You don’t know which creative works until you burn it in production
You launch a creative on Meta Ads, spend €1.500 in 4 days, only to find the hook doesn’t land. Meanwhile, another creative would have performed twice as well. Traditional creative testing is expensive and slow.
We combine pre-testing with synthetic audiences and automatic A/B testing in production. You know what will work before you spend, and you refine in production without having to think about it.
How we build it
1. Pre-test with synthetic audiences
We generate a virtual panel of your buyer persona (10-50 synthetic profiles with Claude). Each one sees the creatives in random order and scores them on expected engagement, clarity, and brand fit. Results in 30 minutes, not 4 days.
2. Pre-launch filtering
Creatives that clearly don’t resonate are discarded before going live on Meta. Those that show promise go into production. 80% of the risk is screened out before spending a single euro.
3. Automatic A/B testing in production
Surviving creatives are launched in Meta Ads with balanced small budgets. The system cuts losers within 48h and scales winners. No waiting to see “what Meta’s algorithm decides”.
4. Semantic analysis of winners
The system breaks down what makes a creative win: hook, CTA, image type, dominant colour, copy length. You learn patterns, not anecdotes, and the next creatives are designed with those patterns in mind from the outset.
What changes when this works
- 3-5x more efficiency in spend (higher ROAS at the same spend, or the same ROAS at lower spend).
- Creative iteration speed: from fortnightly to near-daily.
- Your creative team stops launching creatives blind — they get objective feedback.
- Pause decisions are made in hours, not weeks.
When we do NOT recommend this
- If your ad volume is low (<€2k/mes): there is no statistical signal for a reliable A/B test.
- If you launch 2-3 creatives per quarter: the setup isn’t worth it.
- If your buyer persona is ultra-specific (CTO of industrial companies in Pais Vasco with 50-200 employees): synthetic audiences don’t represent it well.
Honesty about synthetic audiences
Synthetic audiences do not replace the real market. They are a prior filter that reduces risk, not an absolute truth. Final decisions are always made by real A/B testing in production.
Stack we use
- Claude 4.5 para generación de audiencias sintéticas.
- Meta Marketing API + Google Ads API para testing en producción.
- Custom semantic analyzer para descomposición de creatives ganadores.
- Looker Studio para reporting del test.
We start with a diagnostic session
Before quoting the full setup, we run a 90-minute session. We look at your monthly ad spend, your current creatives and your buyer persona together, and we come away with an honest recommendation: whether this service fits where you are right now, or whether it makes more sense to start with something else.
We don’t charge for that session. If you’re interested, get in touch.