What is incrementality testing?
Incrementality testing is an experiment: keep ads away from a random group of people or regions, run them for everyone else, and compare sales. The gap is the lift, the sales your ads caused. No gap usually means the ads were collecting sales you would have had anyway.
By Joris van Huët, Founder & CEOUpdated 6 min read
Run the numbers for your store: the free holdout test planner.
Incrementality testing is an experiment that shows which sales your ads caused. You keep ads away from a random group of people or regions, run them as usual for everyone else, and compare sales. The gap is the lift. If there is no gap, the ads were usually collecting sales you would have had anyway.
That is the textbook answer, and it holds. Google's help calls incrementality experiments tools to understand how effective your ads are at driving an action at a certain point in time. One group sees your ads and the other doesn't. The difference in conversions is the lift.
Tests come in two shapes. A user-based test splits people, and the ad platform decides who sees what. A geo test splits places: in Google's geo experiments, non-overlapping regions are randomly assigned to a control or treatment condition. Both ask the same thing. What would have happened without the ads?
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 |
| All channels in the touched view, added up | 110.4% | Channels sheet, Touched column total |
| Journeys with 1 touch (0.5 days to buy) | 79.5% | Journeys sheet, 1 touch row |
| Paid Social, in all three views | 0.0% | Channels sheet, Paid Social row |
In this store, Direct holds 57.7% of revenue whether you credit the first touch, the last touch or every touch. Three rules, one answer: the journeys that touched Direct both began and ended there. GA4's own models only credit Direct when the whole path is Direct, its help page says. No click-based rule can say what sent those buyers. A friend, a podcast, a video ad nobody clicked: the path is silent.
A holdout is not silent. Switch a paid channel off in some regions, and if Direct sales fall there too, that channel was feeding Direct.
The touched view in the export adds up to 110.4% by design, because a journey that met two channels counts in both. That is credit doing what credit does: it adds up to more than you sold. Each ad platform also counts by its own rules, so their claims can overlap the same way.
Journeys with 1 touch carry 79.5% of revenue in the Journeys sheet, at 0.5 days to buy. In a one-touch path, every model hands that touch all the credit. Only a test can say whether those buyers needed it.
Paid Social sits at 0.0% in all three views on the Channels sheet. That can mean no paid social ran, or that it ran without tags. If it did run, a holdout is how you would find sales it pushed into other rows, such as Direct.
What the export cannot show is cause. It has no spend and no order counts, so it can't price a channel or size a test. It shows where credit landed, while incrementality testing asks what each channel caused.
Why does the usual answer mislead?
The usual answer stops at the definition and adds: ask your ad platform to run a lift test. Good advice, with three catches.
First, each platform tests its own ads. Google's lift study measures the value your campaigns create on top of any other active marketing. Handy, but no platform's test will tell you a rival spends your money better.
Second, the door has a bouncer. Conversion Lift isn't available for all Google Ads accounts. The user-based version asks for at least 1,000 observed conversions and a minimum campaign budget of $5,000 USD.
Third, a test is a photo, not a film. Google frames these experiments around a certain point in time, so a spring result says little about December.
Skipping the test for clever modelling rarely saves you. In 663 large-scale experiments at Facebook, two non-experimental methods tried to recover the measured effect, and neither performed well. At eBay, field experiments found brand-keyword ads had no measurable short-term benefits. The same researchers found paid search returns were a fraction of the usual non-experimental estimates.
What can a lift number not tell you?
A lift result answers one question about one set of campaigns, at one spend level, in one window. It can't say what the next euro would do. That takes a separate test that adds budget in some regions.
It can't see sales that land after the test ends, so the test has to outlast your buyers' decision time. It can't rank channels it didn't test. And a flat result can mean the ads add nothing, or that the test was too small to see them. That is why Google shows a feasibility status before a geo study starts.
Most of all, lift is not profit. A channel can add real sales and still lose money on them. Read the lift next to your margin, not next to the platform's ROAS.
What to do this week
- Find your Direct share. In GA4, open Advertising, then Attribution models under Attribution, and note Direct's share of revenue. Pass: Direct is a small slice, so most journeys have touches a test can probe. Fail: Direct is your biggest row, so plan a holdout on the paid channel you suspect feeds it.
- See which channel's credit swings. In Shopify admin, open the Growth page and select View channel report. Switch the attribution model between first click and last click. Pass: one channel's sales change a lot between models, so it is your first test candidate. Fail: nothing moves, so your journeys are short and only a test will tell you more.
- Ask Google whether it will run the test. In Google Ads, open Lift measurement in the Goals menu and select the plus button. Pass: Conversion Lift appears under Based on geo, with a feasibility status. Fail: it doesn't, so ask your Google account representative or plan a regional holdout yourself.
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). Measuring Ad Effectiveness Using Geo Experiments (Google Research). Get started with attribution (Analytics Help). Set up Conversion Lift based on users (Google Ads Help). Set up Conversion Lift based on geography (Google Ads Help). Close Enough? A Large-Scale Exploration of Non-Experimental Approaches to Advertising Measurement (arXiv). Consumer Heterogeneity and Paid Search Effectiveness: A Large Scale Field Experiment (NBER). Key event attribution models report (Analytics Help). Measuring marketing performance (Shopify Help Center).
Related answers
Frequently asked questions
Is incrementality testing the same as A/B testing?
No. An A/B test usually compares two versions of an ad, and both groups see ads. An incrementality test keeps one group away from your ads altogether, so the gap shows what the ads added rather than which version won.Can GA4 tell me whether my ads are incremental?
Only in part. GA4's data-driven model compares paths with and without a touch, but it shares credit among touches it recorded. A sale from an ad nobody clicked can land on Direct. A holdout is what shows the sales your ads added in total.What does an incrementality test cost?
Mostly the sales you give up in the group that sees no ads, plus the weeks the test runs. Google's user-based Conversion Lift also asks for a minimum campaign budget of $5,000 USD.
Go deeper: Causal attribution, explained.
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Keep reading
Terms in this article
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Attribution ModelAn Attribution Model defines how credit for conversions is assigned to marketing touchpoints. It dictates how marketing channels receive credit for sales.
- 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.
- Holdout TestA holdout test is an experiment where a portion of the audience does not see a campaign. This measures the campaign's true incremental impact.
- IncrementalityIncrementality measures the true causal impact of a marketing campaign. It quantifies the additional conversions or revenue directly from that activity.
- Incrementality TestingIncrementality Testing measures the additional impact of a marketing campaign. It compares exposed and control groups to determine causal effect.
- Lift MeasurementLift Measurement: A method to determine the incremental impact of a marketing campaign by comparing exposed and control groups.