Closed-loop measurement

Closed-loop measurement is attribution in which the same organization that serves or sells the advertisement also observes the purchase, so that exposure and transaction can be joined deterministically without a third-party match.

Retail and commerce media networks are built on it. A grocer knows which loyalty member saw a sponsored product and which loyalty member bought it; a card issuer knows who saw a card-linked offer and who paid with the card. The IAB's 2024 guide on advanced measurement calls incrementality "essential for measuring true marketing impact, especially within retail's closed-loop ecosystem" (IAB, 31 October 2024). The loop closes only inside the network's own estate; a purchase at another retailer, or on a channel the network does not run, is invisible, which is why offsite claims need a clean room or a modelled extrapolation that the IAB/MRC guidelines require to be disclosed (IAB/MRC, January 2024).

Closed loop is not the same as incremental. Deterministic attribution tells you that an exposed shopper bought; it does not tell you whether she would have bought anyway. The standards therefore treat closed-loop attribution (ROAS) and incrementality (iROAS) as separate metrics with separate disclosure rules.

Example: a pharmacy chain serves a vitamin brand's display ad to 2 million app users, matches 180,000 purchases of the SKU to exposed users within a disclosed 14-day window, and reports attributed sales; a 10% randomised holdout shows the exposed group bought 6% more than the control, which is the incremental figure.

Related: incrementality, ROAS, iROAS, retail media network, clean room.

Source: IAB/MRC, "Retail Media Measurement Guidelines" (January 2024); IAB, "Retail Media Advanced Measurement and Data Collaboration" (31 October 2024).

1. IAB, 31 October 2024
2. IAB/MRC, January 2024