What Is a Good YouTube VPH? We Analyzed 2,826 Videos to Find Out

There is no universal “good” YouTube views-per-hour number.

In an OverseerOS analysis of 2,826 public competitor videos from 83 YouTube channels, a fixed VPH threshold produced a surprisingly unreliable picture of what was actually breaking out.

The clearest example:

• 46.2% of breakout videos were below 500 views per hour
• 72.7% of videos above 500 views per hour were not breakouts
• 65.3% of breakouts were below 1,000 views per hour
• Yet 1,000 VPH was already around the top 10% of raw video velocity in the sample

The reason is simple.

500 views per hour means something completely different on a channel that normally gets 60 views per hour than it does on a channel that normally gets 1,200.

That was the strongest pattern in the data.

Among breakout videos running below 500 VPH, the median video was doing only 215 views per hour, but its channel baseline was just 60 VPH.

Among videos running above 500 VPH that were not breakouts, median velocity was 925 VPH, but those channels normally operated at around 1,158 VPH.

So the slower-looking video was almost 2.9× its channel baseline.

The faster-looking video was running at only 0.88× its baseline.

That leads to a much better way to think about YouTube view velocity:

Do not ask whether a video has “high VPH.” Ask whether its VPH is unusually high for that channel, at that stage of the video's life.

Key Findings

The headline result is not that 598 VPH is “good.”

It is that raw velocity without a channel baseline can be deeply misleading.

What Does VPH Mean on YouTube?

VPH means views per hour.

It is a way to express how quickly a YouTube video is accumulating views.

But there is an important problem:

Not every VPH tool calculates it the same way.

For this OverseerOS study, view velocity was calculated as:

Public video views ÷ hours since publication

So if a video had:

• 24,000 views
• 48 hours since publication

its measured velocity would be:

24,000 ÷ 48 = 500 views per hour

This is an average views-per-hour rate since publication.

It is not necessarily the same as a tool showing the number of views gained during the literal most recent hour.

A rolling one-hour VPH can spike or collapse quickly.

Average-since-upload VPH changes more gradually.

So the exact thresholds in this study should not be copied blindly into another tool using a different VPH formula.

The broader finding is what matters:

Absolute velocity needs context.

VPH is not the same as views

A video with 100,000 views can have lower current velocity than a video with 20,000 views.

For example:

Video A

• 100,000 views
• 20 days old
• average velocity: about 208 VPH

Video B

• 20,000 views
• 12 hours old
• average velocity: about 1,667 VPH

Video A has five times more total views.

Video B is accumulating them far faster relative to time.

That is why velocity can surface emerging videos before sorting by total views would.

But even that is only half the story.

You still need to know what is normal for the channel.

How We Analyzed 2,826 Competitor Videos

We started with 2,919 unique public YouTube video records captured through OverseerOS competitor tracking.

To make a relative-velocity study possible, each video needed a valid positive channel velocity baseline.

93 videos did not have one, so they were excluded.

That left:

• 2,826 videos
• 83 distinct competitor channels
• 277 breakout observations
• 2,549 non-breakout observations

Every qualifying video was observed within its first 30 days after publication.

The qualifying records were collected between August 2025 and August 2026.

How velocity was calculated

For each video:

Video velocity = public views ÷ hours since upload

We then compared that velocity with the recent velocity baseline available for its channel.

The relative score was:

Relative velocity = video VPH ÷ channel baseline VPH

If a channel's recent baseline was 200 VPH and a new video was running at 600 VPH:

600 ÷ 200 = 3× baseline

For this study, OverseerOS operationally classified a video as a breakout when its relative velocity was:

greater than 2× the channel baseline

That is an OverseerOS research definition.

It is not an official YouTube threshold.

YouTube does not publish a rule saying that 2× VPH causes a video to be recommended.

Why we used a relative threshold

Consider two videos moving at 500 VPH.

Channel A normally does 100 VPH

500 ÷ 100 = 5× baseline

Channel B normally does 1,000 VPH

500 ÷ 1,000 = 0.5× baseline

Same raw velocity.

Completely different performance signal.

That is exactly the distinction we wanted to test.

We accounted for channels contributing different numbers of videos

The 2,826 videos were not 2,826 independent channels.

The median channel contributed 17 qualifying videos, while some contributed considerably more.

So we did not rely only on pooled video counts.

For the central 500-VPH finding, we also repeated the analysis at the channel level.

The conclusion survived.

Across the 59 channels that produced at least one breakout, the average channel had 46.4% of its breakout videos below 500 VPH.

The median channel had 42.9% of its breakouts below 500 VPH.

And among the 53 channels with at least one 500+ VPH video, the median channel had 60% of those high-VPH videos fail to qualify as breakouts.

So the result is not explained by one prolific channel flooding the dataset.

Finding 1: There Is No Universal “Good” YouTube VPH

A fixed VPH benchmark sounds attractive because it gives creators a simple answer.

100 VPH is good.

500 VPH is trending.

1,000 VPH is viral.

The data did not support that kind of rule.

We tested several raw velocity thresholds.

The tradeoff is brutal.

Raise the threshold and you get fewer false alarms.

But you also throw away large numbers of genuine relative breakouts.

At 500 VPH, almost half of all breakouts disappear.

At 1,000 VPH, nearly two-thirds disappear.

At 5,000 VPH, you find only 17 of 277 breakouts.

Even among the 34 videos moving at least 5,000 VPH, only half were above 2× their own channel baseline.

That is why there is no honest answer like:

“A good YouTube VPH is 500.”

The better answer is:

A good VPH is one that is unusually high relative to the channel's normal velocity, with video age taken into account.

Finding 2: 46% of Breakout Videos Were Below 500 VPH

This is where a raw VPH filter becomes dangerous.

Of the 277 breakout videos:

128 were below 500 VPH.

That is:

46.2% of all breakouts in the qualifying sample.

And:

181 of 277 were below 1,000 VPH.

That is:

65.3%.

Even 37 breakout videos, or 13.4%, were below 100 VPH.

These were not labeled breakout because their raw speed was enormous.

They qualified because their performance was unusual relative to the channel around them.

The channel-level sanity check produced almost the same result

Because some channels contributed more videos than others, we checked whether a few prolific channels created the 46.2% result.

They did not.

Across breakout-producing channels:

• channel-weighted average share of breakouts below 500 VPH: 46.4%
• median channel share below 500 VPH: 42.9%

The pooled result was 46.2%.

That is unusually consistent.

The finding survives when channels, rather than videos, receive more equal influence.

Why low VPH can still be exceptional

Imagine a small competitor whose recent videos usually accumulate around 40 views per hour.

A new upload reaches 180 VPH.

Most creators scanning a feed for “500+ VPH” would never see it.

But:

180 ÷ 40 = 4.5× baseline

Something unusual is happening.

The absolute number looks small.

The relative signal is enormous.

That video may be strategically more interesting than a 2,000-VPH upload from a channel whose normal baseline is 3,000.

Finding 3: Most Videos Above 500 VPH Were Not Breakouts

The reverse problem was just as important.

There were:

546 videos running at least 500 VPH.

Only:

149 qualified as >2× relative breakouts.

That means:

397 did not.

Or:

72.7% of the 500+ VPH videos were not breakouts relative to their channel baseline.

The same issue appeared at 1,000 VPH.

There were:

276 videos above 1,000 VPH.

Only:

96 were breakouts.

So:

65.2% of 1,000+ VPH videos were not relative breakouts.

This is particularly striking because 1,000 VPH was already close to the 90th percentile of raw velocity in the full sample.

Across the 2,826 qualifying videos:

A video could therefore sit around the top 10% of raw VPH in this dataset and still be completely ordinary for its own channel.

That sounds counterintuitive until you remember what is being compared.

A giant channel doing 1,000 VPH may be underperforming.

A small channel doing 200 VPH may be exploding.

Finding 4: A 215-VPH Video Could Be More Interesting Than a 925-VPH Video

This was the cleanest demonstration of why baseline matters.

We split the sample into two deliberately conflicting groups.

Group A: Breakouts below 500 VPH

These were videos that looked relatively slow in absolute terms but still exceeded 2× their channel baseline.

There were:

128 videos across 35 channels.

Group B: Non-breakouts above 500 VPH

These were objectively faster videos that failed to reach 2× their channel baseline.

There were:

397 videos across 38 channels.

Here is what happened.

Read those rows carefully.

The non-breakout group had:

• more than 4× the raw velocity
• more than 8× the total views

Yet the lower-view group was far more unusual relative to its own channels.

The median 215-VPH breakout was running at:

2.87× baseline

The median 925-VPH non-breakout was running at:

0.88× baseline

That is the difference between popularity and outlier behavior.

Raw popularity asks:

“Which video is bigger?”

Relative analysis asks:

“Which video is behaving unusually for the creator that published it?”

For competitor research, the second question is often more valuable.

Finding 5: The 2× Breakout Threshold Landed Almost Exactly at the Top Decile

We also looked at the distribution of relative velocity itself.

Across all 2,826 qualifying videos:

The 90th percentile was:

1.97× baseline

The operational OverseerOS breakout rule was:

greater than 2× baseline

And the actual share classified as breakout was:

9.8%

So in this particular competitor dataset, the 2× rule landed almost exactly around the top decile of relative video velocity.

That is useful context.

It does not mean YouTube has a hidden 2× threshold.

It means that, within this observed sample, being around twice a channel's normal velocity was genuinely uncommon.

The typical breakout was well above the cutoff

The median breakout video was running at:

2.93× baseline

The 75th percentile breakout was:

4.25×

The 90th percentile breakout was:

6.82×

By comparison, the median non-breakout was running at:

0.48× baseline

So the difference between normal and breakout behavior was not merely a few percentage points.

The relative distributions were meaningfully separated.

For a deeper explanation of how to benchmark outliers against a creator's normal performance, see our YouTube Outlier Benchmark Report.

Finding 6: Video Age Changes What VPH Means

Velocity is not independent of time.

That matters especially for the way this study calculates it.

If a video receives a huge burst of views immediately after publication and then slows down, its average VPH will gradually decline as more hours enter the denominator.

So a 12-hour-old video and a 20-day-old video should not be treated as equivalent.

We split the sample by video age.

At the stored observation:

24.1% of videos under 24 hours old qualified as breakouts.

For videos between 14 and 30 days old:

5.3% did.

The median relative score also declined sharply with age.

This does not prove YouTube “decides” in the first 24 hours

That would go beyond what the dataset can establish.

There are at least three reasons to be cautious.

First, the metric itself uses hours since publication in the denominator.

Second, this dataset retains the latest tracked snapshot rather than a complete timestamp-by-timestamp history for every video.

Third, YouTube videos can gain traction late. A video can be rediscovered through recommendations, search, external events, or renewed topic interest.

So the defensible conclusion is narrower:

Relative velocity signals were more concentrated among younger videos in this dataset, which makes early monitoring useful, but the study does not establish an exact “viral decision window.”

YouTube's own recommendation guidance focuses on signals such as viewer appeal, engagement, satisfaction, personalization, topic interest, and competition rather than publishing a universal VPH rule.

VPH is therefore best treated as an observational research signal, not a direct reading of the recommendation algorithm.

How Early Can You Spot a YouTube Breakout?

The data gives us a useful answer, with one important qualification.

Among the 277 videos that were labeled breakouts at their stored observation:

• 24.9% were under 24 hours old
• 42.6% were under 48 hours old
• 71.5% were under 7 days old
• 85.9% were under 14 days old

So a substantial share of the breakout observations were already visible while the videos were still relatively young.

But these are ages at observation.

They are not the exact timestamps when each video first crossed 2× baseline.

That distinction matters.

We cannot say:

“71.5% of breakouts can be predicted within seven days.”

We can say:

71.5% of the breakout observations in this sample were attached to videos younger than seven days when measured.

That supports a practical competitor-research strategy:

watch recent uploads early instead of waiting until total views make the winner obvious.

Finding 7: A Fixed 500-VPH Rule Gets Worse as Videos Age

We tested the 500-VPH threshold inside different age groups.

Even in the first 24 hours, a 500-VPH filter was far from definitive.

Only:

43.0% of 500+ VPH videos under 24 hours old were breakouts.

And the threshold still missed:

one-third of breakouts in that age group.

For older videos the raw threshold became even less discriminating.

Among 14-to-30-day-old videos above 500 VPH:

only 14.1% qualified as relative breakouts.

That is another reason a universal VPH benchmark is the wrong abstraction.

At minimum you need:

1. video age
2. channel baseline
3. comparable recent uploads

So What Is a Good YouTube VPH?

The best answer supported by this study is:

There is no universal good YouTube VPH. A useful VPH is one that is unusually high relative to the channel's own recent velocity baseline, compared at a sensible video age.

If you need a practical hierarchy:

Raw VPH answers: “How fast is this video moving?”

Useful for:

• ranking recent videos by momentum
• scanning many uploads quickly
• spotting high absolute activity

Relative VPH answers: “How unusual is this speed for this channel?”

Useful for:

• finding small-channel outliers
• comparing channels of different sizes
• detecting videos outperforming creator norms
• separating expected popularity from unusual momentum

Age answers: “Is this a fair comparison?”

Useful for avoiding comparisons like:

• 8-hour-old upload vs 20-day-old upload
• launch spike vs mature video
• new video vs evergreen catalog hit

The strongest workflow uses all three.

Is 100 VPH Good on YouTube?

It can be.

In this dataset, the median raw velocity across qualifying videos was approximately:

110 VPH

So 100 VPH was around the middle of this particular competitor sample.

But 37 breakout videos were below 100 VPH.

A channel that normally gets 15 VPH could have a serious outlier at 80 VPH.

A channel that normally gets 2,000 VPH would have a severe underperformer at 100.

Raw VPH alone cannot answer the question.

Is 500 VPH Good on YouTube?

500 VPH is objectively fast compared with most videos in this dataset.

It was above the 75th percentile of raw velocity.

But it was not a reliable breakout threshold.

Among all qualifying videos at 500+ VPH:

• 27.3% were breakouts
• 72.7% were not

And a 500-VPH filter would have missed:

128 of 277 breakouts

So:

500 VPH is high absolute velocity, but it does not automatically mean a video is unusually successful for its channel.

That distinction is critical.

Is 1,000 VPH Viral?

Not by itself.

1,000 VPH was approximately the 90th percentile of raw velocity in this sample.

That makes it uncommon.

But among videos above 1,000 VPH:

only 34.8% qualified as >2× relative breakouts.

And:

181 of the 277 breakouts were below 1,000 VPH.

So 1,000 VPH can describe a very fast video.

It cannot tell you whether the video is outperforming what its channel normally achieves.

A Better YouTube View-Velocity Framework

Instead of asking whether a video passed a universal threshold, use this five-step process.

Step 1: Measure the video's velocity

For an average-since-upload calculation:

views ÷ hours since publication

Example:

40,000 views after 40 hours:

1,000 VPH

Step 2: Find the channel's normal velocity

Take a group of comparable recent uploads.

Avoid letting one giant historical hit define “normal.”

The goal is to estimate what recent performance usually looks like for that creator.

Step 3: Calculate relative velocity

Use:

video VPH ÷ channel baseline VPH

Example:

Current video:

1,000 VPH

Channel baseline:

300 VPH

Relative velocity:

3.33×

Now the number has context.

Step 4: Compare at a similar age

A useful competitor set might be:

• videos under 24 hours
• videos 24 to 48 hours old
• videos 2 to 7 days old

Do not casually compare a 10-hour launch spike with a 25-day-old mature video.

Step 5: Investigate the videos that escape the baseline

Velocity identifies what deserves attention.

It does not explain why the video worked.

Once a video is clearly outperforming, investigate:

• topic
• title
• thumbnail
• hook
• format
• timing
• audience promise
• novelty
• competitor context

That is where velocity becomes strategy.

Why Raw Views Can Hide Emerging Competitors

Suppose you track two channels.

Competitor A

• normal velocity: 50 VPH
• new video: 200 VPH
• total views: 15,000
• relative velocity: 4×

Competitor B

• normal velocity: 2,000 VPH
• new video: 1,200 VPH
• total views: 250,000
• relative velocity: 0.6×

If you sort by views:

Competitor B wins easily.

If you sort by VPH:

Competitor B still wins.

If you ask which video is producing the more unusual signal:

Competitor A wins.

That is why competitor intelligence based only on total views tends to identify opportunities late.

By the time the small-channel video reaches 500,000 views, everyone can see it.

The strategic advantage comes from noticing that 200 VPH was already abnormal for that creator.

This finding fits the larger pattern in our breakout video research: the most informative competitor video is not always the one with the largest raw view count.

What VPH Can Tell You

VPH can help answer:

• Which recent competitor videos are moving quickly?
• Which uploads have unusual momentum?
• Which new videos deserve deeper investigation?
• Which competitors are producing current outliers?
• Which topics may be gaining traction now rather than historically?
• Which videos are outperforming the creator's normal velocity?

It is particularly valuable when you are watching a market continuously.

A list sorted by lifetime views tends to surface old winners.

A velocity-based view can surface current movement.

What VPH Cannot Tell You

VPH does not reveal:

• impressions
• click-through rate
• average view duration
• audience retention
• returning viewers
• traffic-source mix
• recommendation impressions
• viewer satisfaction
• subscriber conversion
• why YouTube distributed the video
• whether the video will continue growing

Those metrics are either private to the channel owner or require different data.

A high relative VPH is therefore evidence that something unusual is happening.

It is not proof of the causal mechanism.

That distinction matters.

A title might be stronger.

The topic could be surging.

The thumbnail might have improved.

A large external site could have linked to the video.

A previous upload may have sent viewers into the channel.

The recommendation system could have found a better audience match.

Velocity tells you:

“Investigate this.”

It does not tell you:

“This exact tactic caused the views.”

How to Apply This With OverseerOS

The research suggests a very specific competitor workflow.

1. Find competitors worth tracking

Use the OverseerOS Viral Channel Finder to find breakout and emerging channels in your niche instead of building your research set only from giant established creators.

A useful competitor set includes channels of different sizes, especially smaller creators producing current outliers.

2. Track their new uploads

The OverseerOS YouTube Competitor Analysis Tool uses Overseer Feed to monitor recent public competitor uploads.

The feed can surface public signals including:

• publish time
• views
• initial velocity
• relative viral score
• breakout status

You can filter by recent upload windows such as 24 hours, 3 days, 7 days, and 30 days.

That matters because this study showed how dramatically age changes the interpretation of VPH.

3. Do not sort by VPH and stop

High VPH is a discovery filter.

Relative performance is the next question.

A 700-VPH video from one competitor may be routine.

A 250-VPH video from another may be its strongest recent outlier.

Use the channel context.

4. Inspect the video causing the signal

Once a video escapes baseline, examine:

• title structure
• thumbnail promise
• topic
• framing
• opening hook
• content format
• timing

Do not copy the execution.

Reverse-engineer the pattern that may be worth adapting.

5. Confirm the pattern across multiple channels

One outlier is evidence.

Several independent creators showing similar outliers is much stronger market evidence.

If a topic or packaging pattern suddenly appears across several breakout competitors, you may be looking at something more important than one creator having a good upload.

That is where competitor tracking becomes a content-research system rather than a leaderboard.

The Practical VPH Checklist

Before calling a YouTube video “viral” because of its views per hour, check:

• What exact VPH formula am I using?
• How old is the video?
• What does this channel normally do?
• What is the video's VPH relative to that baseline?
• Am I comparing similar video ages?
• Is one historical outlier distorting the baseline?
• Is this unusual across several uploads or just one?
• Are similar competitors showing the same pattern?
• What topic, title, thumbnail, or format changed?
• Am I treating velocity as evidence rather than proof of causation?

If you cannot answer those questions, the VPH number alone is not enough.

Limitations

This study has several important limitations.

First, the sample is not a random census of YouTube.

These were public videos from competitor channels tracked through OverseerOS workflows. Creators chosen as competitors may systematically differ from the average YouTube channel.

Second, multiple videos came from the same channels.

We addressed this for the central 500-VPH result with channel-level robustness checks, but the full video dataset should still not be interpreted as 2,826 independent creators.

Third, VPH in this study means:

public views divided by hours since publication at the stored observation

Some platforms and analytics tools calculate VPH using a more recent rolling window.

Those measurements are not interchangeable.

Fourth, the dataset stores the latest retained competitor-video snapshot rather than a complete observation every hour throughout each video's life.

So we cannot identify the exact moment when every breakout first crossed 2× baseline.

We also cannot use this dataset to claim that an early breakout label predicts final lifetime views.

Fifth, channel baselines can evolve.

A creator that repeatedly produces stronger videos will eventually establish a higher normal baseline.

That is desirable for ongoing competitor intelligence, but it means relative velocity is contextual rather than permanent.

Sixth, the historical dataset did not have sufficiently complete duration coverage for a defensible Shorts-versus-long-form comparison, so we deliberately did not publish one from this study.

Finally, this is observational public data.

It does not establish that VPH, relative velocity, or the 2× threshold causes YouTube recommendations.

The 2× threshold is an OverseerOS operational definition for identifying unusual competitor performance.

It is not an official YouTube ranking factor.

Final Verdict

There is no universal number that defines a good YouTube VPH.

Our analysis of 2,826 competitor videos across 83 channels found that:

• 46.2% of breakout videos were below 500 VPH
• 65.3% were below 1,000 VPH
• 72.7% of 500+ VPH videos were not relative breakouts
• 65.2% of 1,000+ VPH videos were not relative breakouts
• the operational 2× breakout threshold landed almost exactly around the 90th percentile of relative velocity
• 71.5% of breakout observations occurred while the videos were under seven days old

The most revealing comparison was even simpler:

Breakouts below 500 VPH

• median speed: 215 VPH
• median baseline: 60 VPH
• relative velocity: 2.87×

Non-breakouts above 500 VPH

• median speed: 925 VPH
• median baseline: 1,158 VPH
• relative velocity: 0.88×

The slower video can be the stronger signal.

So stop asking:

“Is 500 VPH good?”

Ask:

“How unusual is this video's velocity for this channel, at this stage after publication?”

That is the number worth investigating.

And once you find the outlier, VPH has done its job.

The next step is understanding what changed.

FAQ

What is VPH on YouTube?

VPH means views per hour. It describes how quickly a YouTube video is accumulating views. In this OverseerOS study, VPH was calculated as public video views divided by hours since publication.

What is a good VPH on YouTube?

There is no universal good VPH. In the OverseerOS analysis of 2,826 competitor videos, 46.2% of relative breakout videos were below 500 VPH, while 72.7% of videos above 500 VPH were not breakouts. Channel baseline and video age were more useful context than a fixed raw threshold.

Is 500 views per hour good on YouTube?

500 VPH is relatively fast compared with most videos in this dataset, but it was not a reliable breakout benchmark. Only 27.3% of videos above 500 VPH qualified as greater-than-2× channel-baseline breakouts.

Is 1,000 VPH viral on YouTube?

Not necessarily. Around 1,000 VPH was near the 90th percentile of raw velocity in this sample, yet only 34.8% of videos above 1,000 VPH were relative breakouts. A large channel can produce 1,000 VPH routinely.

How do you calculate YouTube views per hour?

For average views per hour since upload:

VPH = total public video views ÷ hours since publication

A video with 12,000 views after 24 hours has an average velocity of 500 VPH.

Some analytics tools use shorter rolling windows, so always check how the specific tool defines VPH.

What is relative view velocity?

Relative view velocity compares a video's VPH with the normal VPH of its channel.

Relative velocity = video VPH ÷ channel baseline VPH

If a video is doing 600 VPH on a channel that normally does 200 VPH, its relative velocity is 3× baseline.

What is a good YouTube outlier score?

There is no universal outlier score across every tool because methodologies differ. In this OverseerOS study, a relative velocity greater than 2× the channel baseline was used as the operational breakout definition. Across the qualifying sample, the 90th percentile was approximately 1.97× baseline.

Is VPH a YouTube ranking factor?

YouTube does not publicly identify a universal VPH threshold as a recommendation ranking rule. Its recommendation guidance discusses factors including viewer appeal, engagement, satisfaction, personalization, topic interest, and competition. VPH is best used as an external performance signal for observing momentum, not as a claimed direct algorithm score.

How early can you tell if a YouTube video is going viral?

There is no guaranteed time window. In this OverseerOS dataset, 24.9% of breakout observations involved videos under 24 hours old, 42.6% were under 48 hours old, and 71.5% were under seven days old. Because the dataset does not record the exact first breakout moment for every video, these percentages should not be interpreted as predictive deadlines.

Should I compare VPH across different YouTube channels?

Only with caution. Raw VPH can be dramatically different between channels because their normal audience sizes and distribution levels differ. Comparing each video's velocity with its own channel baseline is usually more informative for competitor research.

Can a low-VPH video still be a breakout?

Yes. In this study, 128 of 277 breakouts were below 500 VPH. The median of those low-VPH breakouts was only 215 VPH, but their median relative velocity was 2.87× their channel baseline.

Can a high-VPH video be underperforming?

Yes. The dataset contained 397 videos above 500 VPH that did not qualify as >2× relative breakouts. Their median velocity was 925 VPH, but their channels had a median baseline of 1,158 VPH.

What is the best way to use VPH for competitor research?

Use raw VPH to find moving videos, compare it with the creator's normal velocity, control for video age, then investigate the topics, titles, thumbnails, hooks, and formats behind the strongest relative outliers. The goal is to identify unusual market evidence before the winner becomes obvious from total views alone.