Every content plan we write starts with the same question from the client: how long should the videos be? Most advice online picks a number and repeats it. We had something better, real timings for 116 TikToks that beat their own account's median in eight industries, so we checked. The data says the question is slightly wrong.
The short answer
As long as the idea needs, and no longer. In 116 viral business TikToks we timed, the winners ran from 5 seconds to more than 9 minutes, and length on its own told us almost nothing about how well a video did.
What did show up was shape. The breakouts were either a single beat under 30 seconds or a full story over a minute. The middle was thin.
Where the numbers come from
These are TikToks from our open outlier dataset, the same one behind our short-form video statistics. Every video beat the median views of the account that posted it, usually by a lot. We call that ratio the outlier multiplier: a video at 50x did fifty times that account's normal. You can work out yours with the outlier calculator.
- 116 TikToks with a recorded length (of 120 in the set)
- posted in September and October 2026, English captions, at least 10,000 views, one per creator
- eight industries: real estate, car dealerships, barbershops, med spas, restaurants, gyms and trainers, dentists, home renovation
One honest limit up front: we only have lengths for TikTok. We did not time the Instagram Reels in the set, so everything here is about TikTok.
Length versus performance
| Length | Videos | Share | Median multiplier | Videos at 100x+ |
|---|---|---|---|---|
| Under 15 s | 26 | 22% | 46.4x | 11 |
| 15 to 29 s | 23 | 20% | 45.7x | 5 |
| 30 to 44 s | 8 | 7% | 25x | 2 |
| 45 to 59 s | 13 | 11% | 12.3x | 3 |
| 60 to 89 s | 18 | 16% | 62.8x | 7 |
| 90 s to 3 min | 22 | 19% | 87.4x | 11 |
| Over 3 min | 6 | 5% | 18.7x | 1 |
The median length was 45 seconds. Half the videos fell between 17 and 88 seconds, and the average was 67 seconds, pulled up by a handful of long walk-throughs.
Look at the shape of that table. Forty-nine videos ran under 30 seconds. Forty-six ran a minute or more. Only twenty-one landed between 30 and 59 seconds. Of the 40 videos that beat their account by at least a hundred times, 16 were under 30 seconds, 19 were over a minute, and just 5 were in the middle.
Then we ran the simplest test we know: rank every video by length, rank them again by multiplier, and see whether the two orders line up. They did not. The correlation was -0.06, which is statistician for "no relationship". Length against raw views came out at -0.18, a slight lean towards shorter videos picking up more views, but nothing you could plan a calendar around.
Is 30 to 60 seconds a dead zone?
It looks like one, and we nearly wrote that headline. Then we checked who posted those videos.
The median account behind the 30 to 44 second group had 147,000 followers. Behind the 45 to 59 second group, 313,000. The short videos came from accounts with around 23,000 to 80,000. That matters because big accounts already have high medians, so they almost never post a 1,000x video, whatever its length. Across the whole dataset, accounts over 100,000 followers had the lowest multipliers.
So part of what looks like a length problem is an account-size effect. And the middle still produced real winners: a 55-second toner review from a med spa account with 25,000 followers did 1,657 times its normal, and a gym owner confronting a one-star reviewer did 122.6x at 42 seconds.
Our read: a 45-second TikTok is not cursed. It is just the length most likely to be a short idea stretched or a long idea rushed.
What the winners at each length had in common
Under 15 seconds: one beat, no setup. The top three were a dentist account's "watch the tongue while eating" clip (6 seconds, 2,793x, from 773 followers), a renovation clip opening on "eww, you bought a used double wide" (11 seconds, 2,504x) and a home with cabinet doors that open onto nothing (10 seconds, 1,435x, from 906 followers). Each is one surprise, shown once. There is nothing to cut.
15 to 29 seconds: a quick before and after. A body transformation ending in a supplement promo (27 seconds, 1,339x) and a raw reaction to eyeliner tattoo removal (20 seconds, 597x) both had a single change to show and a face reacting to it.
60 to 89 seconds: a question with a payoff. The highest multiplier in the whole dataset lives here: a salesperson asking colleagues around the floor for the highest car payment they ever sold, 69 seconds, 5.8 million views, 2,819 times that account's normal. Every answer is a reason to wait for the next one.
90 seconds to 3 minutes: an expert explaining. The highest median of any band, 87.4x. An aesthetician narrating a green sea peel (96 seconds, 1,779x), a local buffet review (91 seconds, 1,741x) and a dentist explaining at-home whitening instead of veneers (106 seconds, 1,018x). People stay when someone who does this all day is talking them through it.
Over 3 minutes: tours, rarely. Only six videos, mostly property walk-throughs. A rent-to-own duplex tour ran six minutes and still did 258x. Long works if the viewer came to see the whole thing.
Length by industry
| Industry | Median length | TikTok median multiplier |
|---|---|---|
| Real estate | 76 s | 73.4x |
| Barbershops | 55.5 s | 22.3x |
| Dentists | 54 s | 29.9x |
| Med spas | 45 s | 372x |
| Restaurants | 43 s | 14.8x |
| Gyms and trainers | 37 s | 93.2x |
| Car dealerships | 36 s | 45.1x |
| Home renovation | 17.5 s | 65.8x |
Real estate ran longest, which fits: a home is a walk-through. Renovation ran shortest, because the best renovation clip is usually one reveal. Neither tells you the length to use for your next video. They tell you what the format of a winning idea tends to be in that industry.
How we decide the length of a video
This is the process we use on client content, and it is simple enough to use without us.
- Write the idea in one sentence. If the whole idea is one surprise, it is a short video. If it needs a question and an answer, or steps, it is a long one.
- Cut everything before the hook. The first two seconds should already be the idea. Length is earned after that, not before.
- Keep a beat only if it raises a question or answers one. That rule alone will put most videos under 30 seconds or over a minute without you choosing a number.
- Do not stretch a short idea to hit a target. That is how videos end up in the soft middle.
- Measure against your own median. Whatever length you pick, judge it by whether it beat your normal, not by a number someone set for every account.
If you want to see the openings that worked in your industry, our viral hook library lists all 240 videos by hook, with their lengths. And if you are still choosing between platforms, start with TikTok or Instagram Reels for a local business.
Frequently asked questions
What is the best length for a business TikTok?
There is no single best length. In 116 viral business TikToks, length had almost no relationship with performance. The winners were either one beat under 30 seconds or a full story over a minute, so choose the length the idea needs rather than a fixed number.
Do longer TikToks get more views?
Not reliably. In our data, length against raw views showed a slight lean towards shorter videos, and length against each account's normal performance showed no relationship. Videos from 90 seconds to 3 minutes had the highest median multiplier, 87.4x, but the ideas in that band were explainers and reviews, not stretched clips.
Are 30 to 60 second TikToks bad?
No, but they were the thinnest group in our data, 21 of 116 videos, and they scored lowest. Much of that is because big accounts, which score lower on this measure anyway, posted most of them. A 55-second med spa review still did 1,657 times its account's normal.
How long should the hook be?
About two seconds. The top short videos in our data opened on the idea itself, and the top long ones opened on the question or the procedure they went on to explain. Anything before the hook is time the viewer can use to scroll away.
Does this apply to Instagram Reels too?
We cannot say from this data. We recorded lengths for the TikToks in our dataset but not for the Instagram Reels, so these findings are about TikTok only.