Two hundred ads. Ninety days. One question: does creative volume actually correlate with ad performance — or does it just feel productive?
The brand in question sells a $34 home organization product. Small team, no in-house creative, aggressive growth target. They committed to a full quarter of AI UGC production, tracking every render, every launch and every outcome.
Here's what they found.
Eleven winners out of two hundred is a 5.5% hit rate. That sounds low until you consider the alternative: producing three or four creator-based ads per month, where each individual ad carries 25% of the testing budget and the cost of being wrong is enormous.
What Actually Happened
The brand ran a mix of formats. They started with traditional AI UGC video ads at 25–35 seconds for feed placements, then expanded into AI UGC Shorts after month one. By month three, the split was roughly 60% Shorts, 40% standard length.
The winning creative was overwhelmingly in the Shorts bucket. Nine of the eleven winners were under 20 seconds. The brand's prior assumption — that longer ads performed better because they had more "room to sell" — was backwards for paid social on TikTok and Reels.
Four Learnings
Shorts beat feed ads almost every time
Short-form AI UGC outperformed standard-length ads by a wide margin on hook rate and CAC. The shorter format forced tighter scripts, and tighter scripts meant faster hooks — which meant higher retention.
Background mattered more than actor quality
The team initially assumed better AI actors would drive performance. The data disagreed. Ads with casual, native backgrounds — kitchens, bedrooms, real-looking living spaces — consistently beat ads with higher-fidelity actors in studio settings.
Winner variation mattered more than winner production
Once a hook won, reproducing it with small changes (different actor, different opening line, different product angle) was a reliable path to additional winners. The team eventually formalized this: one winning hook generated 4–6 variations before moving on.
Judgment, not output, became the bottleneck
By month three, producing ads was trivial. Deciding which ones to iterate on and which to kill was the real constraint. The brand hired a creative strategist specifically to close this gap — a role that didn't exist on their team before the experiment.
The Tooling Setup
The brand used a pay-as-you-go platform for most of the run, with a short trial of a higher-fidelity tool for comparison. The choice mattered less than expected — what mattered was AI UGC video throughput and the ability to produce Shorts quickly without reworking the workflow between formats.
They also ran a small parallel set through MakeUGC specifically for hook variation, and a premium set through CreateUGC for one high-stakes launch. Neither was wrong — just different use cases.
What They'd Do Differently
Two things, both hindsight-obvious:
Start with Shorts, not standard-length. The team spent the first month producing 30–35 second ads that underperformed their eventual Shorts work. In retrospect, the shorter format should have been the default from week one.
Set a kill rule earlier. The brand kept some ads running for a full week because they looked polished. The data would have killed them by day two. Good-looking creative that underperforms is still underperforming — the polish doesn't earn extended runway.
The Real Conclusion
Two hundred videos in 90 days is not a heroic achievement. It's the new baseline for teams that have switched to AI UGC production. What made the quarter successful wasn't the volume — it was the decision to stop treating each ad as precious.
The eleven winners are the headline. The bigger outcome is that the brand can now find eleven more winners next quarter without re-learning how to do it.