Visual search: The natural evolution in how we search for information
A view from Charlie Davison and Rowan Sullivan

Visual search: The natural evolution in how we search for information

Visual search is expected to be one of the dominant themes at this week's Dmexco 2018 in Cologne. Why does it matter?

Imagine you’re on the Tube and the person in front of you is wearing a really nice pair of trainers. To find them, you could search for "black suede trainers with off-white soles" and leaf through hundreds of possible results. Or, in a world of perfectly accurate visual search, you could find and buy the exact pair instantly from a picture.

Three-quarters (74%) of consumers agree that text based keyword searches are inefficient in helping to find the right product online. This opportunity gap will be explored at Dmexco this week in a number of sessions dedicated to smarter search, and it emphasises that brands need to prepare themselves for visual search.

As the old adage goes, "a picture is worth a thousand words". Humans process visual data better than any other type of data – 90% of information transmitted to the brain is visual. We process information visually, and so it follows that as soon as technology allows, we will search visually too.

As opposed to traditional image search where search engines retrieve images related to a text-based search, visual search uses an image as the search input in order to find more information around the related image.

How does visual search work?

Visual search works by using artificial intelligence to recognise patterns within the image and by matching these patterns to images it has previously learnt from. It then outputs its best understanding of what is in the image, normally a text description of the main object within the searched image.

What are the key benefits of visual search?

Instant gratification: Visual search enables people to discover information about their surroundings instantly by pointing their phone at anything. It’s "like a visual browser for the world around you," says Aparna Chennapragada, vice president of product for AR, VR, and vision-based products at Google.

Effortless discovery: Visual search allows consumers to effortlessly find the exact products they’re looking for or products similar to them that they might be interested in buying, without having to leaf through hundreds of products following a search.

Frictionless conversion: Visual search cuts out the middleman when consumers are looking to purchase. Rather than searching for a keyword in Google, and finding the product within the website Google recommends, consumers can simply search visually through the Amazon app to add an item to their basket.

Emotional connections: For retailers, visual search brings the opportunity for bigger basket sizes. Amy Vener, retail vertical strategy lead at Pinterest, writes, "Visual discovery allows the consumer to generate a more emotional connection, which translates into less price sensitivity."

Who are the main players in visual search?

There are already big players in visual search who are all making moves to get ahead. Firstly, there’s Google with the Google Lens, which is a super-powered, app version of Glass. Point it at things and it will tell you about them. For example, it provides restaurant reviews and can detect everyday objects, text, and articles of clothing.

Pinterest are also in the game with the Pinterest Lens, which offers camera search that recommends ideas based on things people see offline. Pinterest’s Shop the Look feature enables consumers to click into objects they see in an image to shop the items and find similar products. There are now over 600 million visual searches every month across Pinterest Lens.

Then there are the retailers. Asos is one brand that has already implemented a visual search function in their Asos style match. The style match allows shoppers to find clothes they see outside the shop in the Asos catalogue. Just like the Pinterest Lens, the Asos style match lets you take a picture of something your favourite celebrity is wearing and find similar products in the Asos catalogue.

Given the prevalence of visual search nearer the bottom of the path to purchase, a presence on Amazon will also be critical in nailing visual search because of its ability to direct potential consumers directly through to purchase. Therefore brands will need to also cater to Amazon’s visual search platform.

How should brands approach visual search?

Just as they did for traditional search engines and more recently with voice search, any brand that is considering aligning themselves to visual search has to consider the measures that must be taken to ensure that their content appears to consumers across visual search.

Images have always helped brands to sell products but now with image search. If a brand isn’t comprehensively representing products through imagery, it risks not getting found in the discovery and inspiration phase and failing to efficiently convert users. It’s now even more critical to give visual search technology the best chance of surfacing image content appropriately against visual searches. For example, by making your product packaging stand out and having a wide range of product images to help search engines pick out your brand.

Brands should also be optimising their images to improve search engine visibility by ensuing all images have descriptive metadata in the form of filenames and alt tags that are relevant for visual search. In a world where consumers will demand information wherever and whenever they point their camera at a brick and mortar location, local search strategy, such as address details, opening hours, contact details and reviews, also has to be up to scratch.

As with voice search, visual search is another natural evolution in how we search for information, to which brands must now react. With voice search this means helping to build a conversational interface with the answers to respond to all customer queries. For visual search this means going back to the fundamentals of SEO and providing search engines with as many product images and as much information as possible to make their job easier.

Will visual search eventually push out text search?

Not just yet. In the same way that it feels strange asking your phone a question in public, it also feels awkward taking pictures of a stranger to find out what they are wearing. But when an image is easily accessible, the ability of visual search to pinpoint an object from an image will be unbelievably attractive for consumers, and something which brands should start catering for in their search strategies.

Charlie Davison is EMEA SEO director and Rowan Sullivan is EMEA growth manager at IPG Mediabrands agency Reprise