Is this the solution to the serial returns crisis?

“Serial returners” have quickly become the scourge of the retail industry, set to cost brands £5.6 billion over the next five years.

Online fashion retailers like Asos, Zalando, Boohoo and Missguided, who are by far the most affected by serial returners, are inadvertently but entirely responsible for creating this new phenomenon.

Free returns have become such a fundamental part of consumers’ expectations that 72.9 per cent of shoppers now say they’d refuse to order from a retailer which didn’t offer the service. Despite their best intentions, these companies are effectively stuck with a service that is costing them billions.

In April Asos announced a new hardline stance on serial returners, banning any accounts which show signs that shoppers are buying an outfit, wearing it for one night and then returning for free, effectively renting at the expense of the retailer.

But this initiative only targets a minority of serial returners. The real problem for retailers is how to reduce the swathes of people returning items that simply don’t fit.

“It’s not about what your physical measurements are, it’s about what you like, and how it fits.”

“It’s almost just a big matching problem”, True Fit’s head of data science Rhonda Textor told Charged.

True Fit, one of the few tech companies offering a viable solution to the multi-billion-dollar issue, thinks the answer lies in big data.

That’s why it has enlisted the help of people like Textor with a PhD in computer science, more accustomed to analysing satellite images and building artificial intelligence systems for Microsoft than ensuring shoppers buy the correct sized pair of jeans.

“All those techniques are the same techniques that we use on fashion data,” she explained.

“All the data is different but once you map it to numbers, then the techniques that you apply to those numbers are all the same. My background is actually pretty relevant to what I do at True Fit, applying machine learning and AI to data.”

True Fit’s system works by asking users to think of an item they own, from any brand, which fits them exactly how they like.

It will then consult to what it calls the “Fashion Genome”, the world’s most comprehensive collection of garment specifications and detailed style attributes collated from hundreds of millions shoppers, millions of True Fit customers and detailed information from the brands themselves.

“The fashion genome is very much inspired by say the music genome, this idea that it’s not only about the amount of data but how we also enrich that data with information,” Textor continued.

“So, much like the music genome has information about genres and it’s essentially tagged music, we’ve tagged clothing and footwear.

“A big part of this is also the consumer preference data. Since people have been using True Fit for a decade, we have pretty long history of what a lot of shoppers have been doing over a fair amount of time that we can then connect up.”

With just a few bits of information about their favourite fitting tops, trousers and shoes, shoppers will be given a rating on items they’re browsing via the True Confidence platform, which runs on its retail partners websites.

READ MORE: “Wardrobing” costing UK retailers £1.5bn a year

Brands including Clarks, Kate Spade, Macy’s, Levi’s, Ralph Lauren and Carhartt are already using True Fit on their websites.

The more you shop the more accurate your True Fit profile becomes, making it less and less likely you’ll return items that don’t fit. What’s more, you can use other people’s True Fit profiles to buy clothes for them, potentially doing away with the returns peak which plagues retailers after every holiday season.

“Because we have over 100 million users who’ve told us something about the brands they prefer, and because we have hundreds of retail partners who’ve supplied us with transactional data and we work with brands that have a lot of data on individual items, we can kind of do that very complicated mapping,” Textor said.

“It just boils down to this fits in this way, this fits in this way. If you’re this size in this particular brand, then you would be this size in this particular brand because we know that this fits a little bit larger or a little smaller.

“It’s not about what your physical measurements are, it’s about what you like, and how it fits.”

I think that in general, the problem everyone’s trying to solve is that there’s so much stuff, there’s so much content online, and currently the way that we as shoppers find those things is pretty clunky.

Its True Confidence is by now fairly well established, but with big data comes big opportunities, and Textor says the company is now looking well beyond this platform.

“We’re seeing a lot of really cool things in computer vision a lot of companies are using that to power things like visual search,” she explained.

“So we’re now thinking about how does that relate to the data that we have at True Fit and some of the shopping experiences were trying to create? I think that in general, the problem everyone’s trying to solve is that there’s so much stuff, there’s so much content online, and currently the way that we as shoppers find those things is pretty clunky.

“I have to picture something in my mind of what I’m looking for and think about what are all the search terms that are likely to correspond with that, so that’s not very intuitive.”

Instead of simply providing customers with items with similar colours, patterns and shapes as items you’ve photographed, True Fit would enable suggestions based on fit and style based on hundreds of millions of datapoints.

“We conducted a study with shoppers where when an item is out of stock, we ask them do you think this one or this one would be a better substitute?,” Textor explained.

“So maybe you see a dress and one looks similar, but it’s shorter, and the other looks similar but has different sleeves, it forces people to make a choice about which is more similar, even though they’re not exactly the same.

“It turns out in dresses, at least in our study, length was the single most important thing, which was surprising because we thought visually colour and pattern would be more important. In fact, computer vision algorithms, that’s really what they pick up on. But there are some important attributes that are more subtle than that.”

READ MORE: Online returns to rise to £5.6bn

With peak returns season in full swing, retailers will be scrambling to find any edge which can relieve the pressure and costs associated with selling clothes online.

Rather than deal with the problem, True Fit has targeted the source directly and in doing so has established tool which can be used to solve problems far beyond size and returns.

As with many industries, big data is set to play an increasingly vital role in retail over the coming decade, but it’s innovative companies like True Fit and sharp minds like Textor who will shape how the industry adapts and evolves in the 2020s.

Click here to sign up to Charged free daily email newsletter

 

Leave a Reply

Your email address will not be published. Required fields are marked *