News & Views

ASOS / Visual Search

by Contagious Team
This story originally appeared on Contagious I/O, our intelligence tool featuring the most creative and effective ideas in marketing from around the world

Major online retailer adopts visual search into its mobile app



Asos has made a visual search function on its iOS mobile app available to all its UK customers.

The tool appears on the Asos app search bar. Tap the camera-shaped icon and take a picture of an item of clothing, and Asos will search its 85,000-strong inventory for similar-looking products. The search function can recognise items from pictures you take of friends (or strangers), or from snaps of magazine pages. It will also accept pictures uploaded from a camera roll, or screenshots from other apps/websites.

At present, the function is only available on the iOS app, but Asos says that it will soon be available on the Android app, too.



Contagious Insight / 

Mobile first / Asos, with its customer base of 20-somethings, has compelling reasons to create a visual search function for its mobile app. The brand claims that 80% of its traffic and 70% of its orders in the UK come from mobile devices, and its users spend 80 minutes within the app each month, on average.

‘We know this is where our customers are and it’s how they interact with us every day, so we are always looking for ways that are mobile native to make their experience even better,’ says Andy Berks, Asos’ digital product director.

Search party / Online retailers are clamouring to make their UX more attractive and intuitive, and visual search is a popular solution. In 2015 we wrote about Shoes.com using AI to suggest products to customers based on the images that they previously clicked. Similarly, in 2017 West Elm created a tool using Clarifai’s visual recognition API that searches its inventory based on the aesthetics of Pinterest boards.

Image search has been around since at least 2008 (reverse search engine TinEye was one of the pioneers) but it has developed quickly in recent years and it’s significant that a retailer of Asos’ size (12.4 million active customers and revenue of $1.8bn) is adopting the technology.

Visual search is popular because it solves a common frustration: searching for specific products using text queries is difficult. A lot of time it’s barely possible, according to Oliver Tan, the chief executive of ecommerce AI startup ViSenze, who told online publication YourStory: ‘Ninety percent of images on the web cannot be found simply because they aren’t tagged or not tagged well enough. Why are people continuing to search the web using meta-keywords?’

Effectiveness / It’s still early days for the technology, but BloomReach, a tech firm that helps retailers with their customer experiences, has published data indicating visual search’s impact on sales. By monitoring its customers (during a three month period in 2014) BloomReach discovered that visual search was associated with more product views and return visits, and an increase in average spend. Specifically, of the 30.3 million visits to ecommerce (department store) sites that it measured, visitors who used visual search viewed 48% more products, were 75% more likely to make a return visit, and placed orders worth 9% more than those who did not use the technology.