What is Incrementality? Measuring True Growth
Why attribution lies and how to measure the marketing that actually drives new revenue.
In the world of digital marketing, “attribution” is the metric everyone watches. But attribution is often misleading. It tells you which touchpoint a customer interacted with before buying, but it doesn’t tell you if that interaction caused the purchase.
Incrementality is the measure of the true lift: the conversions that would not have happened without a specific intervention.
The Attribution Mirage
Imagine you run a pizza shop. You hand out coupons to people walking in the front door. At the end of the night, you count the coupons and say, “Wow, these coupons generated 100% of our sales today!”
This is exactly what happens when you rely solely on platform-reported ROAS (Return on Ad Spend) or standard attribution models, especially for retargeting or branded search. You are often paying to acquire customers who were already walking in the door.
Correlation vs. Causation
The core problem is distinguishing between correlation (someone saw an ad and bought) and causation (someone bought because they saw the ad).
If a user visits your site, leaves, and then gets retargeted with an ad before returning to purchase, the ad platform claims credit. But if that user was simply checking their bank balance before buying, the ad spent money but generated zero incremental revenue.
How to Measure True Lift
To understand the real value of your marketing, you need to move beyond last-click attribution and look at incrementality.
1. Holdout Tests
The gold standard. Take a random percentage of your audience (e.g., 10%) and prevent them from seeing your ads. Compare the conversion rate of the exposed group vs. the control group. The difference is your incremental lift.
2. Geo-Lift Experiments
If you can’t split audiences by user ID, split them by geography. Turn off spending in a specific region (like a state or city) and measure the drop in total sales compared to a control region that matches its historical performance.
3. Marketing Mix Modeling (MMM)
For complex environments where testing is difficult, statistical models can analyze historical data to determine how different channels contribute to the baseline sales volume.
The Goal: Profit, Not ROAS
When you optimize for incrementality, your reported ROAS might actually go down, but your total profit goes up. You stop spending money on “cheap” conversions that were happening anyway (like branded search terms) and start investing in harder-to-measure channels that actually bring new customers into your ecosystem.
True growth comes from knowing the difference between activity and impact.
Want to Know What's Working In Your Marketing?
Get a free marketing audit. We'll show you exactly where you're losing customers and wasting budget.