Creative testing cadence for finance brands on Meta
How often to ship new ads, how to read the data, and when to kill a 'good' performer.
Most finance brands on Meta in 2024 were either over-testing or under-testing creative, and the consequences of either were expensive. The right cadence for a bridging account at our typical spend levels sat in a narrower window than most agencies were running. This is the discipline we settled on through 2024.
Why the cadence matters
Ship too little and your winners fatigue, CPMs climb, and the account stalls. Ship too much and you can never give a creative enough budget to actually prove itself, so you spend all your time killing ads before the data is meaningful.
For a bridging account spending £15-£40k a month, three to five new creatives per fortnight was the sweet spot. Less than that and the account stagnated. More than that and signal got noisy.
How long to wait before judging
Minimum 500 link clicks per creative before declaring a winner or loser, and minimum 3-4 days of delivery regardless of click volume to clear day-of-week effects. A creative that looks great after 24 hours often regresses, and a creative that looks dead at 24 hours sometimes recovers.
We had a hard rule against killing creatives before the minimum thresholds were hit. It cost us occasional spend on duds. It saved us several winners that would have been killed early.
When to kill a 'good' performer
Frequency above 2.5 with rising CPM on a stable audience is the signal that a winner has saturated. Kill it before it starts dragging account efficiency down rather than after.
The instinct to keep a winner running is strong because the cost-per-result still looks good. The trailing data is misleading - the audience is fatiguing faster than the in-platform reporting catches up. Better to retire a winner one week early than one week late.
Creative testing cadence is more discipline than science. The biggest single improvement we made to account performance in 2024 was not a better-performing creative. It was a more consistent testing rhythm that compounded over months. Cadence beat brilliance most weeks.
Capital Edge publishes one note a month on UK bridging finance, paid acquisition, and AI-led outbound. Written for brokers, by the team running the playbook.