How to Find Viral TikTok
Formats in Your Niche
Search the problems your audience has, follow a small watchlist of creators, and flag videos that beat each creator's usual views, not just big view counts. Keep a structure only when independent creators make it work, then adapt it with your own product and proof and test one change at a time. The output of research is a brief you can film.
A view count tells you a video was seen. It does not tell you whether the video was unusual, whether its structure works for anyone else, or whether you could make your own version. This guide is the method we use to answer those three questions, in that order: Baseline Repeat Fit.
What is a TikTok format, and how is it different from a trend?
A format is a repeatable sequence that shapes what the viewer experiences; a trend is anything gaining attention in a community for a while. A sound, a topic, a visual style or a format can all trend, and they are different things to write down.
"Meal prep" is a topic. "Show a disappointing lunch, reveal one change, demonstrate it, then compare the results" is a format. The song playing underneath is an execution choice.
The distinction matters when a reference looks nothing like your product. A meal-prep video and a software demo can share a structure: show the friction, make one change, prove the improvement. Copying the meal-prep topic would miss the part that transfers. For this month's formats with real examples, see viral content formats; this guide is how to find the ones that fit your niche.
Where should you look for TikTok formats in your niche?
Start from one audience problem, not from "viral videos". Write one sentence before opening the app: "We want to reach people who need ___ and help them ___." For a small home-organizing business that might give searches like "tiny kitchen storage", "renter kitchen organization" and "no pantry meal prep".
Use three routes and keep their evidence separate:
| Route | What it gives you | What it cannot tell you |
|---|---|---|
| Problem and use-case searches | Relevant posts and the audience's own words | How big the niche is |
| Creator Search Insights, where available | Topics people search for, including content gaps | Whether a particular format will work |
| Creative Center trends | Wider hashtag, song and creator context | Organic results for your product |
TikTok describes Creator Search Insights as a tool that surfaces topics people search for and lets creators filter for content gaps: topics that are "searched for often, but aren't featured in a large number of videos on TikTok" (TikTok). Its Creative Center has a Trends section. Treat both as sources of leads, then open the individual posts.
Save the search that led you to each reference. Otherwise next week's research becomes another scroll you cannot repeat.
Which creators should you watch?
A small, fixed watchlist: a handful of creators in your niche and a few in adjacent ones that serve the same audience. Six direct creators and four adjacent ones is a practical start, not a representative sample.
Direct creators show you the familiar problems. Adjacent creators show you other ways to present them: a renter-focused cooking account may have structures a storage brand can use even though it sells nothing similar.
For each creator, review a run of consecutive recent posts, including the ordinary ones. Leaving them out removes the comparison that makes an outlier mean anything. Record the post URL, creator, publish date, the date you looked, views, topic, length, format, and any visible collaboration or paid label. Mark what you cannot see as unknown; public views say nothing about retention or sales.
How do you tell whether a TikTok video is an outlier?
Compare it with the same creator's typical post, not with other creators' views. Use the median of comparable recent posts, so one old hit does not inflate the baseline:
Outlier multiple = the video's views ÷ the median views of the creator's comparable posts
If a creator's comparable videos have a median of 8,000 views and one has 40,000, that video is a 5× outlier. (These are example numbers to show the arithmetic.) Leave the candidate out of its own baseline and note how many posts the median came from. Compare posts of similar age where you can; if you only have today's counts, narrow the date range and label the comparison approximate.
A 3× cut-off is a useful way to triage a busy watchlist. It is an editorial filter, not TikTok's definition of virality.
AdAnt's Strategist runs a faster first pass when it researches a niche: a video counts as an outlier only if its views are at least 5× the creator's follower count, and on TikTok it must also have at least 10,000 likes (on Instagram, at least 50,000 views). Videos that miss the bar are kept and labelled as references, not discarded. The median check above is the stricter follow-up for the candidates you shortlist.
Before you celebrate, check what else could explain the spike: an old post resurfacing, a celebrity, a giveaway, a collaboration, a news event or paid distribution. If you cannot tell whether a post was promoted, record that as unknown.
How do you know a format works beyond one creator?
Find the same structure working for independent creators with their own footage. This is the Repeat step. Search with the audience problem plus a plain description of the structure, and look for separate executions, not reposts of one original or posts from one campaign.
For each candidate, write four lines:
- Topic: what the video is about.
- Hook: what the first visible or spoken moment promises.
- Structure: the sequence that takes the viewer to the payoff.
- Execution: the people, setting, edit, sound or props that deliver it.
Two or three independent examples justify an exploratory test; they do not prove the format caused the results. Record counterexamples too. If the structure performs poorly on several relevant accounts, find out what differs before calling it a winner.
How do you turn a format into your own brief?
Keep the structure, change everything else, and make sure your product can honestly deliver the opening promise. This is the Fit step. Here is an example brief for a home-organizing brand:
| Brief field | Proposed execution |
|---|---|
| Audience | Renters with a small kitchen and no pantry |
| Opening | The cupboard that spills when opened |
| Promise | One usable shelf without drilling |
| Sequence | Problem → install → load → open again |
| Proof | Continuous footage of what fits and whether the door closes |
| Product role | The organizer is the change being demonstrated |
| Next action | Check the dimensions before buying |
| Test variable | Messy cupboard first versus finished shelf first |
The product has to actually fit the cupboard. If the proof needs an unrealistic setup, pick another structure. Film your own footage, wording and demonstration: a structural reference gives you no rights to someone else's video or testimonial.
How should you test a new TikTok format?
Change one major thing per comparison, usually the opening, and review every post at the same ages. Keep the audience problem, the demonstration and the call to action the same. Check results at fixed points such as one day and one week; those are reporting checkpoints, not the end of distribution.
TikTok says its For You feed recommends videos from signals like the videos people like, share and comment on, and that "neither follower count nor whether the account has had previous high-performing videos are direct factors" (TikTok). A small account can test formats; it just needs a fair comparison.
Track business actions separately from views, through tagged links or another attribution method you can explain. If sign-ups are not measured, do not report reach as a sign-up result. When a video disappoints, keep the lesson: the problem attracted the wrong audience, the opening promised more than the demonstration showed, or the result was hard to see.
When is the research done?
When you can name the audience, the evidence, the structure that transfers and the next test. Collecting another hundred links will not fix an unclear creative decision.
To run this on your own product, AdAnt's social content strategy researches your niche, flags outlier videos with their sources, and turns them into briefs. Ask it for the evidence alongside the recommendation, then let your own results decide what gets another version.
Frequently asked questions
How do I research a niche with very few obvious competitors?
Search the problem, setting and desired outcome rather than only the product category, then inspect adjacent creators serving the same people. A niche with few direct sellers can still have a large audience discussing the underlying problem.
Should I copy the video with the most views?
No. Choose candidates by relative performance, audience fit, independent examples and whether you can deliver the same promise. A creator's biggest video can be an unrelated exception.
Do I need a trending sound?
No. Research the structure separately from the soundtrack. If a reference depends on a particular sound, check that it is available and permitted for your use before building the production around it.
Can AI find TikTok formats for me?
AI can collect references, flag outliers and turn patterns into briefs. Ask it for post links, observation dates, the baseline it compared against and what data was missing, then watch the actual videos before you approve its reading of them.