The most comprehensive study of self-checkout losses ever conducted has arrived, and the numbers are worse than the industry expected. The ECR Retail Loss 2026 report, drawing on data from 39 retailers and 1,700 test stores, found that stores with self-checkout machines now record losses 33% higher than stores without them. In the first year after installation alone, shrinkage increases by an average of 22%.
The scale of adoption makes this significant. In stores where self-checkout is available, an average of 54% of all transactions now go through the machines. Self-checkout is no longer a supplementary option tucked into the corner. It is the primary way customers pay.
The Wharton School of the University of Pennsylvania puts the shrink rate at self-checkout between 3.5% and 4% of sales — compared with roughly 1% at staffed registers. According to the ECR data, for every additional 1% of transactions routed through self-checkout, retailers should expect between 0.030% and 0.048% of additional loss. That figure has tripled since the last major study in 2018, when the number sat at 0.01%.
The losses come from a combination of honest mistakes and deliberate theft. For every 10,000 self-checkout transactions, stores recorded approximately 1,000 help calls to staff, 100 card payment failures, and 90 walkaways — customers who abandon the process entirely. Stores in high-density urban areas with younger, higher-income customer bases recorded the worst numbers. An earlier study estimated that losses through self-scan could run up to 170% higher than at staffed checkouts. Non-scanning at self-checkout machines accounted for just 0.44% of self-checkout sales but amounted to 9.5% of all recorded store shrinkage.
The context makes it worse. Total US retail shrinkage hit $142 billion in the most recent full year, a 25% increase from the year before. The National Retail Federation abandoned its annual shrink report altogether, acknowledging the problem had grown too complex to capture in a single survey.
What the self-checkout story reveals is a broader truth about automation in retail. The technology was deployed to solve a labor cost problem. It succeeded — stores need fewer cashiers. But it created a loss problem that, in many cases, exceeds the savings. The machines did not eliminate the cost. They moved it from payroll to shrinkage, where it is harder to see and harder to control.
Three observations for anyone running a retail operation:
The labor savings are real but incomplete. Self-checkout reduces headcount at registers. It does not reduce the need for floor staff, loss prevention teams, or the customer service representatives handling payment failures and item lookup requests. The net saving is smaller than the gross saving, and in high-shrink environments, it may be negative.
Shrinkage is not just theft. Twenty-seven percent of self-checkout losses come from process failures and errors — items scanned incorrectly, weight mismatches, or customers genuinely struggling with the interface. Fixing the technology experience would address a meaningful share of the problem without any additional anti-theft investment.
The machines changed customer behavior permanently. When a human cashier is present, social norms apply. There is a person watching. Self-checkout removed that social friction — and with it, a layer of loss prevention that cost nothing to maintain. Once customers are conditioned to transact without oversight, reintroducing oversight feels punitive rather than normal.
The retailers now pulling back on self-checkout are not doing so because the technology failed. They are doing it because they finally measured the full cost.