Same creative, two audiences: the cleanest A/B test you’re not reading
Every agency argues about creatives: “the video is bad” versus “the audience is wrong”. Most of those arguments are settled with opinions. There’s a cleaner way, and your account already contains it.
The natural experiment
When the same creative runs in two or more ad sets, you have a controlled test nobody planned: content identical, audience different. If the same video does 12.6% CTR in one ad set and 7% in another, the argument is over — the creative works; the second audience (or placement) is the problem.
On one account we audited, five ads were about to be labeled “weak” — until grouping by creative showed each of them performing at least 1.5× better somewhere else on the same account. Killing them would have thrown away proven content. Reshooting them would have wasted a production budget on a targeting problem.
Why nobody reads it
Ad platforms report per ad, and the same creative gets a different ad ID in every ad set. Unless something actively groups by the underlying video or image, the comparison never surfaces. ROASTED groups every report by creative ID and, whenever the gap is 1.5× or more, writes it directly on the weaker ad’s card: the same creative performs better in ad set X — move budget, don’t reshoot.
Do this on your account today
- Pick your three “worst” ads.
- Check whether the same video runs anywhere else in the account.
- If it performs there — you don’t have a creative problem. You have a budget allocation problem, and those are much cheaper to fix.