Every brand running retargeting has seen this: remarketing campaigns show 8–15× ROAS while prospecting shows 2–3×. The natural conclusion is to cut prospecting and scale retargeting. This is the ROAS trap, and it destroys acquisition efficiency over time by systematically underfunding the campaigns that create demand while overfunding the campaigns that capture it.
The ROAS trap
Retargeting shows high ROAS because your retargeting audience is people who were already going to buy. You are paying to serve an ad to a conversion that was happening anyway, the organic search, the direct visit, the email click that would have happened without the retargeting impression. Your ROAS is measuring what happened, not what you caused. The technical term for this distinction is incrementality: the revenue that happened because of the ad, minus the revenue that would have happened without it. Most advertising measurement ignores this entirely.
Incremental ROAS: the real number
Incremental ROAS asks a different question: what revenue happened because of this ad that would not have happened otherwise? The only rigorous way to measure this is a holdout test: withhold the ad from a control group, compare conversion rates between the exposed group and the control. The delta is your true lift. For most retargeting campaigns at growth-stage companies, incremental ROAS is substantially lower than blended ROAS, sometimes by 60–80%. The budget that looked like a great investment is capturing demand you already created through prospecting and brand, not generating new demand.
How to run an incrementality test on Meta
Meta has a built-in Conversion Lift test tool. Structure: (1) Define your test campaign, typically retargeting or a specific prospecting campaign you want to measure, (2) Set a holdout percentage at 20–30% of the audience, these people will not see the ads during the test, (3) Set a minimum budget and run for at least 14 days to reach statistical significance, (4) Compare conversion rates: test group versus control group. The percentage lift is what your ad actually drove. A campaign with 0.3× incremental ROAS is capturing existing demand, not creating new demand, it is a budget drain disguised as a high performer.
What to do with the results
If retargeting shows low incremental ROAS, cut the budget significantly. The conversions will not disappear, they will happen via organic, direct, or email. Reallocate the freed budget to prospecting, which typically shows lower blended ROAS but higher incremental ROAS because it is reaching people who would not have converted otherwise. Run this test once per quarter for any campaign that represents more than 15% of total spend. The goal is not the highest number in the weekly report. It is the most revenue that would not have existed without the ad spend. Those are different things, and most reporting frameworks measure the first while the second is what you are actually paying for.