My previous employer already had a product offering that could do this for a better part of a decade by triangulating with WiFi/BLE and cross referencing with surveillance footage. It was deployed in malls and retail chains.
It generated interesting information, but not interesting enough to be profitable.
We weren't the only ones with this capability either, most major retailers had this level of analytics through surveillance footage that previously existed for loss prevention purposes. Then simply link the data to a rewards number or credit card and you got a stable tracking identity.
I've worked with a major retailer on similar backend systems and can echo the post above - all of them are running these systems and they almost never discuss the specifics (until someone like Walmart sues Everseen and we get a glimpse behind the curtain from the court documents).
If you go to an org's website offering these tools (eg, Everseen mentioned above, RetailNext, etc), they don't directly advertise the full breadth of their capabilities until you have them in a room for a sales pitch. They can combine multiple data streams such that an individual can be traced throughout the store via cameras, wifi, and bluetooth, which gives the retailer an opportunity to sell that information. Did a customer pause in front of the corn chips but then decide not to buy? Print them out a Frito-Lay coupon at checkout and see if you can't get them next time, and Frito-Lay will pay you to do that.
I have no first hand experience outside of North America so I won’t speculate. There is a cost of entry so you need to be moving enough volume in a market already working on razor thin margins. I’d expect that this means it’s only for the regional/national players here.
Sorry, "cost of entry" meaning that the software and the supporting hardware platforms makes it cost prohibitive for a small org that isn't moving a lot of volume from their shelves. Grocer margins are razor thin already.
Theft is only one cause of loss. Stocking/admin/counting mistakes, accidental damage, spoilage, or simply people moving stuff around or misplacing items so they're not where it's expected all fall under loss.
The industry tends to use the even harder-to-understand term "shrink". Not always theft, just any loss of product versus what the books say they should have.
"Theft" is such a value-loaded, moralizing term. It collapses a wide spectrum of socioeconomic realities into a single criminalized label, ignoring the structural inequities that often shape people's choices. When we say "loss prevention", we're deliberately reframing the conversation away from individual blame and toward systems, environments, and institutional responsibility.
Loss prevention isn't about vilifying people - it's about acknowledging that harm occurs within a broader context. It centers the idea that organizations can design safer, more equitable spaces that minimize material loss without resorting to punitive narratives rooted in classism, racism, and centuries-old assumptions about who is "dangerous". Calling something "theft" externalizes accountability onto the most vulnerable actors; calling it "loss" recognizes that institutions have agency, too. And preventing that loss focuses on proactive, compassionate strategies rather than reactive punishment.
It generated interesting information, but not interesting enough to be profitable.
We weren't the only ones with this capability either, most major retailers had this level of analytics through surveillance footage that previously existed for loss prevention purposes. Then simply link the data to a rewards number or credit card and you got a stable tracking identity.