Things That Suck
Self-checkout: unexpected labour in the bagging area
The customer now scans, bags, weighs, troubleshoots and proves they are not stealing. The wage for this role is a slightly shorter queue.

This looks at self-service checkouts from the practical end — what holds up once conditions stop being ideal.
What holds up in practice
- Work has been transferred to the customer without any transfer of value.
- Weight-based security checks are the main source of interruptions.
- Some retailers have partially reversed self-checkout rollouts.
The labour did not disappear, it moved
Scanning, bagging and payment are the same tasks as before, performed by somebody who has not been trained and is not being paid. The retailer saves staff hours, the customer spends them, and the price on the shelf does not reflect the transfer in either direction.
This is the general shape of self-service across industries, from pumping your own fuel to checking in for your own flight. Some of those transfers were genuinely worth it, because the customer gained speed or control in exchange for the effort. Whether this one qualifies depends almost entirely on how often the machine interrupts you, which brings us to the next problem.
The bagging area is a scale, and the scale is a security guard
Most self-checkouts weigh the bagging area and compare the result against the expected weight of what you scanned. Any mismatch beyond a tolerance halts the transaction, which is why a reusable bag, a handbag placed down or a light item triggers it. The system is therefore a loss-prevention device whose false positives are experienced entirely by honest customers.
Tolerances are set by the retailer, and a tighter setting catches more theft and generates more interventions per hundred transactions. The visible cost of that setting falls on shoppers and the benefit accrues to the shrinkage figure, which is a familiar distribution.
Age checks and the item that needs a human
Restricted items require authorisation, which means every transaction containing them stops and waits for an attendant regardless of the shopper. Loose produce requires a lookup by name or code, which is slower than a barcode and depends on knowing what the item is called. Items without barcodes, promotional bundles and anything requiring a manual price all route back to a member of staff.
A single attendant covering several machines becomes the bottleneck, which recreates a queue at the exact point self-service was meant to remove one. The efficiency gain is therefore real for simple baskets and frequently negative for complicated ones.
The theft argument runs in both directions
Retailers report higher losses at self-checkout, from deliberate theft and from ordinary error by people doing an unfamiliar job badly. Some of what is recorded as theft is genuinely mis-scanning, since the interface is unforgiving and the incentive to check carefully is low.
Somewhere in the release notes, several large retailers have publicly scaled back or redesigned self-checkout in response, which is unusual candour about a failed rollout. Others have added cameras, receipt scanners at the exit and staff at the door, each of which reintroduces the labour that was removed.
The end state in many shops is a machine, a supervisor and a guard, which is not obviously fewer people than a till.
Scan-as-you-shop is a different bargain
Handheld scanning genuinely saves time, because the scanning happens during the shopping rather than after it. It works because the security model is different, relying on random rescans of a sample rather than weighing every item.
A random rescan is a real cost when it lands on you and a much smaller expected cost across all trips than constant interruption. The trade is that the retailer learns considerably more about the order in which you shop, which is worth something to somebody. For a large basket it is usually the fastest option in the shop, which self-checkout is only for a small one.
Practices change, and a company that does this today may have quietly stopped by the time you read it.
Using the machines without losing your afternoon
Take the staffed till for anything over about fifteen items, restricted goods or loose produce, because those are the cases the machine handles worst. Put your own bag on the scale before you start scanning where the machine allows it, since that is the most common false trigger.
Scan heavy items first and place them down immediately, as the weight check is most sensitive with a nearly empty bagging area. Keep the receipt, because a disputed charge from a machine is settled by paper rather than by memory. And remember that a queue you join is somebody else's problem, while a machine you use is entirely your own.
The takeaway
Small basket, barcodes only, and put your bag on the scale first.
The fix is usually trivial, which is the most annoying part.
Questions readers ask
Why does it keep saying unexpected item?
The bagging area is weighed and compared against the expected weight of what you scanned. Bags, handbags and very light items commonly trip it.
Is self-checkout actually faster?
For a small basket with barcodes, usually. For a large or complicated one, the interventions typically make it slower than a staffed till.
Also by Lopamudra Nayak
- Your printer has opinions about which ink you are allowed to buyThings That Suck
- The extended warranty is sold at the till for a reasonThings That Suck
- Same box, same price, less crisps: a short history of the shrinking packetThings That Suck
- The resort fee for a resort you did not want to be atThings That Suck





