What is incrementality?
Incrementality is the difference advertising makes. It measures how many conversions, sales or visits happened because of a campaign, compared with what would have happened without it. If 1,000 people bought after seeing your ads but 800 would have bought anyway, only 200 sales were incremental. Those 200 are the true return on your spend.
Incrementality answers a simple question that most reporting skips: did the ads cause this, or just show up next to it?
How incrementality is measured
Every incrementality test compares two groups that are alike in every way except ad exposure. The test group can see the ads, and the control group cannot. The gap in results between them is the lift. Common methods include:
Holdout tests: a random share of the target audience is held back from the campaign.
Ghost ads or ghost bids: the system logs when a control user would have been shown your ad, without showing it, which gives a cleaner comparison.
PSA tests: the control group sees an unrelated ad, such as a public service message. This is simpler but costs media.
Geo experiments: ads run in some regions and not in matched regions, then sales are compared.
How to calculate incremental lift
Incremental lift = (test conversion rate - control conversion rate) / control conversion rate
For example, if the test group converts at 2.4% and the control group at 2.0%, the lift is (2.4 - 2.0) / 2.0 = 20%. With 100,000 people in the test group, that is about 400 incremental conversions. Divide spend by 400 to get the incremental cost per acquisition, which is often much higher than the CPA in platform reports.
Why incrementality matters
Standard reports give credit to ads that appear near a conversion, even if the person was already going to buy. This makes some tactics look better than they are. Retargeting and branded search are common examples, because they reach people who are already close to buying.
Incrementality testing shows where budget actually adds sales. That helps you compare channels on true ROAS, cut spend that only takes credit and scale tactics that create new demand.
A few rules keep tests reliable: run them long enough to reach a meaningful sample, change one thing at a time and make sure the control group really stays unexposed.
Incrementality vs attribution
Multi-touch attribution splits credit for a conversion across the touchpoints a person saw. It describes paths, but it cannot prove that any touchpoint caused the sale. Incrementality uses a control group to measure cause and effect. Attribution is useful for day-to-day optimization, and incrementality checks whether those decisions hold up. Many teams also use brand lift studies for upper-funnel goals, and a data clean room to match exposure with sales data safely.
Incrementality with AdGeeks
AdGeeks traders can help design holdout or geo-based tests for campaigns in DV360, Amazon DSP and Adform, set clean control groups and read the results with you. The goal is simple: put more budget where the ads change outcomes.
