How in-store retail media is measured, and where the proof breaks

A play-log proves that a media player rendered a specific file at a specific time on a specific screen. It is a delivery record, not an audience record. It says the inventory ran, which is a necessary check but not evidence that anyone was standing there.

Play-logs became the default because they are cheap, automatic and hard to dispute. The problem is that they answer a vendor question rather than a buyer question. A brand reconciling spend wants to know reach and outcome, and the log has nothing to say about either.

A zone estimate models traffic past a placement, usually from store traffic counts or sensors, then converts that traffic into impressions. It is modeled reach: reasonable at the store level, weak at the placement level, and impossible to tie to a named shopper.

Estimates are useful for planning and dangerous for pricing. Two placements in the same store can be quoted at similar CPMs while performing very differently, because the model does not see dwell, attention or basket. Buyers should ask which parts of a quoted number are counted and which are modeled.

When the shopper opts in through a QR code or a messaging channel inside the store, exposure attaches to a real person rather than a modeled one. Reach becomes countable, frequency becomes manageable, and follow-up messaging becomes possible with permission already in hand.

HRM builds identification into the store experience, where the relationship with the shopper already exists. That is what separates attributed reach from blind inventory, and it is the reason measured audiences price differently from screens on a wall.