Back to the blog

Visual Search for e-commerce

Wednesday 12th September

Visual Search has only been around since 2014 when Amazon first introduced Amazon Flow. It has come a long way since then with the likes of Pinterest, Google and Bing trying to fight it out to become the market leader in this game changing channel, and yes, visual search will in time become a channel of its own like mobile has, and in up and coming years soon to be voice search no doubt!

Visual Search does come with its challenges for developers as one product can look a millions times different from a search engine depending on the angle, lighting, size of the picture, or other features that appear within the image. But as humans we are pretty lazy, so instead of searching with keywords that generate irrelevant products or having to search for products using a variety of different terms, why wouldn’t we want to simply upload an image to an app and like magic have a selection of products that are exactly what we want to be seeing? As according to visual search company Slyce’s 2015 report, a massive 74% of consumers say text-based keyword searches are inefficient in helping them find the right product online.

But with the image recognition market expecting to grow to $25.65 billion by 2019 (+216% from 2014) has anyone really cracked visual search yet? With the world’s largest retailers snapping it up as part of their e-commerce strategy, this is forming a major gap in the market which search engines and social platforms are eager to fill.


Who does Visual Search well?

Looking for 4 everyday items; a dress, shoes, a fitness watch and an arcade game, which of the top visual search opportunities are really doing what they say on the tin?


In short, Pinterest Lens is by far generating the most relevant products or inspirations and is able to recognise key features of each image, although in terms of usability you have little control as to how each image is cropped. While Bing Visual Search appears to be the most advanced by picking every aspect of an image, this platform is also great for research on products, with the option for “Visual Skills” being brilliant as it allows users to explore that little bit more into the what the product is and how to use it. For example through a visual search of a fruit, the “Visual Skills” feature will be able to generate recipes that include this fruit. Google Lens has a lot of potential and is on occasion; delivering PLA ads within visual search results, but as the visual search feature is only available for users on Android through the Bixby camera there are a lot of limitations.

At the moment Visual Search is not available from a paid advertising strategy point of view. However, brands can align their organic stance in favour of visual search use of multiple images of products both onsite and within social platforms. As well as this, due to the fact that at the moment visual search is more common among mobile apps, ensuring that brands are fully mobile optimised as well would also be key.

Behaviours of Visual Searchers’

Looking at the brands that have already started using visual search capabilities within their app such as ASOS, H&M, Forever21, Target and Macy’s strongly indicates that the core audience falls within Millennials and Generation Z. This is backed up further by the fact that 62% of millennials want visual search over any other new technology, while 70% of Gen Z discover products on social media showing that capturing users while they are browsing on social media certainly is forming a gap in the market for this new channel of advertising.

Article Written by

Caroline Selby

Biddable Manager, Total Media

Caroline has been with Total Media for 18 months, with 4 years Digital experience prior to this, and has been working closely with many retail brands throughout this time.
She has a love for goats and can usually be found watching videos of fainting goats when she isn’t researching the latest trends in the industry.

Liked this article? Let others know