Smarter returns systems must balance automation and trust
Retail returns are becoming a test of how responsibly companies deploy automation. Blue Yonder’s 2026 Consumer Retail Returns Survey found that 89% of UK shoppers would reduce their spending or leave a retailer if their returns were restricted because they had sent back too many items.
That response is significant for companies redesigning fulfilment networks around data-driven cost control. More than 5,000 consumers were surveyed across the UK, US, Australia, France and Germany, including 1,001 in the UK. In the British sample, 52% had selected one retailer over another because of its returns offer, and 35% had stopped shopping with a business after it tightened its policy during the past year.
The UK also has a lower return frequency than the two continental European markets in the survey. Only 25% of British respondents return something at least monthly, compared with 48% in France and 46% in Germany. The operational challenge is therefore not simply volume reduction; it is making decisions that are economically rational and acceptable to customers.
Automation needs a human-centred policy layer
Return-history monitoring illustrates the trade-off. Some 36% of UK consumers consider AI-enabled tracking a fair way to manage costs, while 22% see it as invasive and say they would shop elsewhere. The consequences of personal restrictions vary by age. If told they were returning too frequently and then limited, 62% of Baby Boomers would leave the retailer, compared with 39% of Gen Z.
A resilient returns model should combine customer value, product economics, fraud indicators and environmental impact rather than rely on one frequency threshold. It should also make the outcome explainable to the customer. Such an approach can support both cost control and long-term loyalty.
Product data is part of the sustainability equation
AI shopping recommendations create another source of avoidable reverse logistics. Although 23% of UK consumers have purchased following an AI recommendation, 41% of those shoppers returned an AI-recommended product in the past year. The main causes were a product differing from its description and an incorrect size or fit.
Only 17% of UK consumers trust AI to recommend the correct size, and 37% do not trust it at all. For industrial and retail technology providers, this points to the importance of structured product data: better attributes, dimensions and fit information can prevent transport, inspection, repackaging and markdowns before they occur.
Consumers are open to alternatives that avoid unnecessary movement. Fifty-eight per cent had been offered the option to keep a product instead of returning it, and 67% of those consumers said the experience increased their likelihood of shopping again. The model may be efficient where recovery costs exceed resale value, but it needs controls to prevent abuse. It also responds to a sustainability concern shared by 56% of respondents, who worry about returned goods ending up in landfill.
For Industry 5.0 operations, the lesson is clear: automation should optimise the complete product lifecycle, not merely accelerate a refusal.
