Howwecrackvirality

Almost every brand that comes to us is optimising the wrong ninety percent of the video. They have spent the budget on the edit, the grade, the music bed and the call to action, and none of it matters, because the audience they are trying to reach never got past the first second.

Short-form ranking is a series of gates. The platform shows a clip to a small seed audience and watches what they do. If enough of them stay past the opening, it widens the pool. If they do not, the clip is finished, and no amount of craft in the back half will save it. So the work is front-loaded almost to the point of absurdity.

The three gates

What we actually test

We hold the body of the clip constant and vary the opening. One idea becomes a set of variants that differ only in the first two seconds: a different line, a different first frame, a different delivery. Everything downstream is identical, so the difference in performance is attributable.

This is the part that is hard to do with human creators and easy to do with a fleet. Asking a person to shoot the same video sixteen times with different openings is a miserable ask and produces sixteen increasingly tired takes. Generating sixteen openings costs an afternoon.

You are not looking for the best video. You are looking for the opening that survives the seed audience, and then you feed it.

Hook families that keep working

• The correction — 'I did this wrong for years' — sets up an error the viewer might be making too.
• The overheard — 'my pharmacist told me to stop doing this' — borrows authority without claiming it.
• The cost — 'this cost me $400 to learn' — prices the information, which implies it is worth having.
• The enumeration — 'the third one is the mistake' — creates a countdown the viewer has to stay for.
• The refusal — 'I am not going to tell you the brand' — withholds the obvious thing, which is what everyone stays for.

None of these are clever. That is the point. Clever openings perform badly because they take a beat to parse, and a beat is the whole budget. The families above work because they are legible at speed and they all do the same structural job: open a gap and refuse to close it yet.

Why volume is the strategy, not a shortcut

Hook performance is not stable. What works in a category in March is exhausted by June, because everyone copies what worked and the audience develops immunity. There is no library of winning hooks you can buy once. There is only a rate at which you can discover the current ones.

That reframes the whole problem. You are not trying to make a viral video. You are trying to run enough attempts, fast enough, with clean enough attribution, that you find this month's openings before your competitors do. Volume is not a substitute for judgement — it is what makes judgement possible, because it is the only way to get a signal that is not noise.