. Confusing which verdict is which is where most misreads start.

Impressions come first, and almost nobody looks at them properly. Impressions are the algorithm’s allocation decision: the truest available read on how much the system currently believes in a video. Low impressions isn’t a packaging problem, it’s a consideration problem (that’s Part 2’s territory).

Click-through rate is the market’s verdict on your packaging, within whatever allocation you got. Views are just impressions times CTR. Average view duration and retention are the verdict on the video itself. And then come the two levels you can’t see directly: satisfaction estimates, and what the viewer did next. Those invisible levels are the ones that actually move long-term distribution.

The CTR trap (read this before you ever panic-swap a thumbnail again)

Here’s the dashboard trap that catches even experienced creators: CTR is only interpretable jointly with impressions.

When a video performs well, the system expands it to bigger, colder audiences — people who’ve never heard of you, browsing casually, on surfaces where you’re one thumbnail among twenty. Of course, a smaller percentage of them click. So on a winning video, CTR falls while impressions rise. That’s not decay. That’s what success looks like.

A 12% CTR on 10,000 impressions (warm test audience) and a 4% CTR on 5 million impressions describe the same healthy video at two stages of expansion. Panic-swapping a thumbnail in the middle of that expansion, because “CTR dropped,” is one of the most common self-inflicted wounds on the platform. Judge CTR at comparable impression scales, against your own history, never as a lone number.

Watch time vs. retention, precisely

People use these interchangeably. They’re different verdicts.

Watch time is total minutes: 1,000 viewers watching 5 minutes each is 5,000 minutes. It’s YouTube’s inventory metric, the thing that correlates with ad revenue. Retention is the percentage of the video the average viewer consumes, and the ranking system compares it against videos of similar topic and length. Which means a 6-minute video at 60% retention can beat a 12-minute video at 50%, despite generating less raw watch time.

The strategic implication: length is not a strategy. Length that holds is. Padding a 6-minute idea to 12 minutes buys you raw minutes and costs you the comparison you’re actually being judged on.

Reading a retention curve like an analyst

Your retention graph isn’t a grade. It’s a story with known signatures, and each one has a specific diagnosis:

The first-30-seconds cliff. Some early loss is normal (mis-clicks happen). A steep drop means your opening broke the thumbnail’s promise. The fix lives in the first ten seconds: deliver on the click before you do anything else.

The gradual slope. Healthy. Every video leaks viewers. Don’t fight gravity.

The mid-video cliff. Something specific happened at that exact timestamp: a topic shift, a tangent, a badly handled sponsor segue, a pacing collapse. Scrub to the second and name the cause. The system noticed; you should too.

Bumps and spikes. Moments people rewound and rewatched. This is your audience voting on what your next video should be. Most creators scroll right past their own best market research.

The flat line that lies. A high, flat curve looks like victory. But if people stayed out of sunk cost or background-listening while the satisfaction signals came back weak, you’ve produced the padding signature: strong watch time, hollow experience. This is the exact profile the modern system learns to quietly deprioritise.

The end cliff. Everyone leaves when you signal wrap-up. The moment you say “so, in conclusion,” you’ve told viewers the value is delivered. Put your end screen over content, and bridge straight into the next video instead of rolling credits.

The satisfaction system: the referee you never see

Now the layer that changed everything.

YouTube runs millions of viewer surveys every single day. Post-watch prompts asking whether a video was worth your time, star ratings, homepage panels. And the responses don’t sit in a report somewhere — they feed directly into the recommendation models, which learn to predict satisfaction for every video on the platform, including the ones never surveyed. Alongside surveys: likes and dislikes, “Not interested” and “Don’t recommend channel” clicks (heavy negative signals), shares, and, maybe most importantly, what the viewer does after your video ends. Did they keep enjoying YouTube, or did they close the app feeling vaguely robbed?

That last one has a name: session contribution. And it points at the deepest fact about the modern algorithm. The system’s real question, confirmed in Google’s own research on the platform’s long-term optimisation, is not “will they watch this video?” It’s “if we show this now, will this person come back tomorrow?”

Sit with that, because it inverts the old creator instinct: YouTube would rather a viewer leave early feeling great and return daily than binge two drained hours and quietly stop opening the app. You are not competing for minutes. You are competing for the viewer’s relationship with the platform.

Subscribe now

The garlic bread proof

The cleanest demonstration of this whole layer is a Tom Scott video called “We Sent Garlic Bread to the Edge of Space, Then Ate It.”

Think about what the average creator does with that concept. Film the planning. Vlog the drive to the launch site. Stretch the suspense. Turn a 6-minute idea into a 20-minute “experience” and harvest the watch time. Viewers might even finish it and leave feeling faintly shortchanged, which is exactly what they’d tell the survey.

Tom did the opposite. The garlic bread is on its way to space within the first five seconds. Promise made, promise kept, concept executed cleanly, done. He sacrificed harvestable watch time and bought something the system values more. The video sits at tens of millions of views.

The lesson is not “make shorter videos.” The lesson is that every minute has to pay rent. The satisfaction system was specifically built to find the 20-minute version of a 6-minute idea and stop recommending it.

^{I know this amount of info can be overwhelming; if it is, you can hear it on audio or check out my podcast called the Creators Corner Podcast}

Good abandonment (the early exit that counts for you)

One more idea that reframes “failure”: if someone leaves your tutorial at minute three because they got the answer and went off to do the thing, the system increasingly reads that as a win. An exit into action is not the same as an exit into a competitor’s video, and the models have gotten smart enough to tell the difference. Creators call it good abandonment. A satisfied 20%-watcher who likes the video and comes back tomorrow beats a resentful 60%-watcher every time.

This is also why chapters, which literally help people skip parts of your video, tend to raise satisfaction on tutorials, listicles, and podcasts. Paddy Galloway's hierarchy, from his work across channels doing 10 billion views a year, puts it plainly: satisfaction first, then retention and CTR together. His comparison is the old SEO recipe page that buried the banana bread recipe under 800 words of filler. String-along videos are the same crime, and this whole measurement layer exists to catch them. (One boundary: use chapters for tactical content, skip them for narrative, where they leak the payoff.)

What I’m doing with this

Three habits, straight from the mechanics:

1. Reading impressions before CTR, and CTR only against comparable impression scales. No more single-number panic.

2. Scrubbing every retention curve for the six signatures, logging the exact timestamp of any cliff, and checking my next video for a repeat of the same cause.

3. Auditing my scripts for rent. For every section: does this minute deliver something, or is it inflating the runtime? If I wouldn’t defend it to a viewer who asked “why is this here,” it goes.

The one-sentence version

Your dashboard measures four different verdicts allocation, packaging, delivery, and satisfaction and the invisible one, whether the viewer left glad and came back, is the only one the algorithm is ultimately optimizing for.

Next in this series: the strategy layer how the people behind 50 billion views actually generate ideas, why packaging is half the game, and the 80% rule that keeps channels coherent without making them boring.

YouTube has over 14 billion videos and around 2 billion daily users. When someone opens the app, the system has well under one second to build their entire feed. Carefully asking “how much would this person like this video?” for all 14 billion videos, for every user, every time? Computationally impossible. Not hard. Impossible.

So Google’s engineers split the job in two. And once you understand the split, a whole category of YouTube frustration suddenly makes sense, including, probably, yours.

Stage 1: The bouncer

The first stage is called candidate generation, and its only job is to narrow 14 billion videos down to a few hundred that are plausibly right for one viewer.

It doesn’t watch anything. It doesn’t deliberate. It works off broad behavioural patterns, mainly one called co-visitation: people who watched A tend to also watch B. Every viewer gets represented as a point in a giant “taste space” built from their watch history, and the system grabs the few hundred videos sitting closest to that point. It also pulls in fresh uploads from channels the viewer has recently watched, which is why watching two videos from a channel puts its next upload on your homepage whether you subscribed or not.

Think of it as a bouncer with a guest list. Fast decisions, zero nuance.

Here’s the part that matters: your video cannot rank, cannot be tested, cannot do anything until it gets past the bouncer. And the bouncer doesn’t know your video is good. It only knows whether your channel has a coherent audience pattern it can slot you into.

The invisibility problem

This is where most struggling channels actually live, and almost none of them know it.

If your uploads scatter across unrelated topics, the people who watch you have nothing in common with each other. Your channel has no stable neighborhood in taste space. The co-visitation logic literally cannot find a reliable pool of viewers to test you in. Engineers would call your videos null candidates: not suppressed, not penalized, not shadowbanned. Just never considered.

Ryan Trahan is the canonical case. Before he became a 23-million-subscriber giant, his channel stalled for years. He was making Emma Chamberlain-style vlogs, Cody Ko-style commentary, and athlete content, all on one channel. Three unrelated audiences. No fingerprint. In his own words, he had “pivoted so many different directions.” The videos weren’t bad. The system just couldn’t figure out whose feed he belonged in, so he wasn’t in anyone’s.

Then he committed to one format: narrative budget and survival challenges. One audience, every upload pointing the same direction. Same algorithm, same creator. The data changed, and the channel exploded.

The diagnostic here is simple. If your videos barely get impressions at all, you don’t have a thumbnail problem or a retention problem. Nobody is seeing your thumbnail. You have a coherence problem, and no amount of packaging polish fixes it. Ten consistent uploads start telling the system a story. Fifty give it a reliable model of exactly who you’re for.

(This is something I also had to learn the hard way; I started a separate channel for all my storytelling videos)

Stage 2: The interview

The few hundred survivors go to the ranking stage, where a much heavier model scores each one using hundreds of signals: your entire history with that channel, predicted watch time, predicted satisfaction, device, time of day.

Three details from Google’s published research are worth knowing because they kill popular myths:

The ranker doesn’t optimize for clicks, and hasn’t for a decade. The training math weights positive examples by how long people actually watched, so the model’s output approximates expected watch time, not click probability. Clickbait that doesn’t hold viewers is fighting the literal training objective.

Engagement and satisfaction are predicted separately. The modern architecture scores “will they click and watch” and “will they be glad they did” as distinct tasks, then combines them. You cannot buy durable distribution with engagement alone, because dissatisfaction is modeled independently and counts against you.

Ignored impressions are spent, not free. If your video was shown to someone and they scrolled past, the system quietly demotes it for that person going forward. This is the mechanical reason weak packaging slowly strangles a video, and the reason changing a failing title and thumbnail can genuinely restart one. YouTube never fully stops testing a video, so a flop is a draft, not a verdict.

The testing loop (or: why videos explode on day 3)

Every creator has watched a video do nothing for three days and then suddenly take off, and everyone calls it luck. It’s not luck. It’s a staged experiment, and you can now read the stages.

The system permanently balances two impulses: exploit (serve proven videos, the safe bet) and explore (spend a small slice of traffic testing unproven ones, or the platform calcifies). Your new upload gets the explore treatment:

1. Before any data exists, the system forms a prediction for your video based on your channel’s recent history. Your last 10 to 20 videos set the bar.

2. It seeds the video to a small test audience whose history says they should like it.

3. It measures performance against the prediction. Not against your niche. Not against bigger channels. Against you.

4. Beat the prediction, and allocation expands to a larger, slightly colder audience. Then it measures again, and expands again. Each ring takes time sometimes 48 hours, sometimes a week. That’s the whole day-3 mystery.

So when you are reviewing data for a video, make sure to wait 3 days after publishing the video; in the meantime, let YouTube do its thing.

^{Also side note, if you are interested in knowing how brands choose which creators to work with i highly recommend you check out this podcast, linked 👇}

^PODCAST

The baseline trap

That fourth point hides the most counterintuitive fact in all of YouTube: the bar your video has to clear is your own trailing average. This produces two paradoxes that confuse almost everyone.

First, big channels can be punished by their own history. CaptainSparklez has 11.4 million subscribers and has been on YouTube since 2010, yet gets outperformed per video by Forge Labs at 6 million. Most of those 11 million subscribed in a different era, and when the system tests new uploads on them, they scroll past — which tanks his click-through rate against his own baseline and teaches the system to pull back. Inactive subscribers aren’t neutral. They actively generate negative evidence.

Second, small engaged channels have a hair trigger. A lower baseline means overperformance is easier to register, so the expansion loop fires faster and more often. This is the real mechanism behind every “how did this 50K channel outgrow a 5M channel” story.

Two things follow, and I’d tattoo them on every creator dashboard if I could. Comparing your views to bigger channels carries zero information; your only competition is your own recent past. And a bad upload doesn’t just get low views it lowers the prediction your next video is tested against. Quality control isn’t vanity. It’s baseline protection. (Also commonly known as the 1of10 Theory)

What I’m doing with this as a small creator

Three changes, all directly downstream of the machinery:

1. Coherence before polish. Every upload should sharpen the answer to “who is this channel for,” not blur it. If I want to experiment, I’ll do it deliberately and occasionally, not constantly.

2. A baseline sheet. Last 20 videos: CTR, average view duration, first-7-day views. The averages are the literal bar my next upload gets tested against, so I should know what they are.

3. No mourning, repackaging. Any video where the few people who clicked stayed a long time gets a new title and thumbnail instead of a funeral. At this point I just hired a thumbnail editor to help me with my thumbnails. In the game of YouTube, we need to learn that there are people who just know something better; take advantage of it.

The one-sentence version

Your video passes through two gates a bouncer that decides whether you’re considered at all, and an interview that decides how far you go and most channels that think they’re failing the interview never got past the bouncer.

Next in this series: the metrics layer how to read a retention curve like an analyst, what the satisfaction surveys actually measure, and why YouTube would rather you leave the app happy than stay on it miserable.

Here’s the thing: what you’re describing doesn’t exist. And the mental model you replace it with will do more for your channel than any tactic, tool, or upload schedule. Let me give you that model in four ideas.

(There’s a one-page diagram of all four at the bottom of this post you can save it.)

Idea 1: There is no push. There is only pull.

Creators imagine the algorithm as a distributor: you upload, and somewhere in California a machine decides whether to push your video out to the masses.

YouTube’s engineers have said directly that this is backwards. Nothing is pushed. Instead, every time a viewer opens the app, the system builds a feed for that person by asking one question: of everything available, what would this specific human most want right now?

Todd Beaupré, who runs Growth & Discovery at YouTube, describes the system as “automating word of mouth.” It’s not a broadcaster deciding what to air. It’s two billion personal shoppers, each working for one viewer.

This inversion changes the question you should be asking. “Why isn’t the algorithm promoting me?” has no answer, because nothing gets promoted. The real question is: for which specific viewer, in which specific moment, is my video the single best answer, and does the system have enough evidence to know that? Every diagnosis of a struggling channel is some version of that question.

Idea 2: “The algorithm” is plural

There is no single algorithm. Home, Suggested, Search, Shorts, Trending, notifications, and the TV app each run separate systems with separate definitions of success:

• Home asks: what does this person want before they’ve expressed any intent?

• Suggested asks: given what they’re watching right now, what extends this session?

• Search asks: does this match what they typed and do searchers who pick it stay?

• Shorts asks one brutal question per swipe: watched, or swiped away?

• The TV app rewards long, leaned-back sessions on the biggest screen in the house.

Even the signals get reweighted by context. Beaupré’s example: watch time matters a lot for a podcast on a TV, someone leaving a 2-hour episode after 4 minutes is a bad sign. But someone leaving a song after 30 seconds? They probably just wanted the chorus. Same behaviour, opposite meaning, and the system knows the difference.

So your video doesn’t have one algorithmic fate. It has one per surface, per audience, per device. It can be dying on Home while quietly compounding in Search for years. This is why your traffic-source report isn’t trivia, it tells you which of several different games you’re actually winning.

_{*I know this amount of information can be overwhelming; that is why I have attached a diagram down below for all visual learners :)}

Subscribe now

Idea 3: Everything reduces to two predictions

Strip away all the machinery, and every time YouTube considers showing your video to someone, it’s making two estimates:

1. Will this person click? Given this viewer, this thumbnail and title, this surface, and everything else on screen competing for the same tap. Your visible proxy in Studio: click-through rate.

2. If they click, will they be glad they did? Not just “will they watch long,” but did the watch turn out to be worth it, measured by retention plus satisfaction signals you never see (surveys, likes, whether they kept enjoying YouTube afterwards). Your visible proxies: average view duration and the retention curve.

The system effectively multiplies the two. And that multiplication produces a four-quadrant map that explains every result you’ve ever had:

• High CTR × low retention → the broken promise. Your packaging writes a check the video doesn’t cash. People click, feel misled, and bail. Distribution dies right after the initial test.

• Low CTR × high retention → right video, wrong wrapper. The few who click, love it. The content works; the packaging doesn’t. This is the single most fixable failure on YouTube — re-title it, re-thumbnail it, and let the system re-test. (This was/is a common problem i face)

• Low × low → wrong idea (or wrong audience). No repackaging will save it. Take the lesson, not the L.

• High × high → expansion. Both predictions beat expectations, and the system widens the audience in waves.

The next time a video underperforms, don’t ask “why does the algorithm hate me?” Ask “which of the two predictions failed, and on which surface?” One question is a mood. The other is a diagnosis.

Idea 4: The algorithm is a mirror with a multiplier on it

The last idea is the hardest to internalize because it’s about math, not mechanics.

The system has no opinion of you. It amplifies whatever evidence you feed it, and the amplification is wildly nonlinear. Paddy Galloway, the strategist behind clients doing ~10 billion views a year, tells a story about a video stuck at 1,000 views a day. His team suggested one thumbnail change. Humans rating the two side by side would call the new one maybe 30% better. The video went to 40,000 views a day.

Not 30% better results from a 30% better input. Forty times. Because a small edge in predicted performance compounds through every round of the system’s testing loop, small improvements in the right places (the idea, the title, the thumbnail, the first 30 seconds) produce absurd differences in outcomes. Which means the highest-leverage hour in your week probably isn’t the tenth hour of editing. It’s the hour spent making the promise sharper.

The whole model in one sentence

YouTube doesn’t judge videos, it matches them one system per surface, two predictions per impression, and a nonlinear multiplier on whatever evidence you hand it.

Stop performing for a gatekeeper that doesn’t exist. Start feeding evidence to a machine that’s desperate to find your audience for you.

_{Next in this series: the actual machine, how billions of videos get narrowed to a few hundred candidates before ranking even begins, and why most "algorithm problems" are really invisibility problems.}

Leave a comment

Share

Then a 26-year-old from Alabama makes a horror movie in his parents’ hometown for less than the cost of a luxury SUV, premieres it at TIFF, sells it for $15 million (breaking the festival’s all-time record), and watches it gross over $148 million worldwide.

Production budget for Obsession: $750K

Worldwide box office gross: $148M

Days to film the entire movie: 20

Who is Curry Barker and why does it matter?

Before Obsession, Curry Barker was one half of the YouTube sketch comedy channel That’s A Bad Idea, with around 1.2 million subscribers. He wasn’t a nobody, but he wasn’t a household name. He was a creator the kind of person traditional Hollywood would’ve told to get an MFA, intern at CAA for three years, and maybe get a shot at directing a mid-budget thriller in his late 30s if he was lucky.

Instead, Barker made a found-footage short called Milk & Serial in 2024 with an $800 budget. Eight hundred dollars. It blew up. Then came The Chair, another short that started circulating in industry circles. Hollywood wasn't paying attention to his subscriber count it was paying attention to his craft. His ability to generate tension, get real performances, and make $800 feel like $800,000.

The Barker origin story
Barker has said he's "never had a budget in my life." His entire filmmaking education came from making videos on the internet not film school, not an internship, not a development deal. The constraints that would've disqualified him in the old system became his competitive advantage: lean budgets, scrappy aesthetics, and an audience already pre-sold on his sensibility.

After Obsession premiered at TIFF in September 2025, studios entered a flat-out bidding war. Focus Features won, paying $15 million the highest acquisition price in TIFF history. Barker then signed with UTA. A deal with Blumhouse followed. Then an offer for his next project reportedly hit $10 million before he'd even pitched an idea. He also signed with A24 to direct a Texas Chainsaw Massacre reboot.

“That kind of money used to be reserved for proven franchise directors. Now, it's being floated for a creator who built a loyal online audience.”

This isn’t a one-off. It’s a movement.

Barker’s rise is striking precisely because it isn’t isolated. In February 2026, Markiplier yes, the YouTube gamer with 38 million subscribers released Iron Lung, a self-financed indie horror adaptation. Major box office success. Then came Backrooms, directed by Kane Parsons, a 20-year-old who started posting Minecraft Let’s Plays at age 9. It opened through A24 to $81 million the biggest opening in A24’s history.

And it goes further back than that. David Sandberg, the first YouTuber to break into Hollywood horror, made a viral short called “Lights Out” that became a 2016 feature which grossed nearly $150 million on a $5 million budget. James Wan, one of Hollywood’s most commercially successful horror directors, has called YouTube “the perfect incubator for emerging voices.”

2016

David Sandberg — Lights Out. First YouTuber to break into Hollywood horror. $150M gross on a $5M budget. Opened the door.

2024

Curry Barker — Milk & Serial, an $800 found-footage short, goes viral. Industry starts paying attention to the kid with the sketch comedy channel.

2025

Curry Barker — Obsession premieres at TIFF. $ 750 K budget. $15M acquisition. Record-breaking. $148M worldwide and counting.

2026

Markiplier + Kane Parsons — Iron Lung & Backrooms. Two more YouTubers shatter box office expectations. A24’s biggest opening ever goes to a 20-year-old creator.

What Hollywood has that creators don’t and vice versa

This isn’t a story about creators being better than Hollywood. It’s a story about what each brings to the table, and how the scales are shifting. This is the real change (especially for small creators who are stuck)

Old Hollywood model

• Gatekeepers at every stage

• Enormous development timelines

• Franchise-dependent revenue

• Audience tested after the fact

• Prestige as the entry ticket

Creator model

• Built-in audience before day one

• Ruthless iteration on short-form

• Low-budget forces creative discipline

• Authenticity as the entry ticket

• Distribution solved before production

The thing creators have that Hollywood desperately wants is a pre-sold audience and a gut-level understanding of what makes people click, watch, and share. Barker didn’t learn tension from film school he learned it from watching which thumbnails his audience clicked on. That instinct is worth more than any development deal.

What this means if you’re a creator right now

The lesson isn’t “go make a horror film.” The lesson is structural: the gatekeepers no longer hold the keys. Distribution is democratised. Audiences are findable online before you ever set foot in a studio. Your body of work on YouTube, TikTok, or Substack is your portfolio, your proof of concept, and your audience-building engine all at once.

Barker didn’t cold-email a producer. He made something undeniable on the internet until the industry had no choice but to come to him. That playbook works in film. It works in music. It works in publishing. And increasingly, it works in every corner of entertainment that used to require a permission slip from someone with a corner office.

The gatekeepers no longer hold the keys. Your body of work on the internet is your portfolio, your proof of concept, and your audience-building engine all at once.

Hollywood isn’t dying. But it is being forced to confront the fact that talent no longer needs it the way it once did. Studios are now competing to sign creators, not the other way around. One of them reportedly offered a 26-year-old $10 million before he’d even walked into the room with an idea.

· · ·

The creator economy was always going to get here. The internet gave anyone with a camera the ability to build an audience. The next logical step was always that those audiences would follow their creators wherever they went into movie theatres, onto bookshelves, down fashion runways. Hollywood just didn’t think it would happen this fast.

Curry Barker made a movie for $750,000. It’s made $148 million. He filmed it in 20 days. He’s 26 years old. He learned everything he knows by making videos online.

This is the change.

• My podcast

• My main channel

• Short-form clips

Just one secondary channel.
One lane.
Enough volume to stop speculating.

Same creator.
Same brain.
Same doubts.

I didn’t do this to prove discipline.
I didn’t do it to look consistent.

I did it because I was tired of guessing.

What I thought would happen

Going in, I had a quiet assumption in the back of my mind.

I assumed that if I posted enough:

• My ideas would naturally get better

• Viral moments would eventually be forced

• Growth would accelerate just by showing up

In other words, I believed volume alone would solve the problem.

It didn’t.

What it did do was far more uncomfortable, it exposed exactly where I was wrong.

Subscribe now

Volume doesn’t create results. It creates visibility.

When you only post occasionally, every video feels like an event.

You protect ideas.
You overthink execution.
You attach your identity to outcomes.

At low volume, it’s easy to believe:

• “The algorithm didn’t push it”

• “The timing was off”

• “People just didn’t see it”

Volume kills those excuses.

Once I reached a certain number of uploads, the same failures started recurring. Not randomly predictable.

And that’s when something clicked.

Success isn’t random.
And failure isn’t either.

What repetition actually revealed

1. Most ideas aren’t bad they’re unclear

This was the biggest wake-up call.

Posting at scale showed me that most underperforming videos didn’t fail because:

• The topic was boring

• The editing was bad

• The effort was low

They failed because the promise wasn’t obvious.

The audience couldn’t immediately answer:

• What is this?

• Why should I care?

• Where is this going?

I learned that clarity beats creativity more often than not.

As a creator, your job isn’t just to have an idea, it’s to frame it so the audience can follow along without friction.

When that framing is missing, no amount of effort saves the video.

Subscribe now

2. Effort and performance are loosely correlated

Some of the videos I spent the least time overthinking outperformed the ones I obsessed over.

That was frustrating at first.

Then it was freeing.

Because it showed me something important:

The audience doesn’t reward effort.
They reward clarity and momentum.

Volume compresses that lesson fast.

When you see it enough times, you stop romanticising the process and start respecting what actually moves the needle.

3. Patterns emerge only when ego leaves

Early on, every video felt personal.

If it flopped, it felt like a reflection of me.
If it did well, it felt validating.

After enough uploads, that emotional attachment faded.

Videos stopped feeling like statements.
They started feeling like iterations.

That shift changed everything.

Once ego stepped out of the way:

• I iterated faster

• I killed weak ideas sooner

• I paid attention to patterns instead of feelings

Volume didn’t make me confident.
It made me accurate.

Share

What actually changed in my thinking

Before posting at scale:

• I waited for motivation

• I treated ideas as precious

• I looked for “the one” video

After 117 uploads:

• Motivation became irrelevant

• Ideas became disposable

• Progress became mechanical

I stopped asking:

“Is this good enough?”

And started asking:

“Is this clear enough?”

That question alone raised the floor of my content more than any tactic ever did.

The biggest takeaway

117 videos didn’t give me answers.

They gave me familiarity.

Familiarity with:

• My own patterns

• The audience’s behaviour

• What consistently works

• What consistently doesn’t

And familiarity cleared the noise.

Once you’ve lived through enough reps, you stop needing reassurance.
You stop chasing hacks.
You stop reacting emotionally to every data point.

You trust the process not because it sounds good, but because you’ve seen it work (and fail) enough times to believe it.

Where I’m at now

I’m still early.
The numbers aren’t insane.
The story isn’t finished.

But I’m no longer guessing.

And that alone made the reps worth it.

Volume didn’t make me successful.
It made it clear.

And clarity compounds.

Leave a comment

Best,

Sid

Share SidParikh’s Substack

For the next month and a half, my second channel will become a controlled YouTube experiment, a high-volume data stress test.

I’m publishing one long-form video every single day until the end of the year. Not because it’s romantic or motivational, but because I want the numbers. I want the data. I want visibility into what actually moves a small creator toward the top.

(This is exciting because there was a time where i would be stressing heavily over this but at this point i choose not to lol.)

Anyways,

Every creator talks about “figuring out the algorithm.”
Cool. I’m treating it like a math problem.

Over the next 45 days, I’ll be tracking in detail the following KPIs:

Here is the link so you guys can see this too: 2ND CHANNEL DATA

1. Subscriber Delta

• Daily subs gained

• Daily subs lost

• Net change

• Which videos directly correlate to spikes or dips

This alone will tell me how sensitive my channel is to upload volume, packaging, and topic choice.

2. AVD (Average View Duration)

This is the heartbeat of long-form.

I’ll be calculating:

• Raw AVD (minutes watched)

• AVD % per video

• AVD % across the entire channel

• AVD drop-off points (using audience retention graphs)

If a 7-minute video is losing 40% of viewers at the two-minute mark, I need to know why and how fast it happens.

3. APV (Average Percentage Viewed)

This is the most underrated metric for commentary channels.
If viewers only watch 20–30%, the video is dead on arrival.

My goal is to identify:

• What topics pull 50%+

• What edits increase rewatch loops

• What intros kill retention instantly

4. Click-Through Rate

I’m tracking:

• CTR per thumbnail style

• CTR drop when I post daily vs. spaced-out uploads

• CTR by topic category

I want to see how many impressions I burn through with daily quantity and whether that impression burn actually accelerates the channel or stalls it. Now this is something i have been having a hard time getting right or getting it above of an average 1.86% (ABSOLUTELY VILE!, I KNOW!!)

5. Traffic Source Breakdown

I’m monitoring:

• Suggested vs. browse

• YouTube Search

• Shorts feed cross-pollination

• External clicks from IG and Discord

The goal:
Find the most leveraged traffic source for a channel of my size and also ofcourse i will be including the device type, the new viewers, casual viewers and regular viewers as well, so lets see.

Why I’m Prioritising Quantity Over Quality (Intentionally)

Everyone loves to say “quality over quantity.”
But early-stage creators don’t need that advice, I feel like they need data density.

Quantity gives you:

• More feedback loops

• More retention curves

• More packaging tests

• More topic validations

• More opportunities to discover what actually works

Quality matters, but right now, quantity > quality.
I want to know exactly how many videos I can ship before I hit a breaking point, and how my editing system adapts under stress.

This is me building the ultimate content operating system.

Where This Fits Into My Youtube Journey

My long-term aim is straightforward:
I want to be one of the biggest YouTubers on the planet.

To do that, I’m building two specific verticals:

1. Main Channel is more for storytelling, vlogs, cinematic challenges

2. Second Channel is more commentary, fast-turnaround uploads, testing ground

The second channel is my laboratory.
The main channel is the flagship product.

By pushing the second channel to daily uploads, I’m accelerating my learning curve by months, maybe even years. Every metric collected serves the future of both channels.

This isn’t about going viral tomorrow.
This is about creating a repeatable, scalable system that supports a full media brand.

What Happens After December 31?

By January 1st, I expect to have:

• A full data map of my retention patterns

• A list of my top packaging styles by performance

• A ranking of topics with the highest ROI

• An understanding of how many videos per week I can sustainably produce

• A content system I can scale with editors and a team

If this experiment works, I’ll know exactly how to structure 2026.
If it crashes and burns, I’ll still walk away with clarity.

Either way, it’s a win.

Most creators talk about being “all in.”
I’d rather show the numbers.

The next month and a half is about getting dialled in, gathering data, and making it impossible not to get something out of it.

If you want to follow the experiment in real time, strap in, I’m uploading daily.

Let’s see what the data says.

You guys can check my daily uploads on my Youtube Channel

Talk soon,

Sid

The Before

Old title: “These paranormal videos should not exist.”
Old thumbnail: My face + a still image of what I was reacting to.

(Yikes….)

It was bad. But it also wasn’t doing any emotional heavy lifting.

The problem?
It told what the video was, not why it mattered.
It looked like every other reaction video, predictable, unmemorable, low emotional tension.

The After

New title: “I finally understand why people believe in ghosts.”
New thumbnail: Just the image of the moment, no face.

This subtle swap did three huge things:

1. It reframed the story. The old version said, “Watch these scary clips.” The new one said, “Watch me change my mind.”

2. It humanised the reaction. There’s vulnerability in “I finally understand…” It triggers curiosity to what changed?

3. It replaced explanation with emotion. Instead of information, it sells transformation.

And here’s the kicker: by removing my face, I made the mystery louder.
The viewer’s eye now lands on the “moment,” not the creator. The intrigue sells itself.

The Results

In a week, the numbers spoke for themselves:

• Views: 169 (up 69 from my average)

• Watch time: 8.7 hours (+5.7 more than usual)

• Subscribers gained: +4

• CTR: 2.8% — up from my usual 0.8–1.1%

That’s almost a 3x improvement in click-through rate (small channel things), without changing a single frame of the actual video.
Same content. Different packaging. Entirely new trajectory.

Subscribe now

Will I Be Doing This for Every Video?

Maybe, but not for all of them.

I’m starting to realise that when it comes to thumbnails and titles, there’s no universal formula. What works for one video might flop for another. Especially with reaction videos, tone and context change everything, horror, comedy, internet weirdness, or Reddit content all trigger different viewer emotions.

So no, I won’t just copy-paste this style going forward.
But I will keep experimenting.
Every upload is data. Every thumbnail and title is a test.

The goal is pattern recognition.

What I’ve Learned About Thumbnails and Titles

There are a few standard “rules” YouTubers follow, and now that I’m seeing the results firsthand, they make complete sense.

1. The Title and Thumbnail Should Complement Each Other.
   If your title says it, your thumbnail shouldn’t.
   If your thumbnail shows it, your title shouldn’t.
   Together, they should create one complete thought that hooks curiosity.

   
   Example: If the title says “I finally understand why people believe in ghosts,” the thumbnail shouldn’t repeat “believe in ghosts.” It should show why, maybe a haunting moment frozen mid-frame that completes the sentence visually.

2. The Title and Thumbnail Should Come Before the Video Idea.
   This is the important part and takeaway.
   Most creators (my old self included) come up with a video idea first, and then try to make a title and thumbnail fit. But the pros flip that process.
   They start by asking:

“Would this thumbnail make sense?”
“Would this title make someone stop scrolling?”
If yes, then the idea is worth filming. If no, it’s dead on arrival.

The average YouTube CTR for small creators sits between 0.8%–1.5%. That’s the dead zone. It means YouTube barely tests your content outside your immediate audience. You’re shouting into the void.

When my CTR jumped from 1% → 2.8%, I didn’t just get more clicks —
I got algorithmic validation.
That tiny number shift told YouTube, “People are interested in this show, show it to more of them.” That’s when you realise this platform doesn’t reward effort, it rewards packaging that earns attention.

Share

It makes so much sense now why creators like MrBeast, Ryan Trahan, and so many others spend weeks just perfecting thumbnails before they even hit record. Because at their level, they already know what I’m just starting to fully understand the thumbnail and title are the video.

They’re not afterthoughts. They’re strategy.

That one image and one sentence decide whether millions of people even see your work.
You can have the best storytelling, pacing, and editing in the world but if your packaging doesn’t earn the click, none of it matters.

And that’s what I’m understanding as a small creator:
I’m not just trying to make better videos anymore, I’m trying to make videos that deserve to be clicked on.

Takeaway: The best content doesn’t win. The best-packaged content does.

Talk Soon,

Sid

Now with this shift, Youtube creators such as KSI are doubling down on YouTube. KSI built a career spanning gaming, comedy, boxing and music, but he now describes YouTube as the “heart, the baby, the core” of his empire. He has rebranded channels and released multiple big‑budget videos this year, signalling that he sees YouTube not as a stepping stone but as the main stage (he said this himself as well). This blog post unpacks the data behind YouTube’s mainstream ascendancy, examines KSI’s recent YouTube output and shows why more creators are making YouTube their primary media outlet.

YouTube’s Mainstream Metrics

A number of recent studies highlight how far YouTube has penetrated mainstream viewing:

• Dominant share of TV time: Nielsen’s Gauge report finds that in May 2025, YouTube held 12.5% of total television viewing, compared with 7.5% for Netflix. The platform’s share has increased for four consecutive months and is now the highest of any individual streamernielsen.com.

• Time spent vs. reach: TVision’s State of Streaming 2025 report shows that while Netflix had the highest household reach (71%), YouTube had the highest share of viewing time (21%) and that viewers spent more time with YouTube than any other app. Two‑thirds of U.S. households watched both Netflix and YouTube in 2024info.tvisioninsights.com.

• Ad market and revenue: According to the same TVision study, Hulu and YouTube captured the largest share of streaming ad time in 2024, with YouTube accounting for 11% and Hulu for 13%. MoffettNathanson estimates that YouTube generated $54 billion in revenue in 2024, with operating income just under $8 billion. Analysts project that YouTube’s revenue will soon surpass Disney, prompting them to call it the “new king of all media”ekathimerini.com.

• TV viewing dominance: A New York Times analysis reports that YouTube and Netflix together accounted for 20 % of U.S. TV viewing time in May 2025, but YouTube’s share was 12.5 % compared with Netflix’s 7.5 %, and YouTube’s lead over Netflix has widened from half a percentage point to five points in two years. On average, 7 million viewers watch YouTube on a TV set at any moment, compared with 4.7 million for Netflix.

These metrics make it SO clear that YouTube is no longer the scrappy upstart. It is a dominant, advertising‑supported platform that commands more viewing time on television sets than any other streamer. This environment enables creators like KSI to reach audiences at the scale of mainstream broadcasters.

KSI is one of the most diversified creators in the world, with ventures in music, boxing, beverages and entrepreneurship. In a June 2025 interview with Square Mile, he reflected on his relationship with YouTube. In 2016–2017, he felt he had hit the “ceiling of YouTube,” but he now sees “no real ceiling” because “YouTube is the most viewed content in the world, way more than TV, way more than Netflix”squaremile.com. He emphasised that although he pursues music and boxing, YouTube remains the “heart, the baby, the core” of his career and that he must “never forget about [it]”squaremile.com.

This strategic realignment has tangible manifestations being,

Main‑channel rebrand and subscriber base

• KSI Music (main channel): In 2024–2025, KSI rebranded his main YouTube channel as KSI Music to separate his music releases from general content. According to VidIQ data (updated 15 October 2025), the channel has 24.9 million subscribers and over 6.26 billion total views.

• KSI (formerly JJ Olatunji): His second channel (daily content, gaming and vlogs) retains 16.7 million subscribers and over 4.57 billion socialblade.com.

Why this matters (especially NOW)

KSI’s renewed dedication to YouTube is both a response to and a validation of the platform’s mainstream status. Several takeaways emerge:

1. Creators recognise YouTube as mainstream television. KSI’s remark that YouTube outperforms TV and Netflix, as seen on squaremile.com, reflects the data agencies have quantified: YouTube commands more television viewing time than any other service, according to ekathimerini.com. Creators are therefore allocating TV‑level budgets and production values to their YouTube content.

2. Platform diversification is a hedge, not a replacement. KSI still pursues music, boxing and business ventures, but he views YouTube as the foundation that sustains all other projects. This is in alignment with other creators who see YouTube memberships, brand collaborations and merchandise as integral revenue streams.

3. Advertisers follow the eyeballs. With YouTube capturing 11 % of streaming ad time and delivering billions in revenue, brands are investing heavily in the platform. KSI’s channels leverage sponsorships and product placements (e.g., Prime Hydration) to monetise audience attention.

4. YouTube’s open ecosystem fosters creator autonomy. Unlike Netflix, where content is greenlit and financed by a central gatekeeper, YouTube allows creators to own their content and control upload cadence. KSI’s ability to drop a single, produce behind‑the‑scenes footage and schedule Sidemen shoots illustrates this flexibility.

5. The line between creator and mainstream celebrity is blurring. KSI’s career shows that being a YouTuber no longer precludes mainstream recognition. He is simultaneously a BRIT‑nominated musician, CEO of Misfits Boxing and a mainstream entertainer. As the New York Times notes, even Netflix is signing YouTube talent such as the Sidemen for crossover projects.

The data leaves no doubt: YouTube is now part of the television mainstream, not just a social media platform. When 12 % of U.S. TV viewing time goes to YouTube, and viewers spend more time watching YouTube than any other app, creators ignore the platform at their peril. KSI’s strategy of rebranding channels, investing in high‑production videos and anchoring himself in weekly Sidemen episodes is literally showing us that there is a broader movement where YouTubers treat their channels like broadcast networks.

For aspiring creators and marketers, the message is clear: think of YouTube as a primary distribution channel. Whether you’re dropping a podcast, a product review or a music video, the platform’s reach rivals Netflix and surpasses traditional TV(ekathimerini.com). You need to be able to use cross‑format storytelling, invest in quality, and build community because the future of mainstream entertainment is being forged right now on YouTube. I mean, personally, this is quite revolutionary because now that YouTube is getting more mainstream, it is going to get more saturated continuously. Now, i write this not to scare you but to encourage you to start when you can because there is never a wrong time to begin this journey.

Talk Soon,

Sid

When YouTuber George Clarke announced he was joining Strictly Come Dancing, his post exploded, with over 2 million views, making it the most-engaged reveal of the season.

For context, that’s more engagement than Strictly’s official announcement posts for other contestants.

That one moment says a lot.
Because George didn’t just get cast, he brought his entire audience with him.

Creators Are the New Ratings

For years, Strictly has been one of the BBC’s crown jewels, averaging 10.7 million viewers per episode during its 2020 peak.
But like most traditional TV, numbers have been slowly sliding down to around 8.5 million in 2023.

Now, with George’s involvement, Strictly might finally be reversing that trend.
His following on YouTube and social media gives the show something money can’t buy: attention from the digital generation.

Why TV Needs Creators

Mainstream media has finally realized what the internet already knew: creators drive culture.
They’re not just influencers; they’re distributors.
When George posts, millions listen.
And that organic reach is something even a national broadcaster struggles to match.

This isn’t a coincidence, it’s a strategy.
Casting creators is the new marketing plan.
They bring their audience, their credibility, and their built-in hype machine.

What I’ve Learned About YouTube This Week

Watching all this made me realize something:
YouTube isn’t just a platform , it’s leverage. I personally find it crazy that more people and businesses are not taking more advantage of this!

Creators build real trust, directly with their audience.
They don’t need permission to reach millions.
And now, even billion-dollar networks are relying on that trust to stay relevant.

George didn’t need Strictly to become famous.
Strictly needed George to stay modern.

This is the power shift in full display.
TV isn’t the main stage anymore; it’s a guest spot in the creator economy.

And as more YouTubers like George Clarke step into mainstream arenas, we’re watching the future unfold:
Creators aren’t just joining the media. They are the media.

Now, this whole thing fires me up to double down on YouTube and social media.
Seeing how the industry is shifting in real time, how creators are literally shaping what people watch and talk about, reminds me why I started in the first place. I’m fired up.

Talk soon,

Sid

Leave a comment

Subscribe now

Share SidParikh’s Substack

I’m not only a creator, but I’m a full-time student, I’m working sales to fund my content, I’m streaming on Twitch to generate clips, I’m running a reaction channel that’s actually my most successful one, and I’m still pushing to build the main channel with higher-production, unique ideas.

It’s like playing five games of chess at the same time. And the block isn’t just on the “creative” side, it’s in the logistics of how all these moving pieces fit together.

(just like an editing timeline)

The Juggle

• Sales = cash flow. No sales, no content budget.

• Twitch = content engine. It feeds the clips channel.

• Reaction channel = the funnel. It’s consistent, it’s growing, and it can drive audience toward the main channel.

• Main channel = the big bet. High-production videos, unique ideas, the thing that could really set me apart.

On paper, it’s a smart ecosystem. In reality, it’s exhausting. Each one demands a different type of creativity, and sometimes that just burns out the system before I even hit “record.”

Why the Block Hits Harder

When you’re this stretched, creative block isn’t just frustrating, but it’s also expensive. A block on the main channel delays the content that could push the brand forward. A block on the reaction channel slows the funnel. A block on Twitch means fewer clips. And the irony is, the stress of needing output to fund the system makes the block worse.

Now, You Might Be Wondering…

“Why not quit one thing or the other?”

Here’s the reason I don’t. I’m 21. This is the exact time to build leverage across multiple platforms. Each piece of the puzzle is adding credibility, reach, and long-term brand value.

Of course, it’s going to be hard. Of course, it’s going to be tiring. But it’s the game. I signed up for this. I’m treating content creation like a business, like a media brand and not just as a hobby. That’s why I want to double down, dial in, and build the right systems now. Because this isn’t just about videos, it’s about putting my name out there, building credibility, and making sure people know me as a person, not just as a channel.

What I’m Sitting With

1. The reaction channel is leveraged. It’s the consistent performer, the audience builder, and the easiest way to expand reach.

2. The main channel is the vision. That’s where uniqueness lives. However, it must be sustainable; not every upload can feel like reinventing the wheel.

3. Sales keep the lights on. I can’t ignore the money engine, even though it pulls time and energy away. Until the money I make from content can replace this, we’ll see in the near future about this…

4. Twitch is functional, not primary. It exists to create clips, not become the centrepiece.

This is the paradox: the creative block I’m in right now isn’t just about ideas. It’s about the weight of trying to build an empire before the foundation feels solid.

But I’d rather carry the weight now than regret not trying later. Because the block isn’t the enemy, it’s just part of the journey, especially as a one-man army. And if I can push through it, each piece the reaction channel, the Twitch clips, the main channel, the podcast will stack into the media brand I know I’m building.

At the end of the day, creativity doesn’t just need ideas; it needs infrastructure.

Let me know your thoughts on this!

Talk soon,

Sid

Leave a comment

Share

The Law of Observation

“Great communicators are great observers first.”

For creators, this is a master key. Before you worry about fancy gear or editing tricks, study the people who already have the audience you want. Watch how they talk, how they pace their videos, how they draw viewers into their world. Ask yourself:

• What exactly did they do to connect?

• Why did it resonate?

• How was their story structured?

Viewers today crave less performance and more presence. They want to feel like they’re hanging out with someone real, not watching a showy persona.

Take Arthur TV, a UK creator who has built a devoted community by making videos that feel warm and personal, perfect for a cosy background when you’re unwinding after a long day. He isn’t shouting for attention;, he’s making you feel like a friend.

The Law of Conviction

“If it doesn’t matter to you, it won’t matter to them.”

Conviction is the spark that makes your message stick. If you don’t fully believe in what you’re saying, your audience will sense it immediately. Energy and authenticity can’t be faked.

That means acting on your principles before you talk about them. When you live your message, people feel it. A perfect example is Greg O’Gallagher of Kinobody (almost don’t want to share this one).

Every time he speaks about fitness or lifestyle, you can tell he’s entirely bought in. His certainty is magnetic, and TO KNOW YOUR CONVICTION, YOU NEED TO ACT ON it.

Final Takeaways…

Great communication isn’t just clever words or polished editing. It’s a two-step process:

1. Observe like a strategist: pay attention to what actually engages people.

2. Speak with conviction: share only what you truly stand behind.

Do those two things consistently, and your audience will lean in, not because you performed, but because you meant every word.

Talk soon,

Sid

Leave a comment

Subscribe now

Share SidParikh’s Substack

Instead of sulking, I pulled the data and dissected every mistake i made from a lifeless title to a thumbnail that basically says “guy reacts to a cartoon.”
Here’s the full autopsy and what I’m changing next.

Here’s the raw data: the video is called “TRUE Night Horror Stories.”
Not exactly a scroll-stopping headline.

One of the biggest lessons I’ve learned on YouTube so far is simple but brutal:, make it ridiculously easy for the algorithm and real people to know what your video is about. Your title is the hook, the SEO, and the first impression all in one.

With that in mind, here’s the autopsy and the change I’m implementing.

Current Title Is Too Generic

“TRUE Night Horror Stories” is bland.

• It’s not searchable enough, and people aren’t typing exactly that into YouTube.

• It doesn’t spark curiosity or urgency.

• It buries the hook, “TRUE” isn’t the emotional trigger; the creep factor is.

You Missed the Intrigue Play

The best titles either solve a question or provoke a reaction.
Examples from the TubeBuddy suggestions are decent (not the best but better than what i had):

• “3 Creepy Night Stories That Will Keep You Up!” (fear + specific number)

• “Can You Handle These True Night Horror Animations?” (direct challenge)

Your title doesn’t create that “I have to click” tension. But here is the thing, though, with the rise of AI, spotting an AI title is easier than ever before. Use this carefully, don’t be rash with this. What i’ll be doing in this specific situation is, I’ll be taking inspiration from the AI titles and adding my own twist on it.

Placement & Pacing

• Important words need to be front-loaded: “Creepy,” “Scary,” “Can You Handle” all hook the reader immediately.

• “TRUE Night Horror Stories” starts slow and reads like filler text, now that i think about it.

Thumbnail and Title Synergy

My thumbnail is solid for reaction (face + animation + “wtf?”) again not the best but we’ll get into that later in this post, but the title doesn’t amplify it.
Pairing my dramatic reaction face with a neutral headline wastes that visual energy.

Analytics Confirm It

Five views and flat watch-time = no click-through. Low CTR usually means the packaging (title/thumbnail) isn’t pulling people in.

That being said, let’s talk about the thumbnail.

Ooof…

Anyways,

Here are all 3 of my thumbnails. I mean, I know for a fact that I am not taking advantage of YouTube’s Thumbnail Test here.

Reason being, it’s apparent that I have not changed anything here but the colours in the thumbnail.

Let’s see. On all three thumbnails, we have the text ‘REAL STORIES’ written in white with a black drop shadow, which makes the text stand out. The image of the animated man is good, but there is no context for what it is, who he is, what he represents, or why he looks that way. There is no intriguing element in this thumbnail besides the “wtf?” element.

There is no Narrative Clue. The viewer can’t tell why it’s scary. There’s no hint of the actual story, no shadow, no creepy text, no danger signal. It’s plain.

• Static Visuals: It’s just my face + a cartoon. Horror thumbnails thrive on motion or implied action (a hand reaching, a door opening). Or something along those lines.

• One-Note Emotion: My reaction is surprise, but we don’t know if it’s fear, shock, or disgust. It’s vague.

Right now, the thumbnail says, “guy reacts to a cartoon.”

So, that being said here is my to-do list,

1. Change the title (so much work to do here)

2. Change the thumbnail (add more intrigue to it)

3. If im being honest, get a better video topic idea

Let me know what you guys think about this! Comment down below!

P.S. there is a correct answer for the poll above, you dont have one without the other.

Talk Soon,

Sid

Subscribe now

Leave a comment

Share

Recently, I was reading The 16 Laws of Communication by John C. Maxwell, and one idea hit me hard: credibility isn’t optional, it’s the foundation of everything you create.

As a small creator, it’s tempting to obsess over algorithms, fancy gear, or viral hooks, but none of that sticks if your audience doesn’t trust you first.
In this post, I’m breaking down the Law of Credibility and how it applies to YouTube and social media, so you can start building authority and community before your subscriber count takes off.

The Law of Credibility, to its core, is this: people believe the messenger before the message.

From a creators perspective here is the idea, consistency in uploads, ethical sponsorships and analytics building long-term trust.

There are 5 Key Qualities of Good Communication-

1. BE TRANSPARENT: Being transparent isn’t just about dumping your personal life online, it’s about strategic honesty. Viewers can sense when you’re hiding behind a persona, and the second they feel a disconnect, they swipe away. Transparency means sharing your wins and your screw-ups, explaining why you made certain creative choices, and admitting when something didn’t work.

   For a small creator, this is a growth accelerator. You don’t have the massive production value or brand clout of a big channel, but you do have the ability to be real in a way that scales intimacy. Talk about the late-night edits, the gear you wish you had, the nerves before hitting “publish.” That kind of honesty builds trust and makes every subscriber feel like a stakeholder in your journey.

   Practical moves:

   • Narrate the process: Include a quick behind-the-scenes clip or voiceover about what almost went wrong.

   • Own your learning curve: Share analytics experiments, failed thumbnails, or format pivots.

   • Engage in comments: Respond authentically, even a simple “you’re right, I missed that” shows you’re human.

   Transparency isn’t oversharing, it’s an invitation. It signals, I’m building this with you, which is precisely what keeps early viewers coming back and converts casual watchers into a loyal community.

2. BE CONSISTENT: On YouTube, consistency isn’t just a buzzword, it’s one of the most essential things that builds credibility. Every upload is a public promise. When you keep posting, you train viewers (and the algorithm) to trust you’ll show up. Over time, that reliability becomes your reputation.

   Think of MrBeast: before he was giving away islands, he uploaded hundreds of videos with almost no traction. He famously counted to 100,000, read the dictionary, and tested ideas relentlessly. Those early uploads weren’t about instant virality, they were about proving to himself, to the audience, and eventually to YouTube that he was committed. The quality improved, the ideas levelled up, but the real magic was the volume of work stored up before anyone cared.

   For small creators, that’s the blueprint:

   • Consistency compounds. Each upload sharpens your editing, storytelling, and on-camera comfort.

   • Credibility compounds. The more you show up, the more viewers and the algorithm see you as reliable.

   • Opportunity compounds. Sponsors, collaborations, and recommendations come to the creator who’s already in motion.

   So post often. Document, experiment, iterate. Let your upload history become your résumé. Consistency compounds, and so does credibility.

   Subscribe now

3. GOOD EXAMPLE: Relive what you’ve learned
   Don’t just announce results, walk people through the process. Break down the late-night edits, the failed thumbnails, the tweaks that finally worked. When you share the journey, viewers feel like insiders who are learning alongside you.

   Earn respect with wins and losses.
   Audiences connect when they see the whole picture, not just highlight reels. If a video flops, discuss what you misinterpreted in the title or thumbnail. If an experiment is successful, explain the strategy so that others can learn from it too. This kind of radical transparency builds credibility faster than polished success stories.

   Why it matters for small creators

   • Authenticity signals trust. People stick around for a creator they believe is real, not perfect.

   • Educational value. Viewers pick up tips and feel empowered to try their own projects.

   • Community-building. Sharing failures invites conversation and advice, turning passive watchers into active supporters.

   Practical moves:

   • End each video with a “What I learned” minute. Even if you are a reaction video, talk about what you learned from what you reacted to, if you are an entertainment talk, talk about what you did in the video that you knew, even for a minute or so.

   • Post a short follow-up in the community tab about what went right or wrong.

   • Use captions like “Behind the Scenes” or “Creator Lessons” for easy discovery. (OPTIONAL)

   When you consistently relive lessons and show both sides of the coin, you stop being just a channel and become a trusted guide that someone in your audience respects because you’re not just teaching growth, you’re living it in real time. This is something i have noticed with smaller creators who have scaled up and built hardcore superfans for example, Lab Culture’s JHawk.

These are just my thoughts as another small creator.
I’d love to know what lands for you. Did any of these Credibility tips resonate or spark an idea for your own channel?
Drop a comment or share your own experiences id’d love to know!

Talk soon,

Sid

Leave a comment

YOUTUBE

Share

Share SidParikh’s Substack

Feastables just launched chocolate milk in Walmart, 7-Eleven, and Speedway.
Made with real whole milk, Fairtrade cocoa, and vitamins, it is one step closer to ending child labour in cocoa.

But the bigger story? This is how creator IP becomes a business empire.

👉 In 2024, Feastables pulled in ~$250M in sales and over $20M in profit (Yahoo Finance) — an ~8% margin. For comparison:

• Top CPG brands average 12.2% EBIT margin (Bain & Co)

• Hershey’s net margins hover in the mid-teens (Hershey Co)

Meanwhile, MrBeast’s media business lost $80M in the same year because production costs are insane (Yahoo Finance).

The Feastables Playbook

📊 In 2024, Feastables did ~$250M in sales and over $20M in profit (Yahoo Finance), which comes out to an approximate ~8% net margin.

For context:

• Legacy CPG averages 12.2% EBIT margin (Bain & Co)

• Hershey runs in the mid-teens for net margin (Hershey Co)

Feastables is already operating at a competitive pace while still in hyper-growth mode.

Meanwhile, MrBeast’s media division (YouTube and Beast Games) generated similar revenue but lost approximately $80 million in the same year due to extreme production costs (Yahoo Finance). That’s why Feastables matters, it’s the profitable engine that powers the rest.

Scaling an IP into Stores

• Year 1 (2022): $10M revenue

• 2023–24: expanded into 20,000+ retail doors

• Raised $5M at a $100M valuation

• Targeting $60M+ run rate by 2025 (StartupBooted)

This didn’t happen by accident. Mr Beast leveraged:

• An audience of hundreds of millions → instant distribution.

• Retail partnerships (Walmart, 7-Eleven, Speedway).

• Mission-driven sourcing (Fairtrade cocoa).

The Broader Empire

• 🍫 Feastables (snacks) → $250M sales / $20M profit (~8% margin).

• 📺 Media (YouTube + Beast Games) → ~$250M sales / –$80M loss.

• 👕 Merch → $30–40M sales (20–30% margins).

• 🧸 MrBeast Lab (toys/collectibles) → $30M projected via licensing.

• 🍱 Lunchly (meal program) → 400K+ students served.

• 📊 Beast Analytics → $15M ARR.

• 💻 Digital Products → 70–90% profit margins (LearnWorlds).

The strategy: use high-margin units (Feastables, digital, software) to subsidise content that loses money but grows the brand.

Lessons for Smaller Creators

1. IP before inventory. Don’t launch a chocolate bar before you’ve built community trust.

2. Start digital. Courses, memberships, templates = 70–90% margins, almost no overhead.

3. License to scale. You don’t need to own factories, partner with established suppliers.

4. Attach a mission. Ethical sourcing, sustainability, or impact work builds loyalty.

5. Diversify like a portfolio. Don’t over-rely on one income stream. Each product is an asset class.

6. Reinvest strategically. MrBeast burns $80M+ on content because his empire can absorb it. Smaller creators should focus on profitability from the outset

Mr Beast is showing the blueprint: content drives attention, attention fuels products, products create real-world equity.

And for the rest of us? You don’t need a warehouse full of chocolate to win. You need a strategy to turn attention into assets.

So how do you actually “turn attention into assets”?

👉 Start with one digital product that solves a problem your audience already has.

• If you make career content → create a simple PDF guide or template.

• If you’re in fitness → launch a paid community or training plan.

• If you’re into entertainment → drop exclusive behind-the-scenes or merch as a limited run. (eg. Side+, The Fellas Studios Youtube Channel)

Digital products have 70–90% profit margins (LearnWorlds), almost no overhead, and are the fastest way to prove whether your audience will pay, not just watch. You will be able to survive ‘The Death Valley’ (more on this in another article).

Once you’ve validated demand → reinvest into higher-cost, lower-margin plays (merch, CPG, partnerships). That’s exactly how creators like Mr. Beast seem to be scaling: test cheap, scale proven.

Anyways, if this is something that made sense, consider checking out my YouTube Channel where I create stories, like this one…

Talk soon,

Sid

Subscribe now

Share

Leave a comment

Subscribe now