Why is retargeting ROAS inflated?
Retargeting's reported ROAS is almost always inflated relative to its real incremental impact, for the same underlying reason brand search is: it's re-reaching people who already showed real intent, some meaningful share of whom would have converted anyway without ever seeing the retargeting ad. Crediting the ad with the full value of that conversion overstates what the spend actually produced.
How often should retargeting ads show?
Frequency capping matters more in retargeting than almost any other format, because the same small, already-engaged audience sees the ads repeatedly. Uncapped retargeting frequency commonly produces ad fatigue and a real brand-annoyance cost that rarely shows up in standard platform reporting.
How long should a retargeting window be?
Retargeting window length (how long after a visit someone remains eligible to be retargeted) should match actual purchase-cycle length for the product. A 30-day window on a same-day-decision product wastes spend retargeting people well past the point they were still considering the purchase.
Why Retargeting matters
Retargeting is really useful for closing warm prospects, but its reported efficiency is one of the least trustworthy numbers in a paid media account without incrementality testing. Treating its platform-reported ROAS at face value routinely overstates the channel's real contribution to revenue.
Testing real incrementality
A retargeting campaign reports an 8x ROAS, among the account's best-performing line items. Holding retargeting ads back from a control group of otherwise-eligible visitors for four weeks reveals that a majority of the 'attributed' conversions happened at a similar rate in the control group too. So most of that reported return was purchases that would have happened without the ad. The real incremental lift was a fraction of the reported number.
Common mistakes
Trusting retargeting's reported ROAS without an incrementality test
Retargeting re-reaches people who already showed intent, so a real share of 'attributed' conversions would have happened anyway. A holdout test reveals the real incremental contribution.
Running retargeting with no frequency cap
An already-engaged audience seeing the same ad too often produces fatigue and brand annoyance that standard reporting doesn't capture as a cost.
Using a retargeting window mismatched to the purchase cycle
A window far longer than the product's typical consideration period wastes spend reaching people who've moved past the decision, while a window too short can cut off prospects who are still deciding.
Where we work on this