Short-form video retention looks excellent. Reels average 45% to 65% (Retensis, 2026).
Compare that with long-form video, where anything past 20% is respectable.
Now convert the percentage into seconds.
The optimal short-form length is 21 to 34 seconds (Kapwing, 2026).
So 50% retention means roughly 10 to 17 seconds of attention.
A ten-minute video at 20% retention holds someone for two minutes.
The percentages favour short video. The seconds tell a different story. ⏱️
🎁 Compare Video Editing Trials →
🧾 Principais conclusões em resumo
| Medir | Figura | Fonte |
|---|---|---|
| Instagram Reels retention | 45% to 65% | Retensis (2026) |
| YouTube Shorts retention | 40% to 55% | Retensis (2026) |
| TikTok retention | 40% to 50% | Retensis (2026) |
| Optimal length for engagement | 21 to 34 seconds | Kapwing (2026) |
| YouTube Shorts engagement rate | 5.91% | TechRT (2026) |
| TikTok engagement rate | 5.75% | TechRT (2026) |
| Facebook Reels engagement rate | 2% | TechRT (2026) |
| Videos watched to 41% or more | 59% | PostPlanify (2026) |
⏱️ Percentages Against Seconds
This is the distinction that changes how you read every figure here.
⏱️ Same retention, very different attention
Why short video flatters the metric
The shorter the video, the easier a high retention percentage becomes.
Holding someone for fifteen seconds is a low bar.
A 65% retention rate on a 20-second clip is thirteen seconds.
What this means when comparing formats
Never compare retention percentages between short and long video.
Convert both to seconds first, then the comparison is fair.
Most published comparisons skip that step entirely.
When seconds are the right measure
If your message needs two minutes to land, short video cannot deliver it.
Fifty short videos do not add up to one longer explanation.
Attention does not accumulate across separate viewings that way.
📊 What the Platforms Actually Deliver
Engagement rates differ enough between platforms to matter.
| Plataforma | Engagement rate |
|---|---|
| YouTube Shorts | 5.91% |
| TikTok | 5.75% |
| Facebook Reels | 2% |
The gap between the top two is small. The gap to Facebook Reels is not.
Facebook Reels delivers roughly a third of the engagement of the leaders.
Why the top two are so close
Both serve content by interest rather than by who you follow.
That means a new account can reach people immediately.
It also means yesterday’s success guarantees nothing today.
The reach question people get wrong
YouTube Shorts reports around 200 billion daily views.
That describes the platform, not your prospects.
Enormous platform numbers say nothing about what one business will reach.
⚠️ The Three-Second Statistic
One widely repeated figure deserves a closer look.
65% of viewers who watch three seconds will watch ten, and 45% will reach thirty.
That comes from Facebook research, and it is not recent.
| Preocupação | Por que isso importa |
|---|---|
| Published by the platform | Commercial interest in video |
| Predates short-form dominance | Viewing habits changed |
| Still quoted as current | Date usually omitted |
| Definition of a view differs now | Not comparable |
The underlying point is probably still true: openings matter enormously.
The precise numbers should not be treated as current measurements.
What is safe to conclude
Viewers decide very quickly whether to keep watching.
That principle survives whatever the exact percentages are.
Build the opening accordingly and ignore the decimal places.
A pattern worth noticing
This is the third widely quoted marketing statistic in this series with a dating problem.
Figures get repeated long after their context disappears.
Nosso análise de ROI de e-mail traces an even more extreme example.
🎬 What the Length Data Says
The optimal window is narrower than most people assume.
| Length | Desempenho |
|---|---|
| Under 15 seconds | Often too short to say anything |
| 15 to 30 seconds | 5.8% engagement on Reels |
| 21 to 34 seconds | Highest engagement overall |
| Under 90 seconds | About 50% retention |
| Over 90 seconds | Completion falls sharply |
Twenty-one to thirty-four seconds is a remarkably specific finding.
That is roughly seventy to ninety spoken words.
Writing to that length
Count the words before recording rather than editing afterwards.
Ninety words forces you to make one point, which is the correct number.
Most failed short videos try to make three.
The distribution finding
59% of short videos are watched for 41% to 80% of their length.
Only 30% get past 81%.
So most viewers leave before the end, which changes where you put the point.
🪜 Where Short Video Fits in a Sale
Short video does one job well and another badly, and the distinction matters.
🪜 What each format is good for
Why this framing helps
Businesses judge short video by sales and conclude it failed.
It was never going to produce sales in fifteen seconds.
Judge it by whether people arrived, not by whether they bought.
The handover problem
Most short video has nowhere obvious to send interested viewers.
A viewer decides they want more and finds no route.
Every short video should point somewhere specific.
What the destination should be
One page answering exactly what that video raised.
Not your home page, which asks the visitor to search again.
That single change converts far more of the attention you earned (WifiTalents, 2026).
🎯 Structuring Thirty Seconds
The optimal length leaves room for very little, so structure matters.
| Seconds | What belongs there |
|---|---|
| 0 to 3 | The reason to keep watching |
| 3 to 10 | The single point |
| 10 to 25 | The evidence or demonstration |
| 25 to 30 | Where to go next |
There is no room for an introduction, and none is needed.
Nobody watching a feed needs to know your company name first.
The opening three seconds
State the problem or show the result immediately.
A question the viewer has works better than a claim about you.
If the first frame could belong to any business, it is wasted.
Why one point only
Thirty seconds is roughly ninety words.
Two points means forty-five words each, which explains neither.
Split them into two videos instead.
The ending most people get wrong
Vague endings like “follow for more” ask for very little.
Naming one specific next step performs better.
Tell the viewer exactly what to do and where.
📱 Vertical, Captions and the Basics
Format decisions affect results more than production quality does.
| Choice | Effect |
|---|---|
| Vertical framing | Fills the phone screen |
| Captions burned in | Works with sound off |
| Large readable text | Legible on a small screen |
| Movement in the first frame | Slows the scroll |
| Faces | Attract attention reliably |
| Horizontal video, cropped | Looks like an afterthought |
Repurposed horizontal footage is immediately recognisable as repurposed.
Shoot vertically if the destination is vertical.
Captions are not optional
A large share of feed viewing happens with sound muted.
Without captions those viewers get nothing and scroll on.
Most editing tools now generate them automatically (AutoFaceless, 2026).
The scroll-stopping question
Ask whether your first frame would stop you.
Static text on a plain background rarely does.
Movement, a face or an unexpected image all work better.
🔄 Posting Rhythm and Realistic Output
Advice in this field assumes a full-time creator. Most businesses are not that.
| Saída | Realistic for | Sustainable? |
|---|---|---|
| Diariamente | Full-time creators | Not for most businesses |
| Three a week | A dedicated marketer | Sometimes |
| One a week | Most small businesses | Sim |
| Batched monthly | Busy owners | Sim |
Weekly, sustained for a year, beats daily for six weeks.
Fifty-two videos is a substantial body of work.
The batching method
Film four or five in one session with a single setup.
Edit them together, then release one a week.
This is the only approach most owners maintain past month two.
What burnout looks like here
An enthusiastic start, daily posting, then silence after five weeks.
The account then looks abandoned, which is worse than never starting.
Choose the rate you can hold during your busiest month (PostPlanify, 2026).
📉 Why Views Are So Uneven
Short-form results vary more between posts than any other format.
📉 A typical month of short video results
Why the shape looks like this
Content is served by interest, so any post can reach far beyond your followers.
Most do not, and occasionally one does.
Your average is dragged upward by rare outliers.
What this means for expectations
Judging a single post tells you almost nothing.
Judge across twenty posts, and look at the median rather than the mean.
The median describes the typical post far better.
The trap of chasing the outlier
When one video performs unusually well, the instinct is to copy it exactly.
Often the subject mattered more than the format.
Copy the topic, not the editing style (TechRT, 2026).
🔍 Reading Your Own Numbers
Platform dashboards report many things, and few of them matter.
| Métrica | Worth watching? |
|---|---|
| Visualizações | Only as a rough signal |
| Average watch time in seconds | Sim |
| Retention percentage | Only against your own videos |
| Shares | Strong signal of value |
| Comments asking questions | The best signal available |
| Follower growth | Weak, easily inflated |
| Link clicks | Yes, if you offered one |
Shares and questions indicate someone found it genuinely useful.
Views measure distribution. Shares measure worth.
The seconds measure again
Average watch time in seconds is comparable across all your videos.
Retention percentage is not, because your videos differ in length.
Track seconds and the comparison stays honest.
What comments tell you
Questions in comments are future video topics with proven demand.
They also tell you what your explanation left unclear.
Answer them, then make the video that prevents the question.
💷 The Cost of Doing This Properly
Short video looks cheap and consumes more time than expected.
| Estágio | Per video |
|---|---|
| Deciding the subject | 10 minutos |
| Writing ninety words | 15 minutos |
| Filming with retakes | 20 minutes |
| Editing and captions | 40 to 60 minutes |
| Publishing across platforms | 15 minutos |
| Total | Around 2 hours |
Two hours for thirty seconds of finished video is normal.
Weekly posting therefore costs about two hours a week, indefinitely.
Where the time actually goes
Editing dominates, which is why batching helps so much.
Editing five videos in one session is far quicker than five separate sessions.
Setting up and packing away happens once instead of five times.
Deciding whether it is worth it
Two hours a week is roughly a hundred hours a year.
Ask what else those hundred hours could produce.
For some businesses the answer favours writing or direct selling instead.
The honest comparison
A hundred hours could produce twenty-five substantial articles.
Those keep working in search for years afterwards.
Nosso content timeline research covers how that compounding works.
🏬 Which Businesses This Suits
Short video works far better for some businesses than others.
| Tipo de negócio | Fit | Por que |
|---|---|---|
| Visual products and food | Excelente | The product photographs well |
| Trades showing before and after | Excelente | Transformation is inherently watchable |
| Personal services and coaching | Bom | Trust builds through seeing you |
| Local retail | Bom | Reaches a nearby audience |
| Professional services | Moderado | Advice works, the work does not photograph |
| Industrial and technical supply | Poor | Buyers are not in these feeds |
| Anything with a long buying cycle | Weak alone | Awareness only |
The two strongest rows share one thing: something visible changes.
If nothing about your work is visual, this is the wrong channel.
The before-and-after advantage
A cleaned carpet, a fitted kitchen, a repaired roof.
These need no explanation and hold attention naturally.
Trades often dismiss short video and are among the best suited to it.
The professional services problem
An accountant cannot film the work itself.
What works instead is answering one question clearly.
That is a different format, and it performs less spectacularly (Kapwing, 2026).
When to skip it entirely
If your buyers are specialists who do not use these platforms socially.
Reaching a general audience efficiently is worthless if none of them buy.
Reach without relevance costs the same as reach with it.
🧾 Setting a Fair Expectation
Given the variance, here is what a modest effort typically produces.
| Depois | Realistic outcome |
|---|---|
| Month 1 | Low views, learning the format |
| Month 3 | A sense of what your audience responds to |
| Month 6 | Perhaps one strong performer |
| Month 12 | A body of work and a clearer voice |
Nothing here promises sales, and that is deliberate.
Short video builds familiarity, and familiarity converts slowly.
The realistic best case
Some people recognise you before they ever contact you.
That shortens the conversation when they finally do.
It is real value and difficult to attribute to any single post.
🔁 Reusing Rather Than Multiplying
The cheapest way to sustain output is to stop making everything from scratch.
| Source material | Becomes |
|---|---|
| A customer question you answered | One video |
| A longer video you already made | Three or four short cuts |
| Photos from a finished job | A before and after clip |
| An article you wrote | Its single strongest point |
| A common misunderstanding | A correction video |
None of these requires inventing a new idea.
Running out of ideas is the usual reason people stop.
The correction format
Explaining what people commonly get wrong performs consistently well.
It is useful, mildly surprising, and demonstrates expertise at once.
Every trade and profession has ten of these available immediately (WifiTalents, 2026).
Cutting longer video down
A five-minute explanation usually contains three self-contained thirty-second points.
Extract each one and it stands alone.
One recording session can supply a month of posts (Retensis, 2026).
Building a running list
Keep a note on your phone of every question a customer asks.
Add to it whenever one comes up, without deciding anything.
By the time you sit down to film, the topics are chosen already.
That removes the decision that most often stalls a filming session.
What to do with weak performers
Leave them up rather than deleting them.
Occasionally an older post is served again months later.
Deleting removes that possibility for no benefit at all.
One more source of material
The objections people raise before buying are excellent topics.
Address one directly and honestly in thirty seconds.
Naming a genuine drawback builds more trust than any claim about quality.
Very few businesses do this, which is exactly why it works.
A note on comparing yourself with creators
Full-time creators post daily and treat it as their entire job.
Their results reflect that, and yours will not match them.
Compare your videos against your own earlier ones instead.
🚫 O que esses dados não revelam
Retention percentages are length-dependent. They cannot be compared across formats.
Engagement is defined differently by each platform. The rates are not strictly comparable.
Most sources sell video tools. Leia a ênfase de acordo.
Platform totals are not your reach. 200 billion views is context, not opportunity.
Sector matters enormously. Entertainment and professional services behave differently.
The definition problem again
Engagement rate may mean interactions per view or per follower.
Those produce very different numbers from the same activity.
Check the denominator before comparing any two platforms.
What is missing entirely
Almost no published data connects short-form video to actual sales.
Views, retention and engagement are all upstream of revenue.
That gap is the most important limitation in this whole field.
🏁 Versão curta
Short-form retention of 45% to 65% sounds far better than long-form.
Converted to seconds, 50% of a 30-second video is fifteen seconds.
A ten-minute video at 20% retention holds someone eight times longer.
The optimal length is 21 to 34 seconds, or about ninety spoken words.
YouTube Shorts and TikTok engage at around 5.9%, Facebook Reels at 2%.
The famous three-second statistic is old platform research, quoted as current.
And almost nothing published connects any of these numbers to sales. ⏱️
🚀 Browse Creative Trials by Category →
❓ Perguntas frequentes
What is a good retention rate for short-form video?
Reels average 45% to 65%, Shorts 40% to 55%, TikTok 40% to 50%. Convert those to seconds before judging them.
Why does retention percentage mislead?
Shorter videos make high percentages easier. Fifty percent of thirty seconds is only fifteen seconds of attention.
How long should a short video be?
Between 21 and 34 seconds for highest engagement, which is roughly seventy to ninety spoken words.
Which platform engages best?
YouTube Shorts at 5.91% and TikTok at 5.75% are close. Facebook Reels trails at 2%.
Is the three-second rule real?
The principle holds, but the numbers come from older platform research quoted without its date.
Do most people watch to the end?
No. Only 30% of short videos are watched beyond 81% of their length.
Where should I put the main point?
Early. Most viewers leave somewhere between 41% and 80% of the way through.
Does 200 billion daily views help me?
Not directly. Platform totals describe the platform, not the reach any single business will achieve.
Can short video replace longer explanations?
No. Fifty short videos do not accumulate into one two-minute explanation.
Does short-form video drive sales?
Very little published data connects it to revenue. Views and engagement sit well upstream of buying.
What should a small business do with this?
Use short video to reach people and longer video or text to convince them. They do different jobs.
📚 Referências
Retensis. (2026). Average retention rate for YouTube Shorts, TikTok and Reels in 2026Consultado em 8 de agosto de 2026, em https://retensis.com/blog/audience-retention-benchmarks-2026
Kapwing. (2026). Short-form video statistics: TikTok, Reels and Shorts by the numbers in 2026Consultado em 8 de agosto de 2026, em https://www.kapwing.com/resources/short-form-video-statistics-tiktok-reels-and-shorts-by-the-numbers-in-2026/
TechRT. (2026). Short-form video statistics 2026: key growth factsConsultado em 8 de agosto de 2026, em https://techrt.com/short-form-video-statistics/
PostPlanify. (2026). Short-form video statistics 2026Consultado em 8 de agosto de 2026, em https://postplanify.com/blog/short-form-video-statistics-2026
WifiTalents. (2026). Short-form video statistics: verified 2026 dataConsultado em 8 de agosto de 2026, em https://wifitalents.com/short-form-video-statistics/
AutoFaceless. (2026). Short-form video statistics 2026: performance dataConsultado em 8 de agosto de 2026, em https://autofaceless.ai/blog/short-form-video-statistics-2026
Leituras relacionadas neste site
Nosso video marketing analysis covers why satisfaction is falling while adoption rises. See also our open rate research on metrics that stopped meaning what they used to, and the Diretório de participantes com teste gratuito for testing editing tools free.
Sobre esta análise
Retention percentages are structurally easier to achieve on short videos, and this article converts them into seconds so the comparison with long-form is fair. The widely quoted three-second viewing statistic is identified as older platform research repeated without its date. The absence of published data linking short-form metrics to revenue is stated as the field’s main limitation. Checked on August 8, 2026.
-
Short-Form Video: Reach, Retention and Reality - Outubro 24, 2026 -
Outline Generators: The One Part of Scripting Worth Automating - Outubro 24, 2026 -
Video Content in 2026: What the Engagement Data Shows - Outubro 21, 2026