Recently, I’ve been posting short-form videos on YouTube, Instagram, and TikTok. What bothered me was how inconsistent the view counts were.

View Counts From My Last 7 Videos

Make YouTube for Money: 966 views
YouTuber Minimum Wage: 1,400 views
Hooking Principle (Long Version): 1,000 views
Hooking Principle (Improved): 1,700 views
7 Short-Form Trends: 2,100 views
5 Short-Form Trends: 963 views
Hooking Principle (Short Version): 523 views

The question was simple:

“Same channel, same creator, similar topics. Why is there a 4x difference between 500 and 2,100 views?”

At first, I thought the answer was average retention rate.

If a video is 30 seconds long, then increasing the percentage of the video that viewers watch should improve likes, shares, saves, and ultimately views.

So I focused on retention.

My conclusion was that retention comes from a strong hook. If the hook works, retention follows.

My data seemed to support that idea. Videos that successfully held viewers beyond the first three seconds consistently showed higher retention.

That was my first hypothesis.

But when I dug deeper into the data, I kept finding videos that didn’t fit this explanation.

First Discovery: The Three Conditions for High Views

After organizing the data, I found three metrics that appeared repeatedly in successful videos.

1. “Viewed vs Swiped Away” Above 40%

This appears to be the most important threshold.

During the first few days, YouTube tests a Short with a small audience.

If the “Viewed” percentage falls below roughly 40%, the video rarely receives meaningful distribution afterward.

Above 75% enters breakout territory.

Evidence

Above 40% → All Passed 100,000 Views

Uehara Market: 95.1% → 449K views
Boneless vs Bone-In Chicken: 59.5% → 332K views
Cup or Cone: 59.3% → 178K views
Seasoning Powder: 55.9% → 147K views
No Sweet No Ice: 53.2% → 203K views
Eggplant Side Dish: 52.6% → 126K views
Braised Baby Potatoes: 49.1% → 221K views
Email Address: 40.7% → 161K views

Below 40% → All Stuck Around 1,000 Views

Hooking Principle (Long): 31.1%
5 Short-Form Trends: 29.7%
Hooking Principle (Improved): 28.3%
YouTuber Minimum Wage: 24.2%
7 Short-Form Trends: 24.1%
Make YouTube for Money: 13.8%

The dividing line was almost frighteningly clear.

Forty percent seemed to separate success from failure.

2. Intro Retention Above 110%

Sometimes the retention graph rises above 100% during the first few seconds.

This means viewers are replaying the opening.

From an algorithmic perspective, this is a powerful signal.

It suggests people saw something interesting enough to watch twice.

Example

Braised Baby Potatoes

Intro retention peaked near 210%
Final result: 221K views and 838 subscribers

The first second was watched multiple times.

That alone created a massive signal.

3. Final Retention Above 50% (Estimated)

My data suggests that at least half of the remaining viewers need to stay until the end.

When all three conditions are met, views eventually arrive.

Not always immediately.

But eventually.

This is why YouTube increasingly favors long-term performance rather than instant viral spikes.

The Hidden Truth About Hooks

As I studied videos that generated millions of views, I noticed something interesting.

The strongest hooks were often not editing tricks.

They were keywords.

The first five seconds aren’t primarily about flashy editing.

They’re about whether people already care about the topic.

High first-five-second retention means:

People are interested in the topic.
The algorithm expands distribution.

Low first-five-second retention means:

Interest is weak.
Distribution shrinks.

YouTube already understands this.

The Limitation of My First Hypothesis

Then I discovered videos that satisfied most conditions but still didn’t generate significant views.

Why?

The answer was audience size.

Even excellent content has a ceiling if the total interested audience is small.

Keyword Capacity Matters

Imagine a keyword like “Japanese Women.”

Years ago, perhaps only 100,000 people were interested.

Even if every interested person watched, views would eventually stop growing.

Today, that audience might be 2 million people.

The same video could potentially reach 2 million views.

The keyword itself became larger.

I started classifying keywords by audience size.

Niche Keywords

Potential audience: ~100,000

Big Keywords

Potential audience: ~2 million

Mega Keywords

Potential audience: ~50 million

Essentially mainstream attention.

The ceiling of a video depends heavily on the ceiling of the keyword.

A niche keyword can only grow so far regardless of performance.

Second Discovery: Views and Subscribers Run on Different Engines

This was the biggest surprise.

The videos generating the most views were not necessarily generating the most subscribers.

Example

Uehara Market

449K views
+391 subscribers
0.087% conversion

Rolled Omelet

40K views
+231 subscribers
0.578% conversion

The smaller video converted subscribers at a dramatically higher rate.

Views and subscriptions are not the same game.

What Actually Creates Subscribers?

The answer appears to be psychological triggers.

Action-Based Content

Recipes, tutorials, how-to content

Subscriber conversion:
0.3%–0.6%

People think:

“I want to do this too.”

Identity & Nostalgia Content

School memories, cultural memories, personal identity

Subscriber conversion:
0.08%–0.15%

People think:

“This feels like me.”

Entertainment & Curiosity Content

Interesting but disposable content

Subscriber conversion:
0.02%–0.09%

People think:

“That was fun.”

Then they leave.

Views Formula vs Subscriber Formula

Views Engine

Keyword × Hook × Retention

Short-term game.

Subscriber Engine

Psychological Trigger

Long-term game.

They are fundamentally different systems.

The Four Scenarios of Channel Growth

Growth ultimately depends on two variables:

Keyword Strength
Editing Quality

This creates four possible outcomes.

1. Trendy Keyword + Strong Editing

Explosive views and long-term fans.

The ideal scenario.

2. Trendy Keyword + Weak Editing

Explosive views but weak loyalty.

You earn views today but must keep chasing trends tomorrow.

3. Old Keyword + Strong Editing

Slow growth followed by a breakout.

This is where true fan communities are built.

When the topic becomes relevant again, the channel is ready.

4. Old Keyword + Weak Editing

Little growth.

Little subscriber accumulation.

No meaningful long-term advantage.

The Real Strategy: Build Fans During the Off-Season

The biggest insight from all of this is that fan subscribers can still be created during slow periods.

Even when a topic is out of fashion, there are still interested people.

The audience is smaller, but it exists.

Those viewers matter.

By steadily accumulating loyal fans through strong retention and editing, a channel becomes positioned for explosive growth when the next trend cycle arrives.

Creators who only chase trends never receive this compounding effect.

Creators who build fans do.

Conclusion

Views are driven by:

Keyword × Hook × Retention

Subscribers are driven by:

Psychological Triggers

Editing strengthens both systems.

Trends rise and fall.

Keywords move in and out of popularity.

The channels that survive are the ones that continuously accumulate loyal fans through strong retention and strong content.

Those channels endure the slow seasons and explode during the busy seasons.

YouTube has become a long-term game.

Creators chasing short-term viral hits spend their careers jumping from trend to trend.

Creators who build retention and fandom grow through compounding.