Why Vertical Shorts Convert 4x Better — With Data
Every marketing blog on the internet has told you that vertical video outperforms horizontal. Almost none of them have shown you the numbers, and even fewer have explained why.
We pulled data from public reports (Meta Business, TikTok Creative Center, YouTube Shorts Insights) and combined it with anonymized aggregate data from our own users. Here is the actual picture.
The 4x claim, examined
The headline number — that vertical converts roughly 4x better than horizontal on mobile — is real, but it hides two very different effects:
- Attention capture. Vertical fills the screen. Horizontal video on a mobile feed occupies at most 30% of the visible area. That alone accounts for something like 2-2.5x of the difference.
- Session context. People who are scrolling a vertical feed are already in "consume" mode. People watching horizontal video on mobile are usually doing so from a link they intentionally clicked. Different intent, different conversion baselines.
So the "4x" is partly a real quality difference and partly an apples-to-oranges comparison. Both matter, and both work in vertical's favor.
The three-second rule is actually two seconds
Across platforms, the strongest predictor of a vertical short's performance is retention at the 2-second mark, not the 3-second mark that everyone quotes. If you lose viewers before 2 seconds, no clever hook later will save the video. If you hold them past 2 seconds, you tend to hold them to at least 8.
Practical implication: front-load your visual hook. Do not wait for the "cold open" to finish. The first frame matters more than the first sentence.
Captions are not optional
Across every platform we looked at, videos with burned-in animated captions retained 34-58% longer than videos without. This is not because people cannot hear — most viewers do have sound on. It is because captions give the eye something to track while the message lands. They are a visual scaffolding, not just an accessibility feature.
The best-performing captions share three traits:
- One or two words per frame
- High contrast, thick outline
- Timed to speech rhythm, not sentence structure
Length is a distribution, not a rule
The common advice — "keep it under 15 seconds" — is wrong on the top end. The actual distribution of high-performing shorts is bimodal: strong performance clusters around 7-12 seconds and around 45-60 seconds. The dead zone is 20-35 seconds, where videos are too long to feel snappy and too short to develop a real story.
If your content is close to that dead zone, either tighten aggressively or expand into a proper narrative arc. The middle is where retention collapses.
Repurposing beats producing
One of the clearest patterns in the data: creators who cut vertical shorts from existing long-form content outperform creators who shoot vertical-first, at every experience level. The reason is simple — long-form footage has been selected against a higher bar of quality, and the best 8 seconds of a 20-minute video are almost always better than 8 seconds shot standalone.
If you have a back catalog, cut it up before you shoot anything new.