Our Creative Testing Framework for Meta Ads
The creative testing framework we run across eCommerce ad accounts: batches over single ads, the retention benchmarks we watch, iteration trees, and kill rules.
On this page
Most brands don't have a creative testing problem; they have a creative guessing problem. New ads get launched one at a time, judged on feel, and killed or scaled on numbers nobody agreed to in advance. Creative is the new targeting, and the brands that win treat testing as a system with a scoreboard, not a stream of hunches.
This is the framework we run across our eCommerce ad accounts. It's the same system we teach on camera, and you can run it in a few focused hours a week.
Test batches, not ads
The unit of work is the batch: a few distinct concepts, each cut into several hooks, launched together. Concepts test the angle, and hooks test the first three seconds, which decide everything downstream. The cheapest test in advertising is the same ad body with a new opening line. If you brief creators, the UGC anatomy is the construction each variation should follow.
Read the engagement graph, not your gut
Before an ad earns budget, its retention graph has to earn belief. Across our accounts we look for a hook rate above 40%, roughly a third of viewers still watching at four seconds, and 15 to 17% still there at ten. Those are our internal benchmarks from our own spend, not a platform standard, and yours will drift by format and product. The point isn't the exact numbers; it's that the graph tells you where an ad breaks, so you fix the exact beat that loses people instead of rewriting the whole ad.
Grow winners with iteration trees
When an ad wins, don't frame it; branch it. An iteration tree takes the winner and varies one element per branch: the hook, the creator, the setting, the offer beat. Each branch either beats the parent or teaches you which element carried the win. Over a few cycles the tree turns one lucky ad into a repeatable pattern, and the pattern feeds the next batch's brief. That's the creative flywheel: test, read, branch, feed back.
Keep score with hit rate and kill rules
Two disciplines keep the system honest. First, track your hit rate, winners divided by ads tested, batch over batch; it measures the process, not the ad. Second, write the kill rule before launch, a cost target and a fair spend window, and follow it. The one exception runs through unit economics: an ad buying customers whose cohort pays back deserves more time than its channel metrics suggest.
The 7 Levels, on camera
We map this whole progression, from random testing to a repeatable creative flywheel, as the 7 Levels of Meta ad creatives. The full walkthrough is here:
Want this system run for you?
We run this framework across eCommerce accounts every week: sourcing creators, cutting batches, and reading the graphs. Book a discovery call and we'll audit your current creative and show you what's leaking revenue.
Frequently asked questions
- How many ad creatives should we test at once?
- Test in batches, not singles. A batch of variations built from a few concepts and several hooks gives the account enough to compare, and gives you a hit rate you can actually learn from. One new ad at a time teaches you almost nothing.
- What is a hit rate in creative testing?
- Winners divided by ads tested. It measures whether your creative process works, instead of celebrating one lucky ad. If the hit rate rises batch over batch, the process is learning; if it's flat, you're gambling.
- When should we kill an ad?
- Set the kill rule before the ad launches, based on your own cost targets and a fair spend window, then follow it without renegotiating. The exception is an ad whose cohort economics look better than its channel metrics; that one earns more time.
Want this run for your brand?
Hayes Media builds direct response creative, buys the media, and runs the email & SMS behind it.
Book a discovery call
