When we talk to retailers in the UK and Europe about shelf cameras, the first question is rarely about accuracy or return on investment. It’s usually some version of “what happens to the people in the pictures?”
It’s a fair question, and it’s where we’d start too. A shelf camera sits a few feet from shoppers choosing their groceries and colleagues filling the aisles. Under GDPR and UK data protection law, an image of an identifiable person is personal data, and a retailer that collects it takes on real obligations. Store teams, works councils and unions will want to know what’s being captured. So will shoppers.
So here’s how we think shelf cameras should handle privacy, and how we handle it at Focal.
A shelf camera is not a security camera
Part of the worry comes from picturing CCTV on the shelf. They’re very different tools. Security cameras are mounted high, angled down the aisle and built to watch people over time. Shelf cameras are mounted on the fixture, pointed at the shelf and built to read products: labels, facings and gaps. The angle, the resolution and the purpose are all different.
The shelf is the subject
Shelf intelligence has one job: understanding what is on the shelf. Which products are there, which are missing, which are in the wrong place and which need filling now. None of that requires knowing who walked past.
Obvious, maybe, but it should shape every design decision we make. If the purpose is the product, the system should be built so that people never become part of the data in the first place. Collecting images of people and then promising to protect them is a much weaker position than never keeping them at all.
How Focal’s cameras handle people
Our cameras take still images of the shelf, not video. When a shopper or colleague appears in a frame, every pixel containing that person is stripped out before the image is analysed or stored. We process around six million shelf images a day, and our models are trained to read products. A person in the frame is something to remove, not something to keep.
The cameras themselves are small, battery powered with around two years of battery life, and mounted out of the way of shoppers and staff. They’re designed to blend into the fixture rather than draw attention to themselves, which is also why they don’t get in the way of a busy aisle the way mobile scanning equipment can.
Focal’s systems are GDPR compliant, and we follow the data protection rules in every market where we operate, across North America, the UK, Europe and Australia. We have more than 300,000 cameras deployed worldwide and have had very few complaints.
Focal started in Stanford’s computer vision lab more than ten years ago, and we have built our own cameras and our own models from the beginning. That matters more for privacy than you might think. Because we control every layer, from the hardware on the shelf to the model reading the image, keeping people out of the data is a design decision we make at each step rather than a setting someone else controls.
Even loss prevention starts with the product
The real test of “products, not people” is shrink, because loss prevention has traditionally meant watching people. Theft Spotter takes a different route. It looks for stock that leaves the shelf without a matching sale and alerts the store with the time, date and item, so the team can follow up. The signal comes from the product, and nobody gets profiled along the way.
Questions to ask any shelf camera provider
If you’re evaluating shelf cameras, here’s what we’d ask:
- Does the system capture video or still images?
- Are people removed from images before analysis and storage, or only blurred afterwards?
- Is facial recognition or individual tracking used anywhere in the system, for any purpose?
- Where is image data processed and stored, and for how long?
- What support do you give for DPIAs, colleague consultation and store signage?
- How many cameras do you have live, and what has the complaint history looked like?
You’ll find out pretty quickly whether privacy was designed in from day one or bolted on afterwards.
Better availability shouldn’t cost anyone their privacy
Retailers shouldn’t have to choose between knowing what’s on their shelves and respecting the people in their stores. A camera that only cares about the product is the one you can roll out with confidence, store after store, and explain to anyone who asks. That’s the standard we hold ourselves to, and it’s the one we’d encourage every retailer to hold their provider to.
Frequently asked questions
Do Focal’s shelf cameras record shoppers or store colleagues?
No. Focal’s shelf cameras take still images of the shelf, not video. When a shopper or colleague appears in a frame, every pixel containing that person is stripped out before the image is analysed or stored, so people never become part of the data.
Are Focal’s shelf cameras GDPR compliant?
Yes. Focal’s systems are GDPR compliant, and we follow the data protection rules in every market where we operate, across North America, the UK, Europe and Australia. Because an image of an identifiable person is personal data under GDPR and UK data protection law, we remove people from shelf images rather than collecting them and promising to protect them.
How is a shelf camera different from a security camera?
Security cameras are mounted high, angled down the aisle and built to watch people over time. Shelf cameras are small, battery powered devices mounted on the fixture, pointed at the shelf and built to read products: labels, facings and gaps. The angle, the resolution and the purpose are all different.
How does Focal help reduce shrink without monitoring people?
Focal’s Theft Spotter looks for stock that leaves the shelf without a matching sale and alerts the store with the time, date and item, so the team can follow up. The signal comes from the product, not the person, so nobody is profiled.
