How to go viral on TikTok, and what that actually means

Almost everything written about going viral on TikTok is written as though the algorithm were a judge handing down a verdict on your video. It is not. It is a series of small, cheap experiments, and understanding it as experiments rather than as judgement changes what you do about it.

When you post, TikTok shows the video to a small group of people whose feeds it thinks it fits. Not your followers, particularly: unlike almost every other platform, TikTok has never been follower-first, which is why an account with 200 followers can outrun one with 200,000 on a given day. It watches what that first group does. If the signals are good, it shows the video to a larger group, and it repeats that. Most videos stop after one or two rounds. A very small number keep clearing the bar and end up in front of millions.

So the question “how do I go viral” is really the question “what makes a batch of strangers behave in a way that earns the next batch”. That question has answers.

What the platform is counting

TikTok has never published a ranking formula and never will. But it has said in general terms what its recommendation system weighs, and years of creators comparing notes have converged on the same short list. In rough order of how much they seem to matter:

Completion rate

The share of viewers who reach the end. This is the single most repeated finding in the category, and it is why short videos punish you less: a 15-second video watched to the end beats a 60-second video abandoned at 30, even though the second one collected twice the watch seconds.

Rewatches

A loop counts again. Anything that rewards a second viewing, a detail you only catch the second time, a cut that lands you back where you started, is worth a disproportionate amount.

Saves

A save is the strongest ordinary signal a viewer can send. It says “this is worth coming back to”, which is a much harder thing to fake than a like. Tutorials, lists, recipes and reference content collect saves; opinion content mostly does not.

Shares

Sending a video to one person is what moves it out of your own audience and into someone else’s. It is the mechanism by which a video reaches people the algorithm would never have chosen for it.

Comments

Real ones. A comment that is a reply to another comment is worth more than a wall of emoji, because it means people are still on the video.

Search relevance

TikTok transcribes your audio and indexes your caption and on-screen text. A meaningful share of views on a good video arrive weeks later from search, which is a completely different game from the first-day push.

The first second decides most of it

Every one of the signals above is downstream of a viewer choosing not to scroll. That decision happens fast, before anyone has consciously evaluated anything, and it is made on the strength of what is on screen and in the audio at the moment the video appears.

This is why the highest-leverage edit you will ever make is deleting the beginning. Almost every video that underperforms has a runway on the front of it: a logo, a greeting, a slow push-in, a sentence explaining what the video is going to be about. Cut all of it. Start where the interesting thing starts, and let the context arrive in the second or third sentence, by which point you have already been given the attention you needed to deliver it.

The length question, answered properly

The advice “keep it short” is right for the wrong reason. Short videos do not win because short is good. They win because completion rate is a ratio, and a shorter video has less opportunity to lose people before the end.

Which means the actual rule is: make the video exactly as long as the idea, and no longer. A three-minute video that genuinely holds people for three minutes will do extremely well. A 40-second video padded out of a 20-second idea will do badly. The failure mode is not length, it is dead air, and dead air is easy to find once you look for it.

• Watch your own video back with the sound off, and mark every second where nothing new appears on screen.
• Watch it again with the picture off, and mark every second where nothing new is said.
• The seconds marked twice are the ones to cut. There are usually more than you expect.

What you actually control

It is worth being blunt about the split, because a great deal of frustration comes from working hard on the wrong half.

Volume, and the trap inside it

More posting does correlate with more views, and the effect is not small. It is also the advice that destroys the most accounts, because the mechanism people assume is wrong. Posting more does not persuade the algorithm to like you. It gives you more lottery tickets and, far more importantly, more reps.

The failure is posting more by making each one worse. Five thin videos a week will teach an audience to scroll past your face. Three good ones will not. The number worth chasing is the highest one you can sustain without dropping quality, and for almost everybody that is somewhere between three and five a week rather than daily.

A workable weekly loop

1. Find three videos in your niche from the past fortnight that clearly outperformed the account that posted them. Not the biggest accounts, the biggest surprises.
2. For each one, write down what the first second does, what the shape of the middle is, and what would make someone send it to a friend. That is the whole analysis.
3. Take the shape, not the topic. Copying the topic makes you the second-best version of somebody else. Copying the structure makes it yours.
4. Write the hook first, before the script. If you cannot make the first line interesting, the idea is not ready and no amount of editing will save it.
5. Film it, then cut the first three seconds anyway, and see whether it is better. It usually is.
6. At the end of the month, put the videos in order of performance and look for what the top three share. That list is worth more than any general advice, including this page.

Step six is the one everybody skips, and it is the one that compounds. General advice gets you to a competent baseline. Your own top-performing videos are the only source of information about what works for your audience, and after a couple of months the pattern is usually obvious and frequently surprising.

If you want the first three steps done for you, that is precisely what Viral Pop does: paste a link and it reads the video, tells you what the opening is doing and why the middle holds, and turns it into ideas built for your niche rather than for the account you analysed. But the loop above works with nothing but a notes app and some honesty about your own footage.