Copy article

Self-Checkouts in 2026: Retailers Struggle to Realise Promised Automation Benefits and Consider Rolling Back

ended 07. May 2026

In 2026, retailers aren't removing self-checkouts because they care about shopper frustration. They are removing them because the beancounters finally totted up the cost and these machines are eroding profits. The UK numbers tell a sharp, long running story.

ONS data for the year to December 2025 recorded 509,566 shoplifting offences in England and Wales. In October 2025, the British Retail Consortium (BRC)  Crime Policy Adviser Lucy Whing stated “Retail theft is a major issue for retailers, costing over £2.2bn a year” for reported cases only. Deloitte UK research in 2024 found “a fair percentage of self-checkout loss, circa a quarter (23%), occurs from shoppers unintentionally misscanning a product.” A 2023 Grabango analysis of nearly 5,000 transactions found self-checkout shrink runs at 3.5% of revenue, sixteen times the 0.21% at staffed registers. 

This is not a new problem.  Booths, a Lancashire chain, stripped self-checkouts from 25 of 27 stores starting November 2023. Shrinkage fell. Customer satisfaction climbed from 70 to 74, according to MD Nigel Murray speaking to The Grocer. 

In the US, in 2024, Dollar General stripped self-checkout from 12,000 stores after CEO Todd Vasos called shrink "the most significant headwind in our business." By mid-2025, Dollar General's CFO Kelly Dilts told analysts that shrink improvement after removing self-checkout had become "the gift that just keeps on giving" because of the cost savings realised.

Customers have been clear about self-checkouts for a while. A Toluna and Harris Interactive survey for The Grocer (November 2025) found 

  • 62% of UK shoppers cited technical problems as the biggest drawback
  •  48% said the speed of help when something goes wrong is too slow
  • almost 40% admitted to failing to scan an item at least occasionally. 
  • 51% of British shoppers had avoided self-checkout entirely after just one bad experience. 

Asda's own CFO Michael Gleeson acknowledged in 2024 that the chain had reached a ceiling with self-checkout rollout and pledged to staff more manned tills for longer. 

The UK's Crime and Policing Act, which received Royal Assent last week, scraps the £200 threshold for "low-level" shoplifting, labelling the symptom while the checkout design that still enables it goes unexamined.

We'd like your views:

  • If a quarter of self-checkout losses come from genuine customer scanning errors, not theft, should retailers, who put in the machines, absorb that cost rather than passing it through higher shelf prices?
  • Booths saw satisfaction rise and shrinkage fall after bringing back human cashiers. Is that replicable at Tesco or Sainsbury's scale, or does it only work for a 27-store premium chain?
  • What does this mean for flexible working and entry level jobs for young people? Would they rather work on the checkout or be “TV-screen security guards”?
  • Should staff be expected to potentially put themselves at risk to solve a customer journey failure decided by head office?
  • The replacement technology uses AI cameras to monitor customers in real time. Where does loss prevention end and blanket surveillance of shoppers begin, and who regulates that privacy boundary in the UK?
  • Retail workers returning to checkout areas aren't returning to the same job. They're monitoring screens, not serving people. What does that mean?
  • The Crime and Policing Act removes the £200 shoplifting threshold. Will that reduce retail crime, or does it treat a symptom while the checkout design choices that enabled the losses go unquestioned?

3 responses from the Newspage community

Copy all

Star Quote
Copy

Colette Mason
AI Ethics Consultant at Clever Clogs AI
The machines were sold as faster and more convenient. For a significant number of shoppers, they are neither. Now retailers are feeling the downsides. No doubt self-checkouts had a glossy pitch deck presenting this as a headcount-cutting cost decision dressed as shopper convenience.

Retailers removed paid humans, shouldered expensive store refurbs, and handed the work to acting-unpaid customers. When losses climbed, they didn't absorb the costs, honest shoppers did, through higher shelf prices.

Now the replacement isn't a return to service. It's AI camera systems that track hand movements and lock machines on suspicion. The job entry-level workers are coming back to is a surveillance role, not the customer-facing checkout job it once was.
Copy

Editor
Mitali Deypurkaystha
AI Governance Strategist & CMO at AISensa
509,000 shoplifting offences later, retailers are rediscovering a lesson automation enthusiasts often miss: not every repetitive job should be automated. Removing repetitive admin work so staff can spend more time with people is smart automation. I’ve worked with an accountancy firm that uses automation to free employees for more face-to-face client relationships, improving customer service rather than eroding it. Replacing checkout staff with machines also seemed efficient, but in practice, it hollowed out a human role and turned customer service workers into screen monitors, machine babysitters and part-time security guards. The irony? Instead of freeing staff to help shoppers more, many retailers now have fewer meaningful human interactions and more frustrated customers waiting for somebody to fix a flashing red checkout. The winners in retail may not be the supermarkets with the most automation, but the ones that understand some repetitive jobs hold human relationships together.
Copy

Rohit Parmar-Mistry
Founder at Pattrn Data
Self-checkout was sold as a neat labour-saving innovation. In practice, it has become a live experiment in what happens when retailers automate the easy-looking bit and ignore the messy human part.

The numbers in this brief are the giveaway. If shrink is materially higher at self-checkout, and a chunk of that loss comes from honest mis-scanning rather than organised theft, the machine has not removed cost. It has moved it into complaints, losses, staff interventions and customer mistrust. That is automation theatre: fewer people on the till, more hidden friction everywhere else.

We see the same pattern in AI and workflow audits. The boardroom metric says the process is faster, but the real-world process has no proper exception handling. Retailers should not ask whether self-checkout is modern. They should ask where humans actually add control, confidence and accountability. If a cashier lowers losses and improves satisfaction, that is not nostalgia. That is better system design.