Should we use platform lift studies (Meta, Google) or independent geo tests?
Our answer
Use a platform's free lift study for questions inside that platform once you qualify. Use an independent geo test for sales it can't see, comparisons across channels and big budget moves. Check one against the other now and then.
- Confidence
- Medium
- Strongest proof
- 1 controlled test
- 12 official rules, 5 practitioners, 8 opinions or claims
- Sources
- 16 sources
- 26 statements
- MG
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentThe platform's study is a sound, free experiment. Meta's randomizes people and costs nothing extra once you meet its guide of $5,000 of spend and 500 conversions. Sellforte (marketing mix modeling firm) calls it credible for what it measured, though you can't audit the assignment.
It can't see sales off the platform or compare channels. Meta counts only conversions it can see, so Amazon and retail sales need a geo test that reads your own sales (WorkMagic). Olivia Kory of Haus adds that Meta's study can't be compared across channels.
A geo test costs more to run. Google says geo experiments often need more budget and time than user-based lift. Common Thread Collective (CTC) uses a user-based lift test around Black Friday because it needs a smaller holdout.
Our judgment
Use the platform's study for questions inside one platform and a geo test for total sales and comparisons across channels. We combine Sellforte's starting advice of Meta Conversion Lift for Meta and geo-lift for Google search, Blck Alpaca's split of the two by question and Olivia Kory's point on channel comparisons. Run a geo test when sales land off the platform or before a large budget move, and use one now and then to check the platform's study. That follows Sellforte's advice to validate with occasional matched geo-lift studies, WorkMagic's note on what Meta can see and Rockerbox's Hourglass test, which read only site sales. Charlie Cole's and Cody Plofker's rule against letting anyone grade their own homework argues for the independent check, not against using the platform's study. Sellforte saw no systematic pattern of Meta's result always running higher than geo tests, and Meta's researchers found no meaningful imbalance across 3,204 Lift tests.
Our judgment, built from Sellforte, Blck Alpaca, Olivia Kory (Haus), WorkMagic, Rockerbox, Charlie Cole (The DTC Podcast), Cody Plofker and Meta: our reading of these sources, not a finding from any one of them.
What the sources say
26 statements from 16 sourcesMeta's Conversion Lift has no extra cost. As a guide, the account needs a campaign from the past year with $5,000+ spend and 500+ conversions, plus Conversions API with event match quality above 5.
“There’s no additional cost for the test, but ad campaigns or sets you use for conversion lift tests may have certain budget requirements.”
Meta's Conversion Lift matches conversions from sources such as the Conversions API and offline uploads to test and control users. It recommends the Conversions API and lists no retailer sales source.
“Conversions API is the recommended data source.”
Meta advises against using the post-test window to calculate Conversion Lift results, saying it can add variance and shrink the measured effect once holdouts end.
“While we offer the option, we do not recommend using the post-test conversion window for calculating results.”
PPC Land reports Google's Conversion Lift studies do not accept click-ID-only offline imports or GA4-imported conversions, and Google recommends one to two weeks of conversion data before launch.
“GCLID-only imports are insufficient”
Google says user-based Conversion Lift does not support web conversions imported from Google Analytics or GCLID imports without hashed PII, and asks for 7 to 14 days of conversion data first.
“GCLID imports without hashed PII can't be used in Conversion Lift, because the control group is not served an ad to click on.”
Meta's GeoLift guide recommends daily data, the finest targetable geography, at least 25 pre-test periods across 20 or more regions, history 4 to 5 times the test length, and other media held constant.
“The test duration should cover at least one purchase cycle for the product.”
Stella's pricing page lists a free plan with one mix model and one holdout analysis a year, then plans at $750, $3,000 and $6,000 a month. The $750 plan targets brands at $5M to $10M a year.
“One media mix model and one holdout analysis a year, run on data you upload.”
Google's geo-based Conversion Lift also goes through a Google rep, needs campaigns targeting a single country, and rates feasibility. Google advises against running a study rated low.
“We don't recommend proceeding with low feasibility status”
Meta says A/B tests and lift tests cost nothing extra. A/B tests compare up to 5 versions with no overlap, Conversion Lift needs a good data source such as Conversions API, and lift tests can change CPM.
“Lift tests can affect the overall cost per impression (CPM) of ad campaigns in your test depending on your test type and bidding strategy.”
Google's user-based Conversion Lift needs a Google rep, at least 1,000 observed conversions and a $5,000 minimum budget. Google allows 7-day studies but recommends more than 14 days.
“We've found up to a 17% drop in Absolute Lift in studies with a long conversion lag that run for less than 14 days.”
Luke Austin prefers geo holdouts in general, but for Black Friday suggests conversion lift tests, which need smaller holdouts and carry less risk to the business.
“Geo holdout testing is our preferred sort of approach but it requires larger holdouts”
Google says its free, open-source Meridian GeoX library for geo experiments on any ad platform is now generally available, and its results can feed into a mix model.
“Now, GeoX is generally available globally in Meridian, so you can get started with running independent experiments, or incorporating incrementality results into your MMM to help boost accuracy.”
Sellforte says no test type is best: use Meta Conversion Lift for Meta, geo-lift for Google Search, Shopping and Performance Max, and customer holdouts for email, push and direct mail.
“There is no single best universal test type.”
Google's GeoX guide calls for at least 10 geos for a single-cell test, ideally 50 to 100+, daily data only, and says geo tests usually cost more and run longer than user-based lift tests.
“You should include a minimum of 10 geos for a single-cell design”
Sellforte says a well-powered Meta Conversion Lift test is credible for what it measured, but advertisers cannot audit assignment, so check it with occasional geo-lift studies before calibrating a model.
“We have not seen a systematic pattern in which the Meta result is always higher.”
Blck Alpaca, an agency, says Meta Conversion Lift suits comparisons inside Meta, while geo-lift suits budget moves between channels; it relays about $500,000 a year in spend as a geo holdout threshold.
“Use platform Conversion Lift for comparisons inside one system. Use geo-lift for larger channel and budget questions.”
Cody Plofker's first independent check on TV is the post-purchase survey, scaled up by the response rate (if half of buyers answer, double it), then compared with other attribution models.
“never allow anyone to create their own homework”
Alex Neiman, who sells a course and AI agents, says to size Meta first with a Conversion Lift or geo test, then use campaign holdouts; prospecting usually carries the lift and retargeting flatters itself.
“Prospecting almost always carries it; retargeting almost always flatters itself.”
Elevarus, an agency, says that above about $80K a month per channel you should run 4 to 6 week geo holdouts for incremental ROAS; below about $40K, use Meta's lift tool or skip formal testing.
“Use a geo holdout when you need to test cross-channel effects, when offline conversions are part of the outcome, or when the platform-native option doesn’t cover the campaign type.”
Charlie Cole says brands should never let an ad platform or vendor grade its own homework.
“you should never let any advertiser or vendor grade their own homework.”
CTC says it runs always-on Meta conversion lift tests comparing settings head to head, such as incremental attribution versus 7-day click, alongside third-party geo holdouts and mix models.
Olivia Kory of Haus suggests starting with Meta's free Conversion Lift tool; its weaknesses are signal loss and a method that cannot be compared across channels.
WorkMagic says Meta's Incremental Attribution counts only conversions Meta can see, so sales on Amazon or in stores go uncredited; omnichannel brands need a geo test or a calibrated mix model.
“It only measures what Meta can see.”
In a six-week, 10-market Meta pause run by Rockerbox, Hourglass Cosmetics' site sales showed 0.51 incremental ROAS. Sephora sales are about 8x site sales, so the 4 to 5 total is an untested estimate.
“These results measure ecommerce impact only.”
A study of 3,204 Meta Lift tests and 181,890 A/B tests found Lift tests balanced, while A/B tests deliver to different audiences by design. To test creative alone, keep every other setting identical.
“Lift tests show no meaningful audience imbalance, confirming their causal validity, while A/B tests show clear imbalance, as expected”
TikTok's Conversion Lift Study splits the audience into two groups to measure lift. It is a managed service for eligible accounts that meet minimum spend, and the page gives no spend figure.
“As an exclusive managed service, we work closely with eligible accounts to design and implement studies that align with specific business goals.”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer