Back in 2006, a company named Like.com enabled consumers to upload a photo to search for products. Like.com was popular as a price comparison site, but visual search did not experience wide adoption in the retail community. Google purchased Like.com in 2010.
Fast forward to 2016 and visual search is gaining acceptance. Google’s “Images” search now includes visual search. Last year, Pinterest launched a visual search feature that allows searching part of a pinned photo for products, like the kitchen lamp in the image below. Amazon’s mobile app has included visual search since 2014.
Here are the five primary reasons why visual search is experiencing greater adoption.
1. Advancements in Artificial Intelligence
Artificial intelligence is the ability of software to learn based on data provided to it. Visual search is an artificial intelligence challenge, as software needs to identify the items in a photo (uploaded by a consumer) and execute a search to find matching items. This is not easy, given the billions of images, objects, and shapes.
Take a lamp, for example. In the past, a computer would struggle to identify a lamp as it comes in various shapes and sizes and it was not possible to include every shape and size in the learning data, to identify a lamp.
But recent advances in computing power and Big Data analysis allow more data to be crunched. This means a computer can now have the billions of shapes and sizes an object comes in, to make it easier to identify an object.
Being artificially intelligent, a computer is also learning continuously. If a new variation of a lamp appears in a photo, software can compare it with the billions of other lamps, to identify it. This almost guarantees that the right products will be returned for end users when they execute a visual search by uploading a photo.
2. Increased Accuracy in Real Time
The current artificially intelligent systems are able to produce accurate results — quickly. This speed was missing in the past. For visual search to work for consumers on retail sites, the results needed to show up almost as fast as the text-based search.
Speed has greatly improved in the last couple of years with advanced computing techniques to not only accurately analyze the uploaded photo but also produce the results in real-time. This has made visual search a feature that can be exposed to end-users without worrying about performance.
3. Availability of Visual Search as a Service
Many common ecommerce features gained wider adoption after vendor(s) made them easily available as a service to the retailers, to integrate with their sites. A good example is the ratings-and-review feature by Bazaarvoice. Once Bazaarvoice started offering this as an easy-to-integrate service, the adoption picked up, with thousands of retailers adding product ratings and reviews to their sites.
A similar adoption is happening with visual search. Slyce and Cortexica, as examples, have made it easier for retailers to integrate visual search into their sites and apps. Once the setup is complete, consumers can upload photos to run their searches. CamFind, a mobile app, offers visual search for free.
4. Growth of Mobile Commerce
The rise of mobile computing has made it natural for the users to take a photo and transfer it anywhere. Retailers can benefit by implementing visual search and enabling the consumers to upload a photo and easily search for a product and then buy it.
5. Mainstream Adoption of Visual Search
Larger retailers — Neiman Marcus, Nordstrom, J.C. Penney, John Lewis, Urban Outfitters, Home Depot — now offer visual search on their mobile apps. This has led consumers to expect smaller companies to support it, too. Additionally, visual search vendors (such as the heretofore mentioned Slyce and Cortexica) now integrate with popular ecommerce platforms, to make it easier for smaller retailers.