Should You Only Study YouTube Competitors Your Size? We Analyzed 227 Topic Waves

Most YouTube competitor advice starts with a reasonable rule:

Study channels around your size.

If you have 20,000 subscribers, comparing yourself with a creator at 8 million subscribers can be misleading.

Their normal views are different.

Their audience is larger.

Their production resources may be different.

Their brand recognition is different.

So similar-sized channels are useful benchmarks.

But there is a second question creators often accidentally treat as the same problem:

Who should you study for video ideas?

For that question, limiting yourself to competitors your size may throw away some of the strongest evidence in the market.

We tested it.

OverseerOS analyzed 227 recurring YouTube topic waves across 213 channels and 26 niches. Every topic in the final sample had produced a million-view long-form video on at least two independent channels with current public subscriber data available.

Those 227 topics represented:

534 independent channel-topic wins.

Then we asked a simple question:

How similar in size were the channels that independently won on the same topic?

The answer was surprisingly broad.

74.0% of recurring million-view topic waves crossed at least two subscriber-size tiers.

The median topic wave spanned channels whose current subscriber counts differed by:

6.82x.

And:

• 64.8% spanned at least a 3x size difference
• 53.7% spanned at least 5x
• 41.0% spanned at least 10x
• 27.8% spanned at least 25x
• 8.8% spanned at least 100x

The pattern became even stronger when the topic had succeeded on more independent channels.

Among 50 topics with at least three independent million-view winning channels:

• 100% crossed multiple subscriber tiers
• 80.0% spanned at least a 10x current channel-size difference
• the median largest-to-smallest channel-size ratio was 35.9x

That leads to a much better competitor-research rule:

Use similar-sized competitors to benchmark performance. Use competitors of different sizes to discover and validate demand.

Those are two different jobs.

Your competitor list should reflect both.

Key Findings

• OverseerOS started with 3,983 million-view long-form videos across 891 channels and 42 niches with usable topic intelligence.
• The final analysis isolated 227 recurring niche-topic waves where the subject had produced a million-view video on at least two independent channels with current subscriber data available.
• Those waves contained 534 channel-topic wins across 213 unique channels and 26 niches.
• 74.0% of the 227 recurring topic waves included winning channels from at least two different subscriber-size tiers.
• The median largest-to-smallest current subscriber ratio inside a recurring topic wave was 6.82x.
• 64.8% of recurring topics spanned channels at least 3x apart in size.
• 41.0% spanned channels at least 10x apart.
• 8.8% spanned channels at least 100x apart.
• Among the 50 topic waves with three or more independent winning channels, every wave crossed multiple subscriber tiers, and 80.0% included channels at least 10x apart in size.
• Among the stronger three-channel-plus waves, the median size spread reached 35.9x.
• 59 of 227 topic waves, or 26.0%, included both a currently sub-500K channel and a 1M+ channel.
• A robustness test using only more-specific two-to-three-word topics still found that 65.0% crossed multiple size tiers, with a median size spread of 4.60x.
• The study does not show that subscriber size caused these videos to succeed, nor does it reconstruct every channel's subscriber count on the historical upload date. It shows that repeatable topic demand frequently exists across channels that are very different sizes today.

The Direct Answer

Should you only study YouTube competitors your size?

No.

You should study similar-sized competitors when you need a realistic performance benchmark.

But if you are researching:

• topics
• formats
• breakout ideas
• audience demand
• emerging opportunities
• repeatable content patterns

you should deliberately include channels both smaller and larger than you.

The OverseerOS data suggests a two-part competitor model:

The mistake is not studying big channels.

The mistake is using a big channel's raw performance as if it were your benchmark.

Benchmarking and Idea Discovery Are Different Problems

This distinction explains most of the confusion.

Suppose your channel normally gets:

20,000 views.

A competitor with 30,000 subscribers gets:

120,000 views

on one video.

That may be extremely informative.

Now imagine a 5-million-subscriber channel publishes a video on the same topic and gets:

1.5 million views.

Should you compare your expected views directly with 1.5 million?

Probably not.

But should you ignore that video?

Also no.

The larger channel may provide evidence about:

• topic demand
• viewer psychology
• packaging
• format
• information gaps
• broad audience appeal

Then suppose another 200,000-subscriber channel gets 900,000 views on the same underlying subject.

Now the evidence becomes much more interesting.

The strategic signal is no longer:

A giant creator got a lot of views.

It becomes:

Different channels at different scales are independently finding audience demand around the same topic.

That is what cross-size competitor research can reveal.

How We Analyzed the Data

This study uses public YouTube information collected and analyzed by OverseerOS.

The research question was:

When the same topic produces million-view long-form videos across independent YouTube channels, how similar in size are those winning channels?

Starting corpus

The starting research corpus contained:

• 3,983 million-view long-form videos
• 891 channels
• 42 niches

Each qualifying video had:

• at least 1 million recorded public views
• a known channel
• a known publication date
• a high-confidence niche classification
• usable OverseerOS topic intelligence

We analyzed topics inside niches

We did not treat broad categories as individual topics.

For example:

History

is a niche.

Roman Empire

can be a topic.

Gaming

is a niche.

Minecraft

can be a topic.

Health

is a niche.

Belly fat

can be a topic.

Topic entities were normalized and generic terms were excluded.

A topic was evaluated inside its niche so unrelated uses of the same word would not automatically become one market signal.

Each channel could only confirm a topic once

If one creator made six million-view videos about the same subject, that did not count as six independent confirmations.

For every:

niche + topic + channel

the channel counted once.

This matters because the research question is about transfer across independent channels.

We required cross-channel success

A topic entered the final analysis only when at least:

two independent channels

had produced a qualifying million-view video around it.

The final cohort contained:

• 227 recurring topic waves
• 534 channel-topic wins
• 213 unique channels
• 26 niches

The qualifying winning videos ranged from May 2011 through September 2026.

How we defined channel-size tiers

We used the latest usable public subscriber observation available to the research system and grouped channels into:

1. Under 100K
2. 100K-499K
3. 500K-999K
4. 1M-4.9M
5. 5M+

We then measured how many tiers appeared inside each recurring topic wave.

We also used a continuous metric:

largest current subscriber count / smallest current subscriber count

That second test matters because tier boundaries are arbitrary.

A channel with 490K and another with 510K fall into different tiers despite being nearly identical in size.

A 10x or 25x continuous difference is harder to dismiss as a bucket artifact.

Finding 1: Three-Quarters of Repeatable Topics Crossed Channel-Size Tiers

Among all 227 recurring topic waves:

168

contained winning channels in at least two different subscriber tiers.

That is:

74.0%.

Only 26.0% remained entirely inside one size tier.

This directly challenges the idea that useful topic evidence is confined to competitors at roughly your scale.

A topic can succeed on:

• a 200K channel
• a 900K channel
• a 4M channel

without becoming three different topics.

The audience demand can travel even when channel scale changes dramatically.

That does not mean every execution is transferable.

A smaller creator may need:

• a sharper angle
• a stronger outlier signal
• different production
• more specific packaging

But the topic itself should not be dismissed merely because the first example you find comes from a larger creator.

Finding 2: The Median Winning Topic Spanned a 6.82x Channel-Size Difference

Size-tier categories can hide nuance.

So we looked at the continuous gap between the smallest and largest current channels associated with each recurring topic.

The median was:

6.82x.

The middle 50% ranged from approximately:

2.13x to 29.53x.

The complete distribution was:

The 10x threshold is particularly useful.

Nearly:

two out of five

recurring topic waves contained winning channels whose current subscriber counts were at least an order of magnitude apart.

That is not a tiny scale difference.

A creator who restricts idea research exclusively to near-peers would potentially miss a meaningful part of the cross-channel evidence.

Finding 3: The Most Repeated Topics Were Even More Cross-Size

We then raised the evidence requirement.

Instead of requiring two independent winning channels, what happens when a topic has three or more?

The four-channel and five-channel groups are small.

They should not be overinterpreted.

But the three-channel cohort is much more useful.

There were:

50 topic waves

with at least three independent million-view winning channels.

Every one crossed more than one current subscriber tier.

And:

40 of 50

spanned at least a 10x difference in current channel size.

That tells us something important about repeatable market demand.

Once a subject has produced major results on several independent creators, looking for one "correct" competitor size becomes less useful.

You are no longer studying a creator.

You are studying a market.

Finding 4: More-Specific Topics Still Crossed Channel Sizes

One obvious criticism is that broad topics naturally appear everywhere.

If the topic is:

Minecraft

of course channels of many sizes can cover it.

So we repeated the analysis using only topic entities containing two or three words.

That produced:

60 recurring multiword topic waves.

The results became somewhat less extreme, but the central finding survived.

Even after narrowing the topic definition:

nearly two-thirds still crossed subscriber-size tiers.

That makes it much harder to explain the full result as merely broad subjects appearing everywhere.

Finding 5: Some Winning Topics Bridged Small and Million-Subscriber Channels

Among all 227 recurring topic waves:

59

contained both:

• at least one channel currently below 500K subscribers
• at least one channel currently at 1M+ subscribers

That is:

26.0%.

A smaller subset stretched even further.

Four topic waves contained both:

• a sub-100K channel
• a 1M+ channel

Two contained both:

• a sub-100K channel
• a 5M+ channel

Those final groups are small, so they are examples of possibility rather than universal patterns.

But they reinforce the larger point:

subscriber scale does not define the borders of a topic.

Audience demand can exist across radically different channel sizes.

Finding 6: The Size Effect Was Different Across Niches

We also segmented recurring topic waves by niche.

The samples become smaller here, so these numbers are descriptive rather than rules.

The important takeaway is not that animation creators should monitor channels exactly 30 times their size.

The samples are nowhere near large enough for a rule like that.

The useful finding is that the cross-size pattern is not isolated to one niche.

It appeared across:

• education
• history
• gaming
• psychology
• health
• entertainment-oriented markets

At the same time, the amount of size dispersion varied.

That is another reason a fixed universal competitor rule is weak.

Similar-Sized Competitors Are Still Valuable

None of this means size should be ignored.

If you want to know:

What does good performance look like for a channel like mine?

a 10-million-subscriber channel is usually a poor raw benchmark for a 20,000-subscriber creator.

Similar-scale channels help you understand:

• achievable view ranges
• realistic upload cadence
• production expectations
• typical audience size
• how large an outlier really is
• what growth looks like at your stage

The key is to stop asking one competitor group to solve every problem.

Think of competitor research as having separate lenses.

The Three-Layer Competitor Model

A stronger YouTube research system contains three types of competitors.

1. Peer competitors

These are channels operating around your current scale.

Use them for:

• benchmarking
• realistic comparisons
• identifying what "normal" looks like
• studying channels facing similar constraints

Question they answer:

What is working for creators in roughly my position?

2. Aspirational competitors

These are established channels significantly larger than you.

Use them for:

• mature formats
• large-market topic demand
• packaging ideas
• content systems
• strategic direction

Question:

What does this market look like when executed at a higher level?

Do not copy their raw numbers.

Study the patterns beneath those numbers.

3. Emerging competitors

These are smaller channels producing abnormal results.

Use them for:

• fresh outliers
• underexploited topics
• emerging formats
• unusually transferable ideas

Question:

What is working even without a large existing audience?

These channels can be especially valuable because an extreme outlier on a small channel is difficult to explain with subscriber scale alone.

The best competitor system uses:

peers + leaders + emerging outliers.

Why Raw Views Make Big Competitors Look Less Useful Than They Are

Suppose two videos cover a similar topic.

Channel A

Subscribers:

5,000,000

Video views:

1,500,000

Channel B

Subscribers:

80,000

Video views:

800,000

If you rank by raw views:

Channel A wins.

If you ask which result is more abnormal relative to channel scale:

Channel B suddenly becomes much more interesting.

This is why serious competitor research should normalize performance against the channel's own baseline.

A video should not be interesting merely because:

it has lots of views.

It should be interesting because:

it dramatically exceeded what is normal for that creator.

That is also why our earlier research on which YouTube competitors are actually worth studying found that repeatable outlier behavior is a stronger research signal than simply choosing the most famous channels.

The Better Question Is Not "How Big Is This Competitor?"

Ask three questions instead.

1. Do we serve overlapping viewer demand?

A creator can be your size and still be irrelevant.

If you make documentary-style AI videos and another 100K channel teaches Python coding tutorials, subscriber similarity does not make that channel strategically comparable.

Audience demand comes first.

2. Is this video actually unusual for them?

A 3-million-view video from a creator who usually gets 4 million may tell you less than a 400,000-view video from a creator who normally gets 30,000.

Relative performance matters.

3. Does the pattern exist elsewhere?

This is the strongest question.

If the same underlying demand appears among:

• a small outlier channel
• a mid-sized creator
• a large established channel

you have much stronger market evidence than one isolated viral video.

That is the pattern our 227-topic analysis captures.

Similar Size for Benchmarks, Mixed Size for Discovery

This gives us a simple decision rule.

When benchmarking your own channel

Prioritize:

similar-sized and similar-format creators.

You are trying to understand realistic performance.

When hunting for video ideas

Expand across:

all relevant channel sizes.

You are trying to discover demand.

When validating a topic

Look for:

independent success across different channels.

Cross-size confirmation is especially interesting because it weakens one explanation:

Maybe the video only worked because that creator was already huge.

When studying production strategy

Separate:

transferable mechanism

from:

resource-dependent execution.

A $50,000 documentary may not be reproducible.

Its underlying topic or title promise may still be.

The Cross-Size Topic Validation Framework

When a competitor video catches your attention, use this process.

Step 1: Establish the competitor's baseline

Before analyzing the idea, answer:

• What does this channel normally get?
• Is the candidate video really an outlier?
• How large is the outperformance?

Do not use raw views alone.

Step 2: Define the actual topic

Weak:

Business

Better:

Starting a one-person AI company

Weak:

Psychology

Better:

Why emotionally hurt people stop talking

Weak:

History

Better:

Fall of the Roman Republic

Specificity makes cross-channel validation more meaningful.

Step 3: Search across channel sizes

Do not filter the research to channels exactly like yours.

Look for:

• smaller creators
• similar-sized creators
• larger creators

Ask whether the same demand keeps generating unusual outcomes.

Step 4: Separate topic evidence from execution

If several channels win, identify what stays constant.

Maybe the constant is:

• the subject
• the audience problem
• a recurring question
• an emotional tension
• a format
• a timing event

Then identify what changes:

• title
• thumbnail
• story
• script
• length
• production style

The stable element is closer to the market signal.

Step 5: Find your own reason to enter

Cross-channel validation does not give you permission to duplicate the winner.

You still need:

• a new angle
• new information
• a different case
• better packaging
• more recent evidence
• a distinct argument
• an underserved audience variation

The goal is to validate the demand and then create an original execution.

A Simple Competitor Selection Scorecard

Before adding a channel to your competitor research set, score it on usefulness rather than fame.

A useful competitor is not always:

the channel most similar to you.

It is the channel that answers a useful strategic question.

How to Apply This With OverseerOS

The research suggests a practical workflow.

Step 1: Find competitors across the size spectrum

Use the OverseerOS Viral Channel Finder to discover breakout channels in the market.

Do not build your research list from subscriber count alone.

Look for a combination of:

• relevance
• breakout videos
• channel scale
• current activity

Step 2: Normalize each competitor before comparing them

Open promising channels in the OverseerOS YouTube Channel Analyzer.

Study:

• recent uploads
• top videos
• channel-relative winners
• titles
• publishing patterns
• video lengths

The goal is to understand what counts as unusual for that channel.

That prevents a large creator's normal video from looking more important than a smaller creator's genuine breakout.

Step 3: Look for the same demand across independent channels

Once one topic looks interesting, search beyond that creator.

If another channel of a different size produces a major win around the same demand, your confidence should increase.

If a third independent channel confirms it, you are beginning to see a market pattern rather than a one-channel anomaly.

Step 4: Turn the evidence into an original plan

Move the strongest directions into the OverseerOS Content Planner.

Save:

• the validated demand
• your original angle
• title direction
• thumbnail direction
• supporting competitor evidence

Do not save:

"Copy this 5M-subscriber creator."

The competitor is the evidence source.

Your video is still your own.

What This Means for Small Channels

If you run a small channel, the wrong conclusion is:

Big creators are irrelevant to me.

They are not.

A large creator can reveal:

• major audience demand
• mature packaging
• proven formats
• recurring content categories

The mistake is expecting their absolute numbers to transfer.

Suppose a giant channel repeatedly succeeds with:

retirement mistakes.

You do not need to predict that your version will get 2 million views.

You need to ask:

1. Does the subject also appear on smaller channels?
2. Are any of those videos outliers?
3. Is there a version specifically suited to your audience?
4. Is there an angle the bigger creator did not cover?
5. Can you package it strongly without relying on their identity?

If yes, the large channel was valuable research.

It was simply not your benchmark.

What This Means for Large Channels

The reverse mistake also happens.

Large creators can ignore tiny channels because their absolute view counts look insignificant.

But imagine:

Large competitor

4 million subscribers
800,000 views

Emerging channel

35,000 subscribers
350,000 views

The larger video has more views.

The smaller video's relative performance may contain the more unusual market signal.

That makes emerging channels useful even for established creators.

Small competitors can reveal:

• new packaging
• underserved subtopics
• fast-changing formats
• audience shifts
• ideas that have not yet reached the biggest channels

Channel size should change how you interpret a signal.

It should not determine whether you look at it.

Do Not Confuse Cross-Size Success With Guaranteed Transferability

Our finding is not:

If a topic worked on a big channel, it will work on a small channel.

That would be unsupported.

The study begins with topics that already produced multiple million-view winners.

It does not include every attempt that failed.

We therefore cannot calculate:

the probability that your video will succeed after copying a topic from a differently sized competitor.

The supported finding is:

Among topics that repeatedly produced million-view videos across independent channels, large differences in current channel size were common.

That tells us the borders of useful competitor research should be wider than one subscriber band.

It does not eliminate the need for judgment.

Why Current Channel Size Is a Limitation and a Feature of This Study

There is an important methodological caveat.

The subscriber counts used in this analysis are the latest usable public observations available to OverseerOS.

They are not reconstructed subscriber counts from the exact historical date every winning video was published.

A creator with 5 million subscribers today may have had far fewer when a 2021 video went viral.

That means this study cannot answer:

Did a topic historically move from a 100K channel to a 5M channel?

We would need reliable publication-date subscriber snapshots for that.

We deliberately do not make that claim.

But the current-size measurement still answers the practical competitor-selection question.

If you are choosing channels to research today, it shows that channels currently separated by large subscriber gaps can share the same proven topic demand.

That is the decision creators are actually making when building a competitor list.

Why the Three-Channel Result Matters Most

The two-channel sample tells us cross-size success is common.

The three-channel-plus sample makes the market signal harder to dismiss.

Among:

50 topics with at least three independent million-view winners

every topic crossed more than one current subscriber tier.

And:

80.0%

spanned at least a 10x size difference.

At that point, the useful unit of analysis stops being:

the competitor.

It becomes:

the demand pattern.

That is a major shift.

Beginner competitor research asks:

What is this creator doing?

Better research asks:

What keeps working across creators?

The second question is much more transferable.

Build a Competitor Portfolio, Not a Competitor List

Most creators have a flat list:

• Competitor A
• Competitor B
• Competitor C
• Competitor D

A better structure gives each channel a role.

Peer set

Used for:

benchmarking

Emerging set

Used for:

early outliers and disproportionate performance

Established set

Used for:

mature market patterns and high-level execution

Topic-confirmation layer

Used for:

finding where those sets overlap

This final layer is where the most useful intelligence often appears.

If the same audience demand is winning across all three groups, you have learned something bigger than:

"Competitor A had a good video."

You have discovered:

a repeatable market signal.

The Best Practical Rule

If you remember one thing from this study, make it this:

Match competitor size to the question you are asking.

Need a benchmark?

Study peers.

Need inspiration?

Search broadly.

Need early outliers?

Study smaller and emerging channels.

Need mature patterns?

Study leaders.

Need topic validation?

Look for independent wins across all of them.

Subscriber count is context.

It should not become a research prison.

Limitations

This study has several important limitations.

The dataset contains winning videos

Every topic in the final cohort had already produced million-view long-form videos on at least two independent channels.

This is intentionally a winner-focused dataset.

It cannot tell us the success rate of every topic attempted by every channel size.

Subscriber counts are current observations

The analysis uses the latest usable subscriber observations, not exact historical subscriber counts from each video's publication date.

Therefore the study measures:

current size diversity among channels sharing proven topics

not the historical direction in which those topics traveled.

One million views is a high threshold

A 100K-view video can be an enormous breakout for a small creator.

This study focuses on million-view outcomes because they provide a clear common performance threshold.

That makes the research conservative in one sense and biased toward very successful videos in another.

Topic extraction is imperfect

OverseerOS derives normalized topic entities from available video content.

No automated topic representation perfectly captures human semantic judgment.

We removed generic entities and reran the analysis with more-specific two-to-three-word topics.

The cross-size pattern remained, but the exact percentages changed.

Subscriber tiers are editorial categories

The five size tiers are useful for interpretation.

They are not official YouTube categories.

That is why the study also uses continuous 2x, 3x, 5x, 10x, 25x, 50x and 100x size ratios.

Larger channels are well represented in a million-view corpus

A million-view threshold naturally contains many established channels.

The distribution of channel sizes in this sample should not be interpreted as the distribution of all YouTube creators.

Multiple topics can come from the same channel

The final dataset contains 534 channel-topic wins from 213 channels.

A creator can therefore contribute to multiple topic waves.

We avoid treating 534 as 534 completely unrelated creators.

The primary unit of analysis is the 227 recurring topic waves.

Cross-size recurrence is not causation

A topic appearing on differently sized channels does not prove:

• one creator copied another
• subscriber size did not matter
• the topic alone caused the views
• the same idea will work for you

It shows repeated public success across channels of differing current scale.

Final Verdict

Should you only study YouTube competitors your size?

No. Use similar-sized competitors for benchmarking, but use multiple channel sizes for idea discovery and topic validation.

Across 227 recurring million-view YouTube topic waves, OverseerOS found:

• 74.0% crossed subscriber-size tiers
• the median wave spanned a 6.82x difference in current channel size
• 64.8% spanned at least 3x
• 41.0% spanned at least 10x
• 8.8% spanned at least 100x

And among the 50 topics confirmed by at least three independent winning channels:

every topic crossed multiple size tiers.

80.0% spanned at least 10x in current subscriber size.

That does not mean subscriber count is irrelevant.

It means subscriber count serves a specific job.

Use size to interpret performance.

Do not use it to decide which ideas you are allowed to learn from.

Study peers for realism.

Study bigger creators for mature demand.

Study smaller creators for unusual outliers.

Then look for what keeps working across all of them.

Because the most valuable competitor insight is rarely:

"A channel like mine did this."

It is:

"Different creators keep finding the same audience demand, and I understand how to turn that signal into something original."

FAQ

Should I analyze YouTube competitors with more subscribers than me?

Yes. Larger channels can be useful for studying topics, formats, packaging and mature audience demand. Their raw view counts should not be treated as a direct performance benchmark for a much smaller channel.

Should my YouTube competitors be the same size as my channel?

Not all of them. Similar-sized channels are useful for benchmarking, but the OverseerOS study found that 74.0% of recurring million-view topics crossed current subscriber-size tiers.

Are bigger YouTube channels bad competitors to study?

No. They become misleading when you compare raw results without adjusting for channel context. A large creator's topics and formats can still provide valuable market evidence.

Are small YouTube channels better competitors?

They are especially useful for finding disproportionate outliers and emerging opportunities. But a strong competitor portfolio should contain smaller, similar-sized and larger channels.

How much did channel size vary among winning YouTube topics?

Across 227 recurring topic waves, the median largest-to-smallest current subscriber difference was 6.82x. Forty-one percent of the topics spanned channels at least 10x apart in size.

Can the same YouTube topic work on both small and large channels?

It can. In the OverseerOS dataset, recurring million-view topics frequently appeared across different current channel-size tiers. This does not guarantee the topic will work on any specific channel.

What is the best way to choose YouTube competitors?

Choose competitors based on the research job. Use peers for benchmarking, emerging channels for outlier discovery, established creators for mature patterns, and multiple independent channels for topic validation.

How should I compare a big YouTube competitor with a small one?

Compare each video with its own channel's normal performance instead of comparing raw views alone. A smaller video's lower absolute view count can still represent the more unusual breakout.

How many different-sized competitors should I track?

There is no universal size mix. The important principle is not to let one subscriber tier dominate every research question. Maintain peers, larger references and emerging channels, then analyze recurring patterns across them.

What is stronger than finding one viral competitor video?

Independent confirmation. If the same underlying demand produces unusual results across multiple channels, especially channels of different sizes, the evidence is stronger than one isolated winner.

Does a topic working on a 5-million-subscriber channel mean it will work for my small channel?

No. A large-channel win is evidence of demand, not a guarantee of transfer. Look for smaller or mid-sized independent winners, confirm the topic fits your audience and create an original angle.

Why analyze competitors by outlier performance instead of subscribers?

Subscriber count describes channel scale. Outlier performance tells you whether a video behaved unusually relative to that creator's normal results. For idea research, that abnormality can be more informative than raw channel size.