We studied 194 startup launch videos. The best 20 got 82 million views. The worst 20 got 323.

We scored each video on the craft of the video itself, joined those scores to what the posts actually did — views, likes, bookmarks, replies — and hand-verified that every one is a genuine startup launch, not a big company shipping a feature. One pattern runs through the whole set: the videos that failed weren’t confusing. They were dead.

00 · The formatThe launch video is new. The launch is not.

The launch video traces back to Steve Jobs, who turned a product spec into an event and made the reveal itself the story. But for the two decades after him, that was rare. Most startups launched with a post. Founders wrote it — for Hacker News, for Product Hunt, later for LinkedIn — and picked over every line, maybe with an image or a short blog post attached. Then they spent most of their energy on press: chasing journalists, hiring an agency, and judging the day by how much coverage they got.

That changed in the last two years or so. The launch moved from the post to the video. The style grew up in crypto and the NFT era, when a loud, high-production clip was how a project got noticed on Twitter. It carried over to AI, and now most teams skip the PR agency altogether. They spend the money on the video instead, because the video lets them tell their own story, their own way, in one thing they control.

How founders make it varies a lot. Some shoot the whole thing over a weekend on almost no budget. Others hire a studio and spend tens, sometimes hundreds, of thousands of dollars. For them the video is how they introduce the company to the world: it sets the brand early, builds FOMO with investors, and stakes a claim on what they want to be known for. It can bring in users too. But this is where the data gets honest: whether a viral launch actually turns into users depends a lot on the product. Not every startup benefits the same way, and the numbers here show it.

MethodHow we built the Index, and what it can’t tell you.

We studied 194 real startup launch videos on X, hand-checked so every one is a genuine startup launch, not an established company shipping a feature. We scored each on about 20 craft traits, hook strength and type, pacing, emotional pull, surprise, rewatchability, aspect ratio, differentiation, and more, then matched those scores against how the videos actually performed. The view-ranked findings below use the 155 launches old enough to have settled (30+ days); newer ones count toward the study but sit out of the reach comparisons. Every number here is current as of 17 August 2026, and each one shows up twice from here on: as a sentence, and as a row in the table under its chart, so it’s easy to check or quote.

Read these as patterns, not promises. We report medians, not means, so a few multi-million-view outliers can’t drag a number around. And reach has a lot of causes, the audience, the timing, the size of the account posting, and plain luck, so treat every finding here as a strong correlation, not a guarantee. The one thing this data can’t settle: whether emotional pull causes reach, or whether the founders who can make a video that hits are simply the ones who already have the audience to carry it. Our honest read is that both are true at once.

01 · The clarity trapClear is common. Emotion is rare.

The easy story is that launches flop because nobody understood the product. The data doesn’t back that up. Nearly nine in ten videos were clear. Clarity is table stakes — nearly everyone clears it, winners and losers alike. What almost nobody had was emotional pull: only 3 videos in the whole study hit it hard, and just 42 managed even a medium amount.

So clarity can’t be what separates a 14-million-view launch from a 16-view one, because both were clear. The founders who lost weren’t fighting a comprehension problem. They were fighting the wrong problem.