Fashion turns to data analytics to cut number of returned items

Zurich-based Isabelle Ohnemus was always interested in fashion. As well as working as a broker, she ran a personal shopping service, ordering clothes and selling them from home. She placed orders online but found it hard to know which size was right for her clients. “There are no unified sizes. Designers each have their own way of cutting and sizing,” she says. “It is hugely frustrating. Even though I was measuring my clients in every way, I was still having to return things.”

She tackled the problem by creating EyeFitU. This app lets users create a profile of their measurements and matches this against sizing charts from thousands of brands, filtering online shopping results down to the items most likely to fit.

Sizing is a significant problem for fashion retailers. Poor fit is one of the main reasons clothes are returned, particularly when it comes to items bought online. Returns cost money. Around $642.6bn worth of clothing and footwear is returned to stores globally each year, according to a study by IHL, a research company. About 10 per cent of this is due to poor fit.

Nick Robertson, former chief executive of Asos, told Reuters in 2013 that a 1 per cent fall in returns would add £10m to the UK fashion retailer’s profits.

Retailers have been trying to solve the sizing issue in various ways. A few years ago there was a rash of virtual changing room trials, avatars that could try clothes on for you and automated shopping assistants. Asos, for example, introduced a virtual fitting tool called Virtusize on its site in 2013, which it says reduced sizing-related returns by up to 50 per cent.

However, these tools only went so far. “A majority of them feel a bit clunky, like you are playing a video game that isn’t much fun,” says Jo Allison, an analyst at Canvas8, a consumer behaviour consultancy. Menswear brands such as Thomas Pink and Hugo Boss have had some success with virtual fitting rooms she says, but “women have more variables”.

“Chest size, for example, can hugely affect how clothes fit, and this can be harder to represent.”

Women’s sizing is also an emotive issue, says Donna North, co-founder of Dressipi, which provides personalisation services for retailers such as Top Shop and John Lewis. “Only about 15 per cent of women are happy to put all those measurements into the virtual avatars. Most of us past 25 don’t even want to know what our waist size is, let alone enter it online,” she says.

Ms Allison agrees. “People will often put sizes into these avatars that don’t reflect them. There is a lot of shame attached to sizing.”

Ms North and her Dressipi co-founder Sarah McVittie are trying a different tack. Their personal shopping assistant not only asks for measurements, it learns a shopper’s size by asking what brands’ tops, trousers and dresses fit best. It further judges size by monitoring what is sold and returned. “It isn’t just about sizing — there are also many features and details that can lead to shoppers returning items,” says Ms McVittie. Dressipi collects around 30 to 50 pieces of data for every garment in its system, and can recommend items to go with previous purchases.

Ms North says the service has led to a 3 to 5 per cent decrease in returns at the retailers it works with and a 10 to 30 per cent increase in spend among users. There are plans to extend the service to physical stores.

“Personalisation is the big game in retail right now,” says Trish Young, UK and Ireland head of retail consulting at Cognizant, a technology consultancy. UK and US shoppers surveyed by the company said they wanted retailers to understand them better.

Personalised fitting could speed up as clothes are wired up to the internet of things — the linking of previously unconnected objects to the internet — says Andy Hobsbawm, chief marketing officer of Evrythng, a technology company that this year signed a deal with US branding and labelling company Avery Dennison. They are working to give 10bn clothing and footwear items a unique profile on a cloud database.

Retailers are testing digital tagging of items to monitor supply chains and reduce theft, and most clothing already has identifying information on care labels. “Each item of clothing could have a data record in the cloud, including where it was made, its materials, size and cut,” says Mr Hobsbawn. “You would be able to simply ask the jacket if it is a good fit for you.”