Case #007: The seven-day clock

Case #007: The seven-day clock

A health supplement offer, $3,200 budget, $34 CPA, and one genuinely great video. Days 1 through 6 it ran a $22 CPA — clean profit, and I made the rookie move of assuming a winner stays a winner. I left it running and went to find the next campaign.

Day 7 the CPA was $29. Day 9 it crossed $34 and went negative. The ad hadn't broken; the audience had simply seen it. Frequency had climbed to 4.6, and the same people who converted in week one were now scrolling past it on sight. The creative was fatiguing on a clock I wasn't watching.

The loss window cost me real money — days 7 through 10 spent $640 and returned $510 before I caught it.

The fix turned into a system. I mapped frequency against CPA across that run and found the inflection point sat near frequency 3.8, roughly every six days at my spend level. So I built a rotation: three creatives queued, swap the lead ad before it hits 3.8 frequency, let the rested ones cool, cycle back. Same audience, refreshed eyes.

Full arc, 18 days: $3,200 spent, $5,180 back. 62% ROI — most of it earned after I started treating a winning creative as a perishable thing with a shelf life.

The lesson: a winning ad isn't a winning ad, it's a winning ad for about six days — watch frequency, not just CPA, and rotate before the clock runs out.