What is a conversion lift test?
A conversion lift test is an experiment run inside Google Ads or Meta. A random group is held back from your ads and its conversions are compared with the group that could see them. The gap is the lift: the sales your ads caused.
By Joris van Huët, Founder & CEOUpdated 7 min read
Run the numbers for your store: the free holdout test planner.
A conversion lift test is an experiment that Google Ads or Meta runs on your campaigns. It holds a random group of people back from your ads, lets the rest see them, and compares conversions in both groups. The difference is the lift: the sales your ads caused, which is usually not the number in the ad report.
Google calls the two groups treatment and control. The difference in conversions between them is the lift, Google's help says. Meta uses an intent-to-treat approach. Its test group is everyone eligible to see the ads, including people who never scrolled as far as one. Its control group is not eligible at all.
What comes back is a short list of numbers. Google reports incremental conversions, relative lift, incremental cost per action and, if your conversions carry values, incremental ROAS. Meta reports the lift, a cost per incremental conversion and a confidence percentage. None of them is a click count.
What one store's data shows
One store's anonymised GA4 export, 1 January 2024 to 21 August 2026. It holds shares of revenue only: no ad spend, no order counts.
| What the export shows | Share of revenue | Source cell |
|---|---|---|
| Direct, in last click, first click and touched views | 57.7% | Channels sheet, Direct row |
| Paid Social, in all three views | 0.0% | Channels sheet, Paid Social row |
| Organic Video, in all three views | 0.0% | Channels sheet, Organic Video row |
| AI Assistant, in all three views | 0.0% | Channels sheet, AI Assistant row |
The usual answer makes a lift test sound like a verdict on all your marketing. Read next to one store's export, it is a verdict on the rows you can switch off.
Start with the biggest row. In this store, Direct holds 57.7% of revenue on the Channels sheet, in all three views. GA4 files a visit as Direct when someone uses a saved link or types your URL. No platform can hide Direct from a random half of your visitors. It is where sales land, not a lever you can pull.
A lift test still sees into that row. Meta says its lift tests count all conversions in the test and holdout groups. Google says Conversion Lift ignores the standard attribution rules. So if an ad prompted someone who later typed your address, that sale still counts for the test group. No click required.
Paid Social shows 0.0% in all three views on the Channels sheet. The export holds no spend, so it cannot say whether paid social ran. If it did, a zero in a click ledger is an open question, not an answer. Of these four rows, it is the only one a platform lift test can reach.
Organic Video and AI Assistant also sit at 0.0% on the Channels sheet. Neither is an ad, so no platform can switch it off for a random group. A conversion lift test has nothing to hold back there.
Cause is the one thing the export cannot show. It records where credit landed, not what would have happened without the ads. That is the one question a lift test answers, and only for the campaigns inside it.
Why does the usual answer mislead?
The usual answer stops at two groups and a gap. The trouble starts when the result arrives.
- Relative lift flatters a small base. Google divides incremental conversions by control conversions. Its help warns that very large relative lifts easily occur when the control group converts little. Read the incremental conversions before the percentage.
- Confidence is not size. Meta defines its confidence as the probability that the ads generate incremental conversions greater than zero. A high score says the lift is real. It says nothing about whether the lift is big enough to pay.
- A few campaigns understate the channel. Google says measuring only a subset of your campaigns reduces the chance of detecting lift. The holdback group can still meet the ads you left out.
- The ad report is a different ruler. Meta says lift results are not meant to be compared with campaign results in Ads Manager. Ads Manager counts conversions inside click and view windows. A lift test counts them between its own start and end dates.
When is a conversion lift test worth running?
When the answer can move real money and your account clears the bar. Both platforms set one.
Google's user-based study needs at least 1,000 observed conversions and a campaign budget of $5,000 USD or more. As a guide, Meta wants a campaign from the past year with $5,000 USD or more in spend and at least 500 conversions.
Meta also checks the signal. You need the Conversions API sending an event with an Event Match Quality score above 5, or another supported data source. That second route is unavailable to advertisers mainly targeting countries covered by the EU's ePrivacy directive. For a European store, that means the Conversions API.
Then count the cost. Google lets you hold back between 1% and 50% of the audience, and bigger holdouts carry a higher opportunity cost. Meta notes that lift tests can affect the cost per impression of the campaigns in the test. Patience matters too: Google has found up to a 17% drop in Absolute Lift when long-lag studies run under 14 days.
A test that cannot change a decision is an expensive way to feel informed. Write the decision down before you book it.
What to do this week
- Check whether Meta will run one. In Meta Ads Manager, find your biggest campaign of the past year and note its spend and conversions. Then open Experiments, where Meta says you can create a conversion lift test. Pass: it cleared $5,000 USD of spend and 500 conversions. Fail: it fell short, so a Meta lift test will likely come back thin.
- Check the Conversions API. In Meta Events Manager, open Data Sources, select your dataset and open its Settings tab. Pass: the Conversions API is already set up for your website. Fail: it is not, so set it up there before you book any lift test.
- Ask Google for a power estimate. In Google Ads, open the Lift studies tab under Campaigns, then Experiments, in the Campaigns menu. Select the plus button, choose Conversion Lift, then Based on users, and add your purchase campaigns. Pass: the study power reads 90% or more. Fail: it reads lower, so take Google's budget guidance or treat the result as directional.
Check the homework. Your GA4 Attribution paths export already holds the evidence. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes, for €99 once (excluding VAT), refundable within 30 days. Check the homework
Sources, 1 October 2026: About Conversion Lift (Google Ads Help); About Conversion Lift (Meta Business Help Center); Default channel group (Analytics Help); Differences between Conversion Lift test results and other reporting tools (Meta Business Help Center); Understand your Conversion Lift based on users measurement data (Google Ads Help); Facebook Lift metrics glossary (Meta Business Help Center); Set up Conversion Lift based on users (Google Ads Help); About Experiments (Meta Business Help Center); About confidence in your tests and experiments (Meta Business Help Center); About Meta Events Manager (Meta Business Help Center).
Related answers
Frequently asked questions
Is a conversion lift test the same as a brand lift study?
No. A conversion lift test counts actions such as purchases or subscriptions. A brand lift study asks survey questions about awareness, recall or purchase intent. Both hold a group back from your ads, but only conversion lift measures sales.Why does a lift test show fewer conversions than Ads Manager?
Because it counts something else. Ads Manager credits conversions to ads inside click and view windows. A lift test compares all conversions in the test and holdout groups and keeps only the difference. Meta says the two are not meant to be compared.What confidence level should a conversion lift result have?
Meta treats 90 percent or higher as a statistically reliable lift result. Google suggests aiming for 90% certainty, and calls results between 50% and 90% directional. Below that, treat the number as a hint, not a budget decision.
Go deeper: Causal attribution, explained.
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Keep reading
Terms in this article
- AnalyticsAnalytics is the systematic computational analysis of data. It reveals customer behavior and measures campaign performance.
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Control GroupControl Group is a segment of an audience intentionally not exposed to a marketing campaign, used to measure the campaign's true causal impact.
- ConversionConversion is a specific, desired action a user takes in response to a marketing message, such as a purchase or a sign-up.
- ExperimentsExperiments are scientific procedures that test hypotheses or demonstrate facts. In marketing, experiments like A/B tests determine the causal effect of campaign changes, enabling data-driven decisions.
- Google AdsGoogle Ads is an online advertising platform where advertisers bid to display ads, service offerings, and product listings.
- ImpressionAn Impression counts each time an ad or content displays on a user's screen. It measures exposure, not engagement.
- IncrementalityIncrementality measures the true causal impact of a marketing campaign. It quantifies the additional conversions or revenue directly from that activity.