Most advertisers are measuring the wrong thing and making budget decisions based on the result. Incrementality testing exists to fix that, and the industry is finally moving toward it at scale.
eMarketer and TransUnion data from July 2025 shows 36.2% of marketers plan to increase incrementality spending over the next 12 months. Among retail media buyers specifically, 71% rank incrementality as their most important KPI. And 27.6% of U.S. brand and agency marketers identify expanding incrementality testing as a top measurement priority.
The urgency behind those numbers stems from a specific problem. eMarketer’s research shows that platform-based attribution routinely miscalculates advertising ROI because, when performance is measured within the platform that sold the media, advertisers are relying on a scorekeeper with a stake in the outcome. And today’s media environment has a growing focus on understanding what is truly driving results, not just what is generating activity.
Incrementality testing, hold-out groups, matched-market experiments, and mixed media modeling all produce ROI estimates that are not contingent on any single vendor’s reporting. As these methodologies gain traction, the barrier to entry for incrementality testing continues to fall. Google recently reduced the minimum budget for incrementality experiments from $100,000 to $5,000 by leveraging Bayesian statistical models. What was once accessible only to major advertisers is now within reach of mid-market buyers.
The AI angle is worth noting carefully. A WARC study found that AI is accelerating measurement capabilities, but advertisers still lack a complete framework for ROI evaluation. Technology speeds the analysis. It does not replace the strategic framework needed to interpret and act on it.
At Media Manager, we treat measurement as a media channel in its own right. It requires investment, expertise, and ongoing iteration. Plans that skip that investment may be optimizing blind, and we cannot manage what we do not accurately measure.
