In-Store Retail Media Measurement: Why the IAB Europe Draft Starts From Footfall, Not Ad Plays

Somebody on your team has to type a number into the audience column of the in-store rate card this week, or into the post-campaign report a supplier's brand manager is waiting for. The number that is easiest to find is the play log from the content management system: how many times the 15-second spot ran across how many screens. The number that is hardest to defend is also that one, because a play log counts the screen, not the shopper. The IAB Europe In-Store Retail Media Definitions and Measurement Standards V2, which opened for public comment on 17 September 2026, takes a clear position on this: ad plays are not the starting point for measuring in-store retail media. Footfall is.

What the draft actually asks you to multiply

Read the draft as a buyer would. An in-store impression is defined as visual contact with the ad. The simplified opportunity to see (OTS) calculation then reduces to six inputs:

• Average daily store footfall
• Number of stores
• Number of days
• Number of placements in store
• Aisle penetration
• Share of voice

Four of those six are already on your campaign booking form. Aisle penetration is the one that needs a method, and the draft is pragmatic about where it can come from: beacons in every store, beacons in a sample of stores, inference from category transactions, or a one-off market study. But every chain in the formula hangs off the first input. If the daily footfall figure is wrong by 20 percent, every impression you report to every supplier is wrong by 20 percent, and no amount of precision in share of voice repairs it.

This is a draft, not a finished standard. Comments close on 23 October 2026, and anyone running a retail media network should read it before then, because the methodology that lands here is the one supplier agencies will quote back at you in 2027 negotiations.

The grocery shortcut the draft leaves open

Here is the part worth commenting on. The draft notes that in grocery the starting point for footfall is often a transaction, and it does not require independent verification of the footfall figure. In practice that means a chain can take yesterday's basket count, call it footfall, and feed it into the OTS formula without anyone checking.

A transaction count is not footfall. It leaves out the person who walked the length of the store, stood in front of the endcap screen for six seconds, and left without buying. It leaves out the partner pushing the trolley while one basket is paid for. It leaves out the shopper who came in for a prescription at the pharmacy counter and was never scanned at a till that counts toward the retail media figure. For in-store advertising, these are exactly the people the supplier is paying to reach: the ones who saw the message and are not yet customers of that brand. A method that excludes them undercounts the audience in general and, worse, undercounts it unevenly, because the gap between visitors and baskets differs by format, by daypart and by category.

The direction of the error is not always in the retailer's favour either. Where the convention is one basket equals one visitor, a store with high group shopping is undersold. Where someone has applied a flat uplift factor of 1.3 to baskets to approximate visitors, a quiet weekday store is oversold. Either way, the first question a supplier's analyst will ask is: how did you get the footfall figure? If the answer is tills, the conversation about CPM starts from a weaker position.

Why an annual average breaks in November

The draft's first input is average daily store footfall. The word average hides a second problem. Sensormatic expects the ten busiest days of the 2026 US holiday season to carry 30 to 40 percent of store traffic for the period. A two-week in store retail media campaign booked across those days and priced on a yearly daily average will under-report its audience badly; a January campaign priced on the same average will over-report. Electronics and DIY see the same compression around their own peaks, and fashion sees it around sale weekends.

The fix is not a better average. It is a footfall figure that exists per store, per day and per hour, so that the number of days in the formula is multiplied by what actually happened on those days, in those stores. Suppliers buying dayparts, for example a breakfast cereal brand wanting the 07:00 to 10:00 slot, will ask for hourly footfall sooner or later. Retail media measurement that can only answer at the level of the week is going to look dated quickly.

What a defensible footfall input looks like

The audience number that survives a supplier's questions is built from door counts. Each entrance is counted, in and out, in both directions, aggregated to the hour and stored per store. From there the OTS inputs fall out naturally: the number of days is a sum of real daily counts rather than a multiplication, and placements near a given entrance can be weighted by that entrance's share of traffic rather than the store total.

One thing implementers learn early and buyers rarely ask about: staff. In a large grocery or DIY store, colleagues cross the customer entrances dozens of times a shift, collecting trolleys, bringing stock from the yard, going on break. Uncorrected, staff movements can add several percentage points to the daily count in exactly the stores where the margin matters. Ask your counting provider how staff are excluded, whether by badge, by zone logic or by a documented adjustment, and ask to see the figure before and after. If the answer is a shrug, the footfall input is softer than it looks.

The same discipline helps when the sales window changes. Reporting that already runs on daily counts is indifferent to whether the supplier wants 30 days or 28; a chain that recalculates averages by hand is not. We covered a category-level version of this in Turning Pharmacy Shelf Space Into Measurable Retail Media, where the dispensing counter and the front shop have very different visitor profiles and a single store-level figure hides both.

Three questions to settle before 23 October

• Where does our footfall input come from today? If it is tills or an estate-wide estimate, decide whether that is the position you want on record when the draft becomes a reference document.
• Can we produce it per store, per day, per hour? Not in principle, but as an export a supplier's agency can open on Monday morning.
• Should we comment? The absence of a verification requirement for the footfall input favours chains with weak data in the short term and penalises the whole channel in the long term. Retailers with counted traffic have an interest in saying so.

For readers weighing whether counting at the door is worth the investment: Vemco Group has more than 55,000 installations in 98+ countries, counting 60+ million people daily, and more than 20 years of doing it. Contractually we commit to a minimum of 96 percent counting accuracy; in practice, where lighting, store layout and visitor behaviour allow, it typically runs at 98 to 99 percent. Local support means the engineer who calibrates the staff exclusion for your flagship is reachable when a supplier queries the numbers.

Frequently asked questions

How is an in-store retail media impression defined? In the IAB Europe draft open for comment until 23 October 2026, an in-store impression is visual contact with the ad. That rules out counting screen plays as impressions, since a play with no shopper in front of it generates no contact.

How do you calculate opportunity to see for in-store screens? The draft's simplified formula uses six inputs: average daily store footfall, number of stores, number of days, number of placements in store, aisle penetration and share of voice. Footfall is the base that everything else scales, so its quality determines the quality of the final figure.

Why is store footfall the starting point for in-store audience measurement? Because the audience for in-store media is everyone who enters the store, not everyone who buys and not every time a screen refreshes. Starting from footfall ties in-store audience measurement to a count of people, which is how other media channels are bought.

Is a transaction count the same as footfall? No. A transaction counts baskets, not people, and excludes everyone who saw the screen and bought nothing, as well as companions of the person who paid. The draft accepts transactions as a common grocery starting point, but a door count is the figure that holds up to supplier scrutiny.

How do retailers track shopper engagement with in-store media? Most start with counted entrance footfall per store and per hour, then add aisle penetration from beacons, transaction inference or a market study to estimate how many of those visitors reached the placement. Sales uplift over the reporting window, which the draft moves from 30 to 28 days, is layered on top rather than used as the audience figure itself.

If you need to replace a transaction-based footfall input with per-store, per-hour door counts before the next supplier negotiation, or want to understand how staff exclusion and entrance weighting affect your OTS figures, talk to Vemco Group about counting at the entrances of your retail media estate.