Self-Checkout Theft Statistics The Verified 2026 Loss Data, With the Viral Numbers That Do Not Hold Up Flagged.
Self-checkout loss is real, but most of the numbers shared online are not. The credible research comes from the ECR Retail Loss Group, which analyzed billions of transactions across dozens of retailers to measure how much self-checkout actually adds to shrink. This page collects those verified figures, explains what a basis point of shrink really means, and calls out the viral statistics that trace back to marketing surveys rather than audited retail data.
How Much Do Stores Lose to Self-Checkout Theft?
The most credible measurement comes from the ECR Retail Loss Group, not viral posts. Its research found that stores where 55 to 60 percent of transactions run through fixed self-checkout can expect shrinkage roughly 31 percent higher than stores using no self-checkout, and that every 1 percent of fixed self-checkout use adds about 1 basis point, one hundredth of a percent, to a store's shrink rate. In one video-audited study of about 72 billion euros in transactions, non-scanning at self-checkout was 0.44 percent of self-checkout sales, which added up to roughly 9.5 percent of all store-recorded shrinkage.
Those are increases on top of an existing loss rate, not a claim that a third of sales walk out the door. A store with base shrink near 1.5 percent of sales that runs most transactions through self-checkout is looking at shrink closer to 2 percent, which is a serious margin hit at retail volumes but a long way from the headline figures. The numbers that get shared most, like a "$100 billion" global loss or "36 million Americans have stolen from self-checkout," come from vendor and consumer surveys, not audited data, and are flagged further down this page.
ECR Retail Loss Group self-checkout research. Figures are increases relative to a store's baseline shrink.
The Verified Self-Checkout Loss Figures
| Finding | Figure | Source and basis |
|---|---|---|
| Shrink lift where fixed SCO is 55 to 60% of transactions | ~31% higher | ECR Retail Loss Group, cross-store analysis |
| Shrink increase per 1% of fixed SCO utilization | at least ~1 basis point | ECR aggregate estimate |
| Non-scanning at fixed SCO | 0.44% of SCO sales; ~9.5% of all store shrink | ECR study, ~72 billion euros of video-audited transactions |
| Stores running SCO with no weight-check system | 147% higher losses | ECR case study of one retailer |
| Scan-and-go error rate, full re-scan audit | 43.4% | ECR scan-and-go research, 20,000 audits |
| Shrink increase per 1% of scan-and-go utilization | ~11 basis points | ECR estimate; higher risk than fixed SCO |
All figures are from ECR Retail Loss Group research on self-checkout and scan-and-go. They measure how much loss rises relative to a store's baseline shrink, not the total loss rate itself. A basis point is one hundredth of one percent of sales.
What These Numbers Actually Mean
The most important thing to understand is that these are relative increases. When ECR reports shrink 31 percent higher at heavy self-checkout stores, that is 31 percent on top of a base rate, not 31 percent of sales. If a chain's baseline shrink is 1.5 percent of sales, pushing most transactions to fixed self-checkout moves it toward 2 percent. On tens or hundreds of millions of dollars in sales, that half a point is a large number, which is why retailers care, but it is nothing like the doomsday percentages that circulate online.
The non-scan finding is the clearest window into how the loss happens. When ECR video-audited a huge volume of self-checkout transactions, items that simply never got scanned came to 0.44 percent of self-checkout sales. That sounds small until you see it made up about 9.5 percent of everything the stores lost to shrink. Some of that is deliberate theft and some is honest mistakes at a machine with no cashier to catch a missed barcode, which is exactly why controls that verify the basket matter more than blaming shoppers.
Scan-and-go, where customers scan with their own phones as they shop, carries more risk than fixed self-checkout in ECR's data: a full re-scan audit found a 43.4 percent discrepancy rate, and the loss estimate is roughly 11 basis points per 1 percent of utilization versus about 1 for fixed self-checkout. The lesson is not that self-service is doomed, it is that the amount of verification between the customer and the exit is what decides the loss rate.
Self-Checkout Loss in the Wider Shrink Picture
Self-checkout is one contributor to total retail shrink, not the whole story. The last full industry figure comes from the National Retail Federation, which put shrink at 1.6 percent of sales, about $112.1 billion, for fiscal 2022, split across external theft including organized retail crime, internal theft, and process or paperwork errors. The NRF discontinued that survey in 2024, so any confident "2024 or 2025 national shrink rate" you see is an estimate rather than an official number, and should be treated with caution.
On the crime side, FBI data for 2024 recorded about 1.27 million shoplifting offenses, up 8.9 percent, even as overall property crime fell. It also reported that only 15.9 percent of property crimes were cleared, meaning more than four in five went unsolved. Self-checkout sits inside that environment: it removes a cashier who used to be a natural checkpoint, so the gap it opens shows up both as deliberate theft and as the friction-free mistakes an unattended lane invites. The upside is that the same gap is very measurable, and the controls that close it are well understood.
Because self-checkout loss surfaces as a difference between what the store recorded selling and what it actually banked, catching it is partly a finance exercise, comparing register totals against the day's deposits so an unexplained gap gets noticed early. Retailers that turn each day's bank statement into a clean spreadsheet to reconcile against point-of-sale totals spot a widening shrink problem sooner than those that only look at an annual audit. Loss prevention and the books are reading the same leak from two ends.
What Actually Reduces Self-Checkout Loss
ECR's research points to verification, not suspicion, as the fix. Weight-check systems that confirm a bagged item matches what was scanned are a baseline control; the case study where losses ran 147 percent higher was a deployment that had none. Staff presence matters too: an attendant covering a bank of machines both deters walk-off and helps honest shoppers who hit a snag, and survey work suggests a visible staff member would deter a large share of would-be thieves. Receipt and basket checks at the exit, used consistently rather than randomly, add another layer.
Camera-based analytics close the loop between the lane and the loss report. AI watching the self-checkout area can flag the specific behaviors that drive non-scan loss, an item passed around the scanner, a produce code entered for a pricier product, a cart pushed straight past the reader, and surface the clip to a supervisor in near real time instead of surfacing it in next quarter's audit. Paired with the register data, that turns a vague shrink number into a reviewable incident. Our shoplifting detection AI and broader retail theft prevention run on the cameras a store already has over its checkout area, so the controls above get an automated set of eyes without new hardware.
Numbers We Will Not Print
Several self-checkout statistics get repeated everywhere without a credible source. We are leaving them out on purpose, and here is why, so you can spot them elsewhere.
"$100 billion lost globally to self-checkout"
This round global figure appears on aggregator and vendor blogs with no underlying audited study behind it. Real measured loss is expressed as a shrink lift over baseline, not a single global dollar total.
"36.3 million Americans have stolen from self-checkout, 20 million plan to again"
These come from a consumer opinion survey, not loss data. Self-reported intent in a poll is not a measurement of theft, and the precise-looking counts imply an accuracy the method does not support.
"Self-checkout shrink is 3.75 percent of inventory" / "up to 4 times a cashier"
These specific multipliers circulate without a traceable primary source. ECR measures the effect as basis points of shrink per percent of utilization, which is far more modest than a flat 4x claim.
"52 percent of self-checkout theft is malicious, 48 percent accidental"
This clean split has no verifiable source. Non-scan loss clearly mixes deliberate theft and honest mistakes, but no credible study assigns it a precise malicious-versus-accidental percentage.
Self-Checkout Theft: Questions
How much do stores lose to self-checkout theft?
The credible measurement is a shrink lift, not a fixed dollar total. ECR Retail Loss Group research found stores where 55 to 60 percent of transactions run through fixed self-checkout can expect shrinkage about 31 percent higher than stores with none, and each 1 percent of self-checkout use adds roughly 1 basis point of shrink. Non-scanning at self-checkout measured 0.44 percent of self-checkout sales, about 9.5 percent of all store shrinkage.
Is theft higher at self-checkout?
Yes, the loss rate is measurably higher, though not by the extreme multiples often quoted. ECR data shows shrink rises with self-checkout use, roughly 1 basis point for every 1 percent of fixed self-checkout, because removing a cashier removes a natural checkpoint. Scan-and-go carries more risk still. The size of the increase depends heavily on what verification, such as weight checks and staffing, the store puts in place.
What percentage of self-checkout transactions involve theft?
In ECR video-audited data, non-scanned items came to 0.44 percent of self-checkout sales, which mixes deliberate theft with honest scanning mistakes. That share made up about 9.5 percent of all the shrink those stores recorded. Claims that a large percentage of shoppers steal at self-checkout usually come from opinion surveys about intent, not from audited transaction data, and should not be treated as measured theft rates.
Do self-checkouts increase shrinkage?
The evidence says yes, in proportion to how much you use them and how little you verify. ECR estimates at least about 1 basis point of extra shrink per 1 percent of fixed self-checkout utilization, and a case study of a deployment with no weight-check system found losses 147 percent higher. The increase is controllable: weight verification, staffing, exit checks, and camera analytics all reduce it.
How do stores reduce self-checkout theft?
The controls that work are about verification, not suspicion. Weight-check systems that confirm the bagged item matches the scan, an attendant covering the machines, and consistent receipt or basket checks at the exit all lower loss. Camera-based AI over the checkout area adds automated eyes, flagging non-scans and produce-code swaps to a supervisor in near real time so the loss becomes a reviewable clip instead of a year-end number.
Are self-checkouts being removed because of theft?
Some retailers have scaled back or removed self-checkout at certain stores, citing loss and customer experience, while many others keep expanding it because of labor savings. The data suggests the decision is really about net economics: self-checkout cuts labor cost but adds shrink, and whether it pays off depends on how well a store controls that added loss. It is a tradeoff to manage, not a settled retreat.
Related Solutions and Guides
Retail Theft Prevention
AI that flags theft behavior on your store cameras.
Shoplifting Detection AI
Catch non-scans and concealment at the lane in real time.
Retail Theft Statistics
Sourced US shrink and shoplifting numbers, with citations.
How to Reduce Grocery Store Shrink
Practical steps to cut loss across a store.
Retail Video Analytics
Traffic, conversion, and loss prevention in one platform.
Employee Theft Statistics
Sourced ACFE and NRF figures, myths flagged.
Watch the Self-Checkout Lane With AI
Surveillant runs on the cameras already over your checkout area and flags non-scans, concealment, and produce-code swaps to a supervisor in near real time. Published pricing: $39 to $42 per camera per month, with a free forever plan for one camera.
Works with the IP cameras you already own. No credit card required to start.