Conversion lift study requirements: what Meta and Google ask for
The broader decision is covered in incrementality for mid-size ecommerce brands.
What does a platform lift study do?
It measures lift: the difference in results between a group that could see your ads and a comparable control group that could not.
Meta's documentation says it randomises the audience and assigns accounts to a test group or a control group1. Google's help centre calls Conversion Lift an incrementality tool and offers two types: one builds groups from users, the other from locations2.
The output is a result you can turn into incremental ROAS (iROAS): revenue caused by the ads, divided by spend. That is a stricter number than platform-reported ROAS.
One control method you will meet in this context is ghost ads: the system records when a control-group user would have seen your ad and shows them a different one instead.
Why do e-commerce brands use them?
Because they are the nearest thing to a platform-approved answer to "did the ads cause these sales?"
They are also the easiest to start. No R code, no regional sales data to prepare. The platform has the data.
That convenience has a catch, covered below.
What are the requirements?
Here is what we can say, and what we cannot.
Meta Conversion Lift
The Marketing API guide says Conversion Lift measurement is currently limited and that access goes through your Meta representative; measurement partners can also set studies up for advertisers1. The same guide says a results breakdown only displays with at least 100 conversions across test and control combined1. Third-party guides quote entry thresholds that differ from each other. We do not repeat them. Ask Meta for the current requirements for your account.
Google Ads Conversion Lift
Google Ads Help's About Conversion Lift page lists no budget thresholds and tells you to contact your Google account representative2. Its page comparing lift types says the geo-based budget requirement tends to be higher, and the split is noisier with no breakdown by demographics3. It also says user-based studies rely mostly on opt-in traffic, so measurement gaps can affect them3. Trade press reports that Google lowered its budget threshold and added feasibility ratings4. Ask your Google representative for the current amount.
What is a conversion lift feasibility rating?
A feasibility rating (high, medium or low, as trade press describes it) is a pre-test estimate of how likely a conclusive result is. It does not guarantee one4.
What decides whether a study can see anything?
In every case, the key technical question is the minimum detectable effect (MDE): the smallest lift the design can reliably detect. If your spend is too small to see a plausible lift, the study will come back unclear.
How do we implement it?
Check the conditions
Is spend large enough for the MDE? Is conversion tracking trustworthy? Can you hold ads back from a control group? If not, wait.
Ask the platform
Ask your Meta or Google representative for the current requirements on your account. Do not rely on thresholds quoted in third-party guides.
Design around a decision
Choose the question, the cells and the duration so that the answer can change the budget.
Read a range, then compare
Treat the result as a range, not a single number. Where possible, compare it with a test the platform did not design.
When should you not use them?
Skip them, for now, when:
What does it cost in effort?
No money figures here. The real cost is what you give up: the control group does not see your ads, so any sales those ads would have driven there are lost for the test period.
The effort:
- Access. Getting the platform to enable it.
- Design. Choosing the question, the cells and the duration.
- Clean tracking. The result is only as good as conversion data.
- Reading. Someone has to interpret a range, not a single number.
Pitfalls
Treating access as a guarantee
Eligibility is not a result.
Reading feasibility as a promise
It is an estimate4.
Taking "No lift" as failure
Google says it does not prove the ads failed6.
Trusting the platform's "new customer" label
That customer may already be in your CRM5.
One platform, one view
Each platform measures only itself.
Quoting thresholds from third-party guides
They disagree.
Our take: Selective
We rate platform lift studies Selective. This is our view, not a market fact.
They are worth running when you can get access, the design is feasible, and the decision depends on the answer. They are not a neutral referee. An AdExchanger piece argues that a platform's incrementality reporting may be directionally useful but should not decide on its own whether sales are incremental or customers are new5. A customer the platform has never converted may already be in your CRM5.
Read results with care. Trade press reports that Google's interface shows "No lift" when its certainty is low, and that Google says this does not prove the ads failed6.
Our view: where possible, compare a platform result with a test the media seller did not design or calculate.
FAQ
Who can run a Meta conversion lift study?
Meta's API documentation describes access as limited, via a Meta representative1.
What is the difference between user-based and geo-based lift?
User-based splits people. Geo-based splits regions. Google says geo needs a higher budget and is noisier3.
Does a feasibility rating guarantee a result?
No. Trade press describes the ratings as likelihoods of a conclusive result, not promises4.
Does "no lift" mean the ads did not work?
Not necessarily. Google says a "No lift" label does not prove the ads failed6. It can mean the test could not tell.
Can I trust a platform's own lift study?
It is useful evidence, but the platform designs and grades it. Trade press warns against letting a platform decide alone what is incremental5, so compare with a test the platform did not design where you can.
Sources
- Lift Study (Meta Marketing API)Primary source
- About Conversion Lift (Google Ads Help)Primary source
- Comparing lift types (Google Ads Help)Primary source
- PPC Land: Google lowers incrementality testing threshold to $5,000Blog or media
- AdExchanger: Beware the platform incrementality trapBlog or media
- PPC Land: Conversion lift (przegląd)Blog or media
Let’s find what’s actually working.
Next step. Want a view on whether a lift study fits your budget? Talk to us about your budget decision. Or read about our Measurement & incrementality service. You can start with a free measurement audit.