marql

Retail Shrinkage: What It Costs a Store Chain and How to Cut It, Cause by Cause

Shrinkage is stock you paid for that never became a sale. Theft gets the headlines, but errors in the stock record make up a large share of it, and they are the cheapest part to fix.

Key takeaways

  • Shrinkage is stock a retailer paid for and could not sell: stolen, damaged, spoiled or lost to errors in receiving and recording.
  • Grocery runs above the retail average, and in the UK the published theft figures count only part of the loss.
  • Every pound of shrinkage comes straight off profit, so a small rate weighs heavily on a thin-margin store.
  • Theft is not the whole story: process failures and errors are about a quarter of shrink in the largest survey, and a wrong record also empties shelves.
  • Find shrinkage store by store and category by category, starting with lines whose stock record cannot be true.

Retail shrinkage is the stock a store paid for that never turned into a sale. Some of it walks out of the door, some is damaged or goes out of date, and some never existed except in the stock record. For a grocer it is not a rounding error: independent grocers in the US reported shrink at 3.9% of sales against a gross margin of 27.9%7, a loss equal to about a seventh of everything they earned on the goods they sold. This guide covers what counts as shrinkage, how large it runs, what causes it and what a store chain can do about each cause.

1. What counts as shrinkage

Shrinkage is the gap between the stock a retailer should have and the stock it can sell. Part of it is known when it happens: a breakage written off, a theft caught on camera, a tray of salads past its date. The rest, unknown loss, appears only when a count finds less than the book says. The rate is the loss over a period divided by sales over the same period, valued consistently at cost or at selling price; the stock take guide works the formula through with an example and covers how to count and reconcile.

The causes fall into five groups: theft by customers, theft by staff, errors in the process (deliveries, transfers, pricing and recording), waste and damage, and a remainder nobody can explain. The largest survey that splits them is American, and its retailers attribute the shares like this:

Where shrink comes from, US retailers, financial year 2022

External theft
36%: shoplifting, burglary, robbery
Employee theft
29%
Process, control failures and errors
27%: receiving, pricing, recording
Unknown and other
7%
National Retail Security Survey 2023, 177 retail brands. The shares are the retailers' own attributions.

2. How big it is

The same survey put average shrink at 1.6% of sales, up from 1.4% a year earlier, with pharmacy, grocery, department and mass-merchandise stores above 2% on average1. UK grocers do not publish a comparable rate, but the direction is clear. Ten of the largest UK retailers, with £192bn of turnover between them, told ECR Retail Loss in 2023 that their shrinkage was on average a third higher than in 2018, and more than half higher for some; they put most of the rise down to shop theft2.

UK crime figures show only the part that is recorded. The British Retail Consortium counts 5,451,000 incidents of customer theft costing £408m in the year to August 2025, but only verified losses, and says the true cost is likely to be much higher3. Convenience stores feel it per shop: the Association of Convenience Stores estimates crime cost the sector £354m, or £7,137 per store, an 11p "crime tax" on every transaction, on top of £6,213 per store per year spent on prevention. Its members estimate 5.8 million shop thefts, against 519,381 recorded by the police4.

3. Why a small rate matters

Shrinkage comes off profit, not off sales. Take a chain of six shops selling £3m a year at a 30% gross margin. A shrinkage of 2% of sales, valued at cost, is £60,000 a year: stock that was bought, delivered, shelved and paid for, and never sold.

  1. Gross margin on £3m of sales£900,000

  2. Shrinkage at 2% of sales−£60,000

  3. Margin the chain keeps£840,000

An illustration with round numbers, not a chain's data.

At a net margin of 3%, earning that £60,000 back through trading would take another £2m of sales, two thirds of the chain's turnover. That is why loss prevention pays for itself faster than most growth plans in a small chain, and why it is worth knowing which part of the loss is which before spending on any of it.

4. What to do about each cause

Customer theft. Organised offenders go for alcohol, cosmetics, meat, baby formula and medicines3. Keep those lines in sight of the till, in smaller facings, tagged or in locked cases where the volume justifies it, and count them more often than the rest of the range. Convenience retailers rate updated CCTV highest among their measures4, but a camera only helps if someone reviews it and reports what it shows.

Self-checkout. A study of 39 retailers with €1 trillion of turnover found that store losses rose by 22% on average in the first year after self-checkouts went in. Missed scans happen in 1% to 4.8% of self-checkout transactions and cost about €2.50 each, and retailers' own estimates of how much of the loss is theft range from 6% to 80%: much of it is customers struggling with awkward items and barcodes. The fixes are mostly cheap. An on-screen prompt to rescan led customers to correct themselves in more than 80% of cases, personal display monitors cut losses by 12%, and one retailer cut walkaways by 28% with exit gates5.

Staff theft. In the US survey it is the second-largest cause1. The signals sit in till data: refunds, voids, no-sales and staff discounts by cashier and by shift. Compare them across stores and across people doing the same job, and follow up the outliers rather than the averages.

Process errors. A delivery booked in full but short on the pallet, a transfer booked out of one store and never into the other, a case of twelve booked as one unit, a product sold under the wrong code: each moves the record away from the shelf without anything leaving the building. In 11 stores of a UK grocer, 39.4% of stock records showed more than was on the shelf, and perishables were less accurate than the rest6. These errors are cheaper to fix than theft and they do double damage: the stock take reports them as loss, and in the meantime the system sees stock that is not there and orders nothing. In the store the researchers had recounted in full, sales rose by about 11% over the next two months, almost all of it on lines whose record had been too high6. The out-of-stock guide shows how to spot those lines from sales data, and the guide to investigating a shortage how to trace one discrepancy to its cause.

Waste and damage. In grocery, fresh food that goes out of date is a large part of known loss. Ordering closer to what each store sells and reducing prices before the date do more here than security; the food waste guide covers both.

5. Find it store by store

A chain-wide shrinkage rate tells you how much, not where. Break it down by store and by category, and compare each store with its own history and with similar stores in the chain. A store whose margin slips while its sales hold, or a category that loses more at one store than at the others, is where to look first. Between counts, the stock record itself gives the earliest signals: a negative stock figure, which no shelf can have; stock that has not moved for weeks on a line that used to sell; and a store that keeps receiving a product the till never records selling. Lines that hold stock with no sales are also where the dead stock guide starts.

Here is what such a signal looks like in a real store. Yammy Way is a candy chain in Romania that uses marql. On 17 November 2026, the order marql prepared for its Brașov store had three lines with a negative stock record in the chain's ERP. On 20 November it had one, on a different product, while the orders of the chain's other three stores had none:

17 NovProductJohny Bee 3D Gummy Cup Cake 18 gStock in the ERP−3Sells a day0.7What marql showsRead as an empty shelf; check first, before the order goes
20 NovProductFINI Barril Pirata B/200UStock in the ERP−4Sells a day0.7What marql showsRead as an empty shelf; check first, before the order goes
Yammy Way, Brașov store orders in marql, 17 and 20 Nov 2026: one negative stock record from each day.

A shelf cannot hold minus four. The store sold units the system never saw arrive: a delivery or a transfer that was not booked, or the product received under one code and sold under another. Left alone, such a record either blocks the next order or turns into a "loss" or a "gain" at the next count that has nothing to do with theft. Weighed goods add their own blind spot: Yammy Way sells pick-and-mix by the 100 grams, so its stock sits in the trays and in unopened boxes in the back, and only the trays show up at the till. The Yammy Way and Pralio story shows what counting both in kilograms turned up.

A note on the evidence: the cause shares and the grocery rate are American, the UK theft totals count only recorded or verified incidents, the convenience figures are extrapolated from 43 retailers representing about 11,000 stores, the UK grocers' index is their own report, and the self-checkout figures come from the university's summary of the full report.

How marql shows it

marql does not put a shrinkage figure on a store: that takes a count. What it shows is where the record and the shelf are likely to disagree. In each store order that the AI Inventory Manager prepares from the chain's POS and ERP, a line with a negative stock record, or with a stock reading that is old or missing, is marked to check first and listed before the others, and a negative record is treated as an empty shelf rather than as stock. Analyze shows stock at cost against the snapshot a month earlier and the stock that has had no sale in the last 60 days, and its profit and loss has a line for losses, which fills in once a source sends the chain's write-offs. A person checks the shelf, corrects the record and sends the order.

  1. National Retail Federation and Loss Prevention Research Council (2023-09-26). 2023 National Retail Security Survey, release. lpresearch.org
  2. ECR Retail Loss (2023-07-12). UK Grocery sector reports a 33% increase in Unknown Loss (Shrink). ecrloss.com
  3. British Retail Consortium (2026). BRC Crime Report 2026, covering 1 September 2024 to 31 August 2025, pages 14 and 15. brc.org.uk
  4. Association of Convenience Stores (2026-03). ACS Crime Report 2026. acs.org.uk
  5. University of Leicester (2026-06-17). University of Leicester study reveals rise in retail losses linked to self-checkouts (on Hopkins, M., Self-Checkout Loss Report 2026, ECR Retail Loss). le.ac.uk
  6. Rekik, Y., Oliva, R., Syntetos, A. and Glock, C. H. (2026). Inventory record inaccuracy in grocery retailing: impact of promotions and product perishability, and targeted effect of audits. Journal of Business Logistics 47(3), e70079. Figures as in the authors' manuscript, arXiv:2506.05357. doi.org/10.1111/jbl.70079
  7. The Shelby Report (2026-07-30). Annual FMS-NGA study: independents held firm in fiscal 2025 (2026 U.S. Independent Grocers Financial Study, FMS Solutions and the National Grocers Association). theshelbyreport.com

TopicsRetail shrinkage, Loss prevention, Shop theft, Stock records

All articles

Blog

Frequently asked questions

Shrinkage is stock a retailer paid for and could not sell, because it was stolen by customers or staff, damaged, spoiled, or lost to errors in receiving, pricing and recording. Part of it is known when it happens; the rest shows up only when a stock take finds less than the record.

Subtract the counted stock from the book stock for the period between two counts, value both on one basis (cost or selling price), and divide by net sales for the same period. Compare stores and periods only on the same basis.

There is no published UK benchmark. The US National Retail Security Survey put average retail shrink at 1.6% of sales for 2022, with grocery above 2%, and US independent grocers reported 3.9%. Your own stores compared with their history and with each other are the more useful benchmark.

Known loss is recorded when it happens, such as a write-off for damage or out-of-date food, or a theft that was seen. Unknown loss is the gap a stock take finds between the record and the shelf, with no event to explain it.

In the largest survey that splits it, retailers attributed 36% of shrink to external theft, 29% to employee theft and 27% to process and control failures and errors. In grocery, waste of fresh food adds a large known loss on top.

On average, yes. A 2026 study of 39 retailers found store losses rose 22% in the first year after self-checkouts were installed. Much of the loss is missed scans by honest customers, and prompts to rescan, personal display monitors and exit gates reduce it.

Break the loss down by store and category, correct stock records that cannot be true (such as negative stock), count high-risk and perishable lines more often, review refunds and voids by cashier, protect the lines thieves target, and order fresh food closer to what each store sells.