In-store screens start buying audiences
Grocery TV is bolting 3,500 of Esri's behavioural and psychographic attributes onto 35,000 in-store screens, so buyers can select stores that over-index for a behaviour instead of stores in a postcode. It's the cheapest targeting upgrade in place-based media — and a beauty network can do a version of it with a spreadsheet.
On 29 July, Grocery TV announced a partnership with geospatial analytics firm Esri, adding more than 3,500 behavioural and psychographic audience attributes to a network of 35,000 screens across 6,700 stores. The pitch is short: stop buying stores by location and start buying them by who shops there. A planner can now ask for the stores that over-index for shoppers planning a major purchase, or for organic preference, and get a screen list back.
No new hardware. No new measurement. The inventory didn’t change — the description of it did. That is worth paying attention to, because describing inventory better is the one upgrade available to a network of any size.
What happened
The mechanic here is worth being precise about, because it’s easy to mis-hear as surveillance and it isn’t.
Esri’s Market Potential data is area-level modelled propensity: for a given trade area, how much more or less likely the resident population is to exhibit a behaviour than the national average. Grocery TV joins that to store locations, so each store inherits an index per attribute. A buyer then filters screens by index. At no point does the system identify a shopper, track a device, or need a consent dialogue — the targeting happens before anyone walks in, in the store-selection step.
That’s the structural insight. Digital targeting resolves at the impression; place-based targeting resolves at the venue. Once you accept that, the entire “DOOH can’t target” complaint turns into a data-joining problem rather than a measurement problem — and joining public area data to a venue list is a task, not a technology.
The commercial effect is on the brief. A CPG buyer with a health-and-wellness audience can now be answered in the language they already use, from an inventory list, without a new metric being invented. That’s usually what unlocks budget: not better delivery, but a spec the buyer’s planning tool can accept.
What it means for beauty
- Every beauty network already has this data and mostly doesn’t use it. A salon list plus public area-level data gives you an index per venue — median income, age profile, category propensity — before a single screen is measured. That’s the raw material for the endemic advertiser map and for the venue counts we model in beauty venue base by country. Selling “42 salons in three cities” is a list; selling “salons whose trade areas over-index for premium skincare” is a media product.
- It’s a targeting model that survives the privacy question. Area-level indices need no consent, no device graph and no face in a camera — the durable advantage we set out in the cookieless advantage of DOOH, and the reason we push operators toward venue-level rather than person-level claims in privacy and consent for in-venue measurement.
- This is what a media kit should contain. Not screen photos — indices, trade-area profiles, and a filterable venue list. That’s the build in building a media kit that sells, and it’s what turns packaging and pricing for advertisers from a rate card into a proposal.
- Retail media keeps moving toward the model beauty needs, from the other direction. Grocery TV is buying its way to venue-level audience logic because it can’t do person-level in an aisle. Beauty place-based starts there natively. The two economies and where they diverge are set out in retail media vs place-based DOOH and beauty DOOH vs in-store retail media.
- Index-based selection is also how you price differentially. If a subset of venues over-indexes for a buyer’s audience, that subset is not worth the network average — the argument behind tiered pricing in how to price your inventory.
The caveat that keeps us honest
Two limits. First, modelled propensity is not observed behaviour: Market Potential describes the population of a trade area, not the people who actually walked into that store on a Tuesday, and the gap between the two is largest exactly where a venue draws an unrepresentative crowd — which is most beauty venues. Second, this is a partnership announcement with no performance data: no lift, no CPM uplift, no case study is published, so the claim that audience-indexed planning sells better is commercially plausible and evidentially empty for now. Nothing beauty-specific is asserted here — the attribute categories named are automotive, financial services, health & wellness, telecom and CPG, not beauty — and any beauty read-across rests on our own modelling in beauty DOOH market sizing, not on this release.
Related: The cookieless advantage of DOOH · The endemic advertiser map · Retail media vs place-based DOOH · Building a media kit that sells · Packaging and pricing for advertisers · How to price your inventory