Paid Social
The first frame is the whole ad
Ninety-four percent of the decision happens before your logo appears. Here is what we test in frame one, and what we stopped testing entirely.
Lampy StudioHead of Performance6 min
Look at enough matched pairs of winning and losing ads in the same account and the same variable keeps separating them. It is not length, not format, not spend. It is what happens in the first 0.8 seconds.
That is not a new claim. Every platform's creative guidance says something like it. What is new is how specific the pattern gets once you have enough matched pairs to look at.
What actually worked in frame one
- Texture or material at macro distance, before any face appears
- A number on screen that contradicts an assumption the viewer holds
- Motion that starts mid-action — no establishing, no wind-up
- A hand entering frame doing something with the product
- On-screen text that names the objection instead of the benefit
The unifying idea: frame one should answer 'why is this on my screen' before the viewer consciously asks it. Every opener above does that with information, not with production value.
What we stopped testing
Logo stings. Aspirational establishing shots. Slow push-ins on a founder about to speak. Category explainers. These test reliably below account average across published creative studies, and we do not intend to spend a client's budget re-confirming it.
1.1%
Beauty category median CTR — the number a texture-first opener has to beat
How to run this on your own account
Take your three best-performing ads. Cut the first second off each and replace it with four different openers, changing nothing else. Run them against a holdout for fourteen days. You will learn more about your audience from that one test than from a quarter of full-concept testing.
If the first frame does not carry information, the rest of the ad is a rendering exercise.
The uncomfortable part is that this makes creative cheaper, not more expensive. Once you know which openers work, the rest of the ad becomes a versioning problem — and versioning is the one part of production that scales without a proportional cost increase.