Big question
When a platform reports a sale, how much of it did the media cause?
The position
A platform-reported return is a claim on sales rather than a measure of effect. Rank channels only inside one measurement you control, and move money on a holdout or a calibrated model.
Why it is live
Every seller reports the sales it can associate with its own exposure, and association is not causation. The IAB and IAB Europe incremental measurement guidance from November 2025 draws that line explicitly, which a market observing the line would not need written down.
The case for using reported returns
DocumentedThey are computed on real exposures and real orders, and they are the only figures available at the speed a plan runs.
The case against ranking on them
InferredEach return was produced by a different party, over a window that party chose, against sales that party could see a different share of, so moving the money changes the number you moved it for.
Where the weight sits
Decisively against ranking channels on reported returns. The arithmetic is correct and the comparison is invalid, and that sentence is the whole answer.
What would change this
One party with no stake measuring every channel the same way in the same month, or a clean room standard that makes the windows and the counting rules identical.
Where this is taught
Connected Evidence
Sources
- IAB and IAB Europe incremental measurement in commerce media, 3 Nov 2025secondary