Home goods DTC · 4× creative output, 58% CAC drop on Meta
Meta CAC had doubled in six months. The brand was shipping 4–6 new ads per month, none testing against a clear hypothesis. Fatigue was setting in across the audience by week two of every launch. The team thought they needed a creative agency. They needed a production pipeline.
- Mapped top-performing past creative into 6 hook categories — problem, authority, founder, result, comparison, social proof
- Built a production cadence shipping 5–8 ads per week against hypothesis tests, not vibes
- Pre-tested hooks via 3-day micro-budget tests before committing to full production
The ads the brand had been shipping all looked the same after a while — variations of the same founder-in-kitchen format the team liked. Meta’s algorithm needs diversity of hook, not diversity of background color. We audited twelve months of performance and mapped every winner into one of six hook categories. Three of those categories had been under-produced against. That’s where we went first.
Production got rebuilt around hypothesis. Every new ad started with a written brief naming the hook, the audience objection, and the metric we expected to move. No more “just make more founder videos.” Scripts went through a 24h feedback loop with the media buyer before shoot. Shoots happened weekly, not monthly.
The real unlock was micro-testing. Before committing to full production, we’d run three-day whisper tests — $200 against a rough mock-up — to validate the hook. About 40% of hooks didn’t clear. Those never went to production. The 60% that did had much better hit rates once produced.
Four months in, creative output was 4.1× higher and CAC had dropped 58%. Not because the creative got prettier. Because every ad was testing something specific, and the losers got killed before they spent real money.