Guest Feature: How Edeka’s Netto adds transparency to loss prevention
Netto Marken-Discount, part of Edeka Group, took an innovation-led step to make shrink measurable – without slowing the checkout experience. The discounter’s deployment of AI-powered computer vision from Trigo has just earned Netto the Reta Award 2026 in the Artificial Intelligence category, with Trigo recognised as Top Supplier Retail 2026. But the award matters less than what the technology revealed – data that is now adding transparency to Netto’s loss prevention effort and delivering a proactive, real-time strategy.
German grocery retail is caught in a familiar bind. Labor shortages – particularly among sales staff and cashiers – have made self-checkout not just convenient but necessary. Trigo’s analysis indicates self-checkout concentrates a significant portion of avoidable loss – spanning friction-driven mis-scans and deliberate scan avoidance. Economic pressures have compounded the problem, driving both increased theft attempts and rising operating costs simultaneously.
Most European grocers know shrinkage is a problem. Fewer have a full view on exactly where it happens, when, or how. Estimates from security, operations, and finance rarely align. “The data you don’t have is the data you can’t act on,” says Thilo Freund at Trigo. “Most loss prevention strategies are built on assumptions. The question is whether those assumptions are correct.”
The visibility gap
European grocers see the opportunity to get ahead of the problem. AI-based technology can detect incorrect or non-scanned products in real time. Employees receive immediate alerts, customers see a discreet pop-up if an item needs to be rescanned.
This allows staff to respond quickly before losses occur. Crucially, AI-based systems replace subjective assessments of in-store inventory discrepancies with a uniform, data-based foundation – creating the opportunity to identify recurring patterns early and deploy targeted prevention measures. What the data revealed challenged conventional assumptions: a significant share of losses originated not at checkout, but on the sales floor itself.
Where checkout monitoring hits a wall
This insight points to a fundamental limitation in how most systems approach the problem. It points to an architectural limitation in most SCO loss prevention solutions. If a customer conceals a product in aisle seven and walks calmly to self-checkout, a system watching only the till has no visibility into that item. You cannot generate an alert for something you never saw.
Trigo developed its product-processing expertise while building an autonomous checkout, where accuracy requirements are unforgiving. Those learnings now underpin the company’s loss prevention approach – adapted to work with a retailer’s existing camera infrastructure, no new hardware or store renovations required.
Three-layered approach
The technology operates across three layers. The first validates the scan, catching operational errors, intentional fake scans, or mis-scans. For example, if a customer scans incorrectly and the barcode doesn’t register, a prompt nudges them to retry. The second verifies the basket by identifying items present at checkout but never scanned – products left in the trolley or sitting on the counter. The third tracks concealment on the sales floor itself, connecting what happened three aisles back to what appears – or doesn’t – at the till.
“This is what separates transaction monitoring from genuine loss prevention,” says Thilo Freund. “If you’re only watching checkout, you’re watching the end of a journey that started elsewhere.”
From alert fatigue to actionable intelligence
Detection alone doesn’t solve the problem. When every alert carries equal weight, associates stop paying attention. Peak hour, queues building, multiple flags firing simultaneously – staff clear them without looking up. Trigo’s approach differentiates by risk. A missed scan on a low-value item triggers a simple customer prompt. Concealment of premium spirits escalates to full associate intervention. The response scales with the actual risk – staff are only interrupted when it matters.
A multi-retailer study validated this model. Analysing 1,000+ verified theft cases across major European grocers, Trigo found concealment happens early – 80 per cent of commonly stolen items are hidden in clothing or bags before they ever reach self-checkout. Checkout-only monitoring is tuned for the visible 20 per cent – it misses where most loss actually occurs.
What comes next
When theft does surface at SCO, the patterns are subtle but repeatable: items left in the basket/bagging area without scanning (31.7 per cent) and fake scans where products are waved over the scanner but don’t register (27.3per cent). Category signals are concentrated too – beverages (22 per cent), fresh produce (19 per cent), and bakery (10 per cent) and incidents peak on weekday afternoons and evenings, especially Thursday afternoons. Data like this now informs how retailers respond including intervention rules, staffing, and placement decisions.
Retail’s labour and margin pressures aren’t easing. But for the first time, retailers have a way to see exactly where and how losses occur – and act before they walk out the door. That’s not a promise. For Netto and other leading retailers, it’s already an operational reality.
