Retailers “want to be easy to work with”, says Michaela Weber, interim General Manager, EMEA at BigCommerce; but the catch-22 is that reducing friction on the road to purchase can lead to a higher rate of return – while reducing returns friction can lead to losses from returned goods.
“It’s an overall really tricky question – you want to build loyalty with your customers … but you also want to be margin-positive.”
With an estimated 16.9% of retailers’ annual sales in 2024 estimated to have been sent back, totalling $890 billion with of goods, according to a report by the National Retail Federation and Happy Returns, retailers are undeniably facing a dilemma. Do they risk losing out on consumer goodwill by being stricter with returns – or accept the inevitable losses?
Fortunately, advances in technology have only added to the tools at retailers’ disposal for improving their returns prospects, particularly with the advent of generative AI and augmented reality. Weber spoke to Econsultancy about the options that retailers have for tackling returns today, including ways to address ‘expectation mismatch’, personalisation, the state of AR try-on, and the role played by omnichannel.
One new method that has emerged with AI advancement is the ability to analyse customer feedback to pinpoint the reasons for product dissatisfaction or return.
Weber cites “expectation mismatch” as a major contributing factor to product returns – at the extreme end, this can look like people accidentally ordering dollhouse furniture from Temu or Shein, believing it to be a full-sized item; but it can also look like someone ordering clothing and finding the fit to be too small.
“The best thing is having really clear descriptions – and that’s where we see the use of AI tools [that] can help understand your customers’ feedback: you also get return information,” says Weber. “If I’ve got a new jacket, and people are sending it back at high rates because they say it’s too small … you often see a note [on the product page]: ‘If you’re between sizes, order a size up; or, ‘This jacket runs small’.
“There are easy wins you can add as a retailer to help end customers understand the fit of a product.”
Generative AI can also enable more consistent product descriptions across tens of thousands of SKUs that would be difficult for retailers to standardise by hand. “Having something that ‘learns’ best practice – you can train your AI tool to understand the types of information that should be in all the product descriptions,” Weber explains. “Maybe you’re not writing all of them – it’s an impossible task to keep up on, if you’ve got 20,000 SKUs on your site.
“Being able to go through and have it flag where there’s issues, it’s not descriptive enough [is invaluable] – this also relates into your Google Shopping ad listings, ensuring those come up with the right keywords.”
Econsultancy Editor Ben Davis recently wrote in praise of M&S’ return UX, noting that the retailer drills down in its return form to find out what specific elements of a piece of clothing led to a fit issue. Collecting more detailed data at the point of return also gives retailers much more to work with in this regard, which can feed into future design and product description improvements.
Retailers don’t have to take a one-size-fits-all approach to addressing returns. Asos, for instance, has rolled out a returns charge solely for customers with a high return rate – which is applicable unless they keep £40 worth of items. Other targeting can be carried out based on product and price
“You can do some targeted pricing – doing things like final sale, if you have seasonal items; retailers can run promotions where you have tighter return deadlines, to try to ensure that people have a higher intent to purchase – they’re not just impulse purchasing,” Weber suggests.
“If you see you have high return rates on a product, you can just increase the price to account for that; you could add in another five percent. [You] can say, ‘Actually, this product is appealing to a demographic that has a higher return rate – maybe I’ll try to get them [to commit to] a minimum spend to get free shipping, or incentivise them to increase their basket size and retain some goods.”
Some brands have found ways to match customers with the right type or size of product, enabling them to buy with more confidence and making them more likely to retain their purchase. Weber cites the example of BigCommerce client SportsShoes, a retailer for running shoes, clothing and equipment that developed a tool to help customers find the right shoe for their needs.
“They were seeing people struggle to buy tennis shoes online. So, they have a ‘Shoe Finder’ tool [which asks]: ‘Are you going to be walking, are you going to be running? Are you running on a road, are you running on a trail? How much running do you do?’
“You can submit a video of yourself running for analysis” says Weber, with the tool then recommending shoes “after you come through the flow.” “ I think that’s a really elegant way [to] provide some guidance for a purchase that’s a little more considered. They also have a chatbot,” she added
Dressipi, a provider of machine learning-driven personalisation tools for fashion ecommerce that helps to push tailored recommendations to customers that they are more likely to keep, is another strong example of this principle in action – the company was recently acquired by Mapp Marketing Cloud.
Weber praises this type of personalisation, which helps “to ensure that, in your search bar, you’re matching results to what the likely persona is of the person behind the screen. And it’s very successful.”
Augmented reality technology can play a similar role in ensuring that customers are buying the right product and that they know what to expect, although Weber acknowledges that usage of AR tools isn’t “very well established” among consumers. Moreover, while AR works well for certain verticals like home furnishings, makeup, eyewear, headgear, and other accessories, the solutions for clothing, where AR could potentially be most transformative, are still lacking.
“No-one’s really cracked the market” for AR clothing try-on, Weber says, “and I think some of the big brands have decided it’s just too tough.”
One significant hurdle is privacy considerations – “People don’t want to upload pictures of their body to the internet.” Another is the heightened potential for consumer disappointment if an item still falls short of expectations: “If you make a strong recommendation to someone on their size, and then it doesn’t work? Then they’re upset.”
The complexity and cost involved is also likely to be a disincentive to brands – even considering the potential value in terms of tackling returns.
However, some retailers are still iterating on this technology in the hopes of a breakthrough – such as Zalando, which has built on a previous trial of virtual fitting room technology by enabling customers to create a 3D avatar based on their body measurements for better fit representation. The virtual fitting room is currently being piloted in 14 European countries with a selection of upper garments from Levi’s.
Zalando’s Director of Product Size and Fit, Pelin Anlu Bedirhanoglu, said, “We have already observed up to a 40% reduction in return rates with our virtual fitting room technology”, which was first trialled in 2022. “While these results are limited due to the testing phase of the technology, they are a promising indicator of its potential to enhance both customer and brand partner satisfaction.”
Retailers have different UX levers that they can pull to nudge consumers towards a certain outcome; for example, they can introduce some friction earlier in the customer journey with a shipping charge, which may lead customers to take more care in selecting items.
Incentivising larger baskets may also up the likelihood of a positive margin from each customer. But Weber cautions that each of these can lead to unintended outcomes.
“This is the catch-22: customers are so smart. You have to be careful what customer behaviour you incentivise,” she says.
“If you add a free shipping threshold, you’re incentivising higher basket sizes – that doesn’t mean you’re forcing them to keep the goods, though.
“…There are a lot of tools around personalisation and targeting that companies can use to help them understand – who are these customers that are highest-margin? Ultimately, what I need to do as a brand is to have more high-margin consumers.
“So, whatever the behaviours are that help me capture more of those customers is healthy. It’s not good if I add customers that are going to cost me money – they just churn and burn through high return rates.”
Omnichannel retailers can make the most of their store footprint to encourage customers to bring returns in-store: in his article, Ben Davis reports that M&S will nudge consumers to make their return in a clothing and homeware outlet as the processing is “fastest”.
Offering store credit for a return is another option that benefits retailers by locking in future customer loyalty. Weber observes that both this and in-store returns introduce elements of friction, as a customer may not want to actually physically return an item.
However, retailers can highlight the benefits of these options to consumers to make them more attractive – as M&S does with its clothing and homeware in-store returns, or with e-gift cards, which it again signposts as a much more rapid option.
Overall, Weber says of brands considering whether and how to make use of their omnichannel presence: “What we preach to brands is certainly: the more channels you’re in to meet consumers where they are and where they want to interact with you, the more successful you can be.
“Buy online, return in store; buy online, pick up in store – these actually are somewhat sophisticated technologies, because you need to tie in your in-store and your ecomm inventory and reconcile everything.
“But it’s worth it, because then you have a more consistent brand experience.”
If you’re a marketing leader or work in L&D and are interested in how you can upskill your team in 2025, get in touch, or attend our next Capability Forum.