Beyond Last-Click: How We Prove Incremental ROAS in Retail Media

Retail media is evolving from a gold rush of first-mover advantage to a focus on measurable business outcomes.
Incrementality—the isolation of ad-driven sales from organic behavior—has become the top metric for 71% of advertisers, yet many still rely on legacy attribution.
Traditional methods are manual, revenue-limiting, and incompatible with machine learning.
Moloco's AI-native onsite ads platform uses ghost bidding, a randomized controlled trial (RCT) technique, to measure incremental return on ad spend (iROAS) during live campaigns.
Ghost bidding identifies when an ad would win an auction but withholds it from a control group, comparing behavior between exposed and eligible-but-unexposed users.
This ensures clean causal measurement without disrupting optimization.
Results from early experiments show iROAS of 253% to 1,609%, with incremental conversions per exposed user ranging from 4% to 29%, revealing wide variation across partners.
Best practices include ensuring statistical significance, defining clear populations, counting all conversions, controlling for external factors, and conducting regular tests.
As platforms integrate incrementality into ad serving, privacy-compliant, cookie-less methods will become more accessible.
For advertisers, moving from volume-based metrics to value-based decisions is key to breaking growth ceilings.