Why 42% of viral TikToks won't sell your app

The numbers:

• 5,896 TikToks that founders and our team saved as winners
• 4,671 of them were real outliers, 10x or more their account's normal views
• 1,978 of those outliers (42%) had comments with no interest in a product at all

Genuinely viral. Useless for selling an app.

Where the numbers come from

Handler keeps a library of TikToks people flag as worth copying. Every video gets scored against its own account's median views, and Handler reads the comments to decide whether viewers want the thing being shown or just liked the video.

That's two of the three checks in the VSC framework, the filter I built while growing Flamme and that we now run on every brief at Agniverse Media. Viral, Scalable, Convertible. So we ran the whole library through it.

Finding 1: raw views hide the best ideas

1,052 of the 4,671 real outliers had fewer than 100,000 views.

Sort a niche by views and you lose more than one in five of the formats that actually broke out, because they came from small accounts. 800K views on an account that usually gets 4M is a normal day. 40K on an account that usually gets 2K is a signal.

Finding 2: viral isn't convertible

42% of real outliers failed Convertible. The comments were about the sound, the joke or the feeling. Nobody asked what the app was.

It doesn't get better with size. Among outliers with a million views or more, 49% failed.

A period-tracking video in the library did 313K views at 153x its account's normal. The comments were people wishing each other luck. Lovely, and nobody asked about the tracker.

Finding 3: the format changes the odds

Share of outliers whose comments showed interest in the product:

• App demo: 72%
• Slideshow: 60%
• Talking head: 59%
• Silent UGC: 48%
• Text-heavy: 33%

App demos win because the product is the story. Silent UGC pulls huge views, but half the time people are there for the mood, so the hook has to point at your app or the views go nowhere.

What we do with this

Nothing gets filmed until it passes all three checks:

• Viral. 10x its own account, and the format has worked for more than one creator.
• Scalable. A creator who isn't you can make 30 versions from a brief.
• Convertible. The comments ask about the product.

Then every hook gets 5 to 10 videos and the numbers decide: kill, mutate or scale. Of the 87 hooks on our board that have reached a decision so far, 21 got scaled. The point of VSC isn't that every idea wins. It's that the losers stay cheap.

The full method, with the scoring rules, the board numbers and my walkthrough video, is on the VSC framework page.