“Ryan Gosling is my favourite actor so I’m going to look like him,” says the Asos chief executive, Nick Beighton. “Here’s a picture of Ryan looking cool so I’d like something to make me look like just like him. There we go, a printed T-shirt, add it to bag and away we go.”
Beighton is not waving a magic wand, he’s demonstrating new technology that promises to change the way we shop. Again. From seeing something you like to having a parcel winging its way to your home is now possible within seconds as new technologies reboot retailers’ websites for the smartphone age.
Last month Asos, which is aimed at twentysomethings, quietly began testing a feature that allows customers to upload photos snapped from the pages of magazines or from social media to search for a similar outfit. Want to look like Lily James, star of Baby Driver? Take a snap of her pictured in a sequinned gown in this week’s Grazia, feed it into the Asos app and up pop 100 similar styles, starting with a £16.50 Miss Selfridge slip dress.
About 10% of its customers have access to the software, which flicks through the 85,000 products in Asos’s digital wardrobe in seconds to find the best match to the look in the picture.
Rich Jones, Asos’s head of product and user experience, says the retailer does not yet know whether visual search will change millennials shopping habits: “We don’t know yet, that is why we are doing the test to understand what works with them.”
A number of other major high street players are experimenting with visual search including John Lewis and the shopping centre operator Hammerson. Both have used FindSimilar, an app developed by the tech company Cortexica that was spun out of Imperial College London in 2009. The app lets shoppers upload a picture to search products on sale at the retailer or more widely across the shopping centre.
“Our algorithms are modelled on what the brain does,” says Iain McCready, the Cortexica chief executive, who said its technology mapped 1,500 key points from an image. The app gets better at identifying a type of dress or top shoppers prefer over time, explains McCready “It can learn what a dog looks like and then gets better at it. Give it 100,000 pictures of dogs and it will get very good at identifying what kind of dog it is.”
Shopping centre operators such as Hammerson are fighting back as physical stores are challenged by the rate at which retail sales are transferring to the internet. “Our own research shows that on an average shopping journey customers use a combination of offline and online shopping channels,” explained David Atkins, Hammerson chief executive. “Initiatives such as FindSimilar are designed to further blur the lines between the two.”
McCready says the FindSimilar trial at Hammerson’s Brent Cross centre was well received and that it was going to be rolled out to all 20 of Hammerson’s UK centres.
Last week John Lewis permanently added the tool to its website after 90% of customers said they found it was useful. And, more importantly, they were buying the items of clothing that it had helped them to find.
Voice search is being brought to the masses by Alexa, the digital assistant hiding in Amazon’s Echo speaker. Its new Echo Look goes even further with a voice-controlled selfie camera that enables Alexa to extend her advice to fashion tips – a development likely to force retailers to offer similar services.
McCready can see a world where both visual and voice search will become commonplace, saying Cortexica already has the technology that would enable shoppers to scan the internet for products contained within a video clip. “It’s a bit like your senses,” says McCready of the competing technologies. “When you use them all together it becomes very powerful.”
What most analysts agree on is that the pace of change in retail is increasing. “What you have got to recognise is that there is a massive transformation taking place,” says Tom Adeyoola, the chief executive of Metail, the Cambridge-based startup that has just raised another £10m to fund expansion.
For many shoppers, buying clothes online is a very hit-and-miss affair. Sizes vary between outlets and getting the right fit means many clothing items are returned – which is bad for the shopper and the shopkeeper. But Metail’s software means shoppers can create an accurate 3D model – called “Me Models” – after inputting their vital statistics, enabling them to “try on” clothes.
Adeyoola predicts clothing will go the same way as music and TV, with shoppers expecting websites to offer a “Netflix-style experience” where they log on to be presented with an edited selection of the things they want to buy: “Retailers need to move away from being retailers and start thinking like a consumer tech business.”
guardian.co.uk © Guardian News & Media Limited 2010