Retargeting almost always shows the best ROAS in the account. It also almost always deserves the least credit for that ROAS. Someone who already visited your product page, already added something to cart, or already searched your brand name was probably going to convert with or without the retargeting ad reminding them. The dashboard gives the ad full credit anyway.
The test we run before trusting retargeting numbers
We've held out a portion of a retargeting audience from ads entirely — a genuine control group, not just a smaller budget — on several accounts, then compared conversion rates between the group that saw ads and the group that didn't. The gap was consistently smaller than the platform's reported ROAS implied. On one account, the "true" incremental lift from retargeting was about a third of what the dashboard suggested. The rest was sales that would have closed regardless.
This isn't a reason to kill retargeting. It's a reason to stop treating its reported ROAS as gospel when deciding how much budget it deserves relative to top-of-funnel spend that's actually generating new demand instead of just capturing demand that already existed.
Where retargeting budget is usually genuinely wasted
Audiences retargeted for too long. Someone who visited a site 45 days ago and never returned is a much colder lead than the dashboard's audience definition treats them as. Long look-back windows dilute genuinely warm retargeting audiences with a lot of people who have completely lost interest.
No frequency capping. We've inherited accounts where the same person was shown the same retargeting ad 20+ times in a week. Past a certain frequency, more impressions don't increase conversion — they increase ad fatigue and occasionally outright annoyance, while still burning budget with each impression.
Retargeting people who already converted. Sounds obvious, and it's one of the most common issues we find in inherited accounts regardless. Excluding recent purchasers from retargeting audiences is a five-minute fix that immediately stops wasting spend on people who have nothing left to buy.
What we do instead
We size retargeting budgets based on estimated incremental lift, not raw reported ROAS — meaning smaller than the dashboard alone would suggest, with the difference reallocated to prospecting campaigns actually reaching new audiences. We shorten look-back windows to focus spend on genuinely warm traffic. We cap frequency aggressively and exclude recent converters by default, not as an afterthought.
None of this makes the retargeting ROAS number in the dashboard look as good. It does make the account's actual growth — new customers acquired, not just existing intent captured — look considerably better, which is the number that was supposed to matter in the first place.