Meta's Conversion Lift test measures the incremental effect of your Facebook and Instagram advertising on people who could have seen your ads. A true incrementality test, typically a geo holdout, measures the effect of all your advertising on business outcomes regardless of platform.
The name "Conversion Lift" sounds like it measures everything. It does not. It measures the incremental effect of advertising within Meta's ad ecosystem, as reported by Meta. That is useful but limited, and it has a significant conflict of interest built in.
How Meta's Conversion Lift works
Meta randomly assigns a portion of your target audience to a holdout group that does not see your ads during the test. At the end of the test, Meta compares the conversion rate of the exposed group to the holdout group and reports the lift.
Meta has visibility over who did and did not see ads, and it controls both the experiment and the reporting. The methodology is sound in principle, but the execution and reporting are entirely within Meta's control.
Where Conversion Lift falls short
- Meta reports lift within its platform using its own data. It cannot account for the cross-channel effect of your overall media mix.
- Meta has a financial incentive to show positive lift results: positive results encourage continued spending.
- Conversion Lift does not suppress all digital advertising, only Meta ads. If the holdout group is still seeing your Google or programmatic ads, the comparison may be contaminated.
- Meta's measurement relies on its own pixel and conversion data, which is subject to signal loss from iOS privacy changes and ad blockers.
What a true incrementality test does differently
A geo holdout test suppresses your Meta advertising in a set of geographic regions while keeping it running in matched regions. You then measure actual business outcomes, such as total revenue or online purchases, rather than conversions as reported by Meta.
This approach is platform-independent. The measurement is based on your own first-party sales data, not Meta's attribution system. The holdout is geographic rather than audience-level, which removes Meta from the equation entirely.
In independent tests comparing Meta's Conversion Lift results to geo holdout results, geo holdouts consistently produce lower, and likely more accurate, lift estimates. The gap is often 30 to 50 percent. This does not necessarily mean Meta is being dishonest: the methodologies genuinely measure different things. But you should not treat platform lift results as a substitute for independent measurement.
When to use each approach
Use Meta's Conversion Lift as an operational diagnostic. It is useful for comparing ad set performance, testing creative variations, and getting a directional signal quickly without the complexity of setting up a geo test.
Use a geo holdout test when you need a reliable, defensible answer to the question: "Is our investment in Meta advertising generating real incremental business value?" This is the question that should drive budget decisions.
The portfolio view
Even if Meta is genuinely incremental at the platform level, it does not mean your current spend level on Meta is optimal. You could be getting incremental lift from Meta while simultaneously over-investing relative to other channels that generate higher iROAS.
Platform-level lift tests cannot tell you about the opportunity cost of your spending. Only a cross-channel view, combining geo holdouts across channels with a marketing mix model, gives you that.
Is Meta's Conversion Lift test free to use?
Meta does not charge directly for running a Conversion Lift test, but the holdout group does not see your ads during the test period, which may reduce conversions from that group. The cost is in the potential revenue you forgo from the holdout, not a platform fee.
How long does a Meta Conversion Lift test take?
Meta recommends a minimum of two weeks and typically four weeks for reliable results, depending on your conversion volume. For low-volume advertisers, the test may need to run longer to reach statistical significance. Meta provides a pre-test power estimation tool in Ads Manager.
Should I use platform lift tests to calibrate my MMM?
With caution. Platform lift results tend to be higher than independent geo holdout results, so using them directly as calibration inputs may inflate your MMM's channel coefficients. If you use platform lift results, consider applying a discount factor or weighting them less heavily than independently run geo tests.