Visual Searches: A Lens on Pinterest’s Interests
In the digital world one activity tends to rule over any other, at least as a starting point. Long gone are the days of finding the address of a new page from a friend in the know or just guessing the obvious, today search is king. Naturally in a world where almost everyone is carrying a camera it makes perfect sense to migrate your searches from the world of a few well-chosen words to that of a single picture … I hear they are worth quite a few words anyway! Visual search uses images be they photos, screen captures or digitised impressions as the basis for an online search either via direct comparison or detection and dissemination.
This has some great practical implications, don’t know what something is? Take a picture of it and search for its name online. Got someone in police custody who won’t tell you their name? get that mugshot online and find their Facebook page. Trying to sell something and want as much advertising reach as possible? … you get the idea.
Lots of big platforms have taken on the challenge of creating and curating their own visual search engines. Most notable being Google, Amazon and Bing, big hitters in the all-round and product search world. However, one of the visual searches gaining the most popularity, and in the news for its continuing enhancements of its visual search, is the online memo-board Pinterest.
The most modern visual search technology uses AI to understand the content and context of its input images and returns a list of not only related results but results with some synergy to the original. This has a veritable raft of uses in eCommerce, especially for interior design and fashion retailers. Visual search allows the seller to suggest items related by style, theme or even material to the shoppers search giving them a broader power to influence sales across ranges rather than just on a specific line.
Pinterest has gone through several iterations of visual search tools beginning in 2017 with Lens, fairly swiftly followed by the introduction of Shop the Look. This became an automated service capable of recommending beyond the initial purchasers findings, for the first time removing human input from the process of buying associated items on Pinterest. Soon after, personalised results for Lens Your Look arrived automatically creating lists of items that matched your style.
In June Pinterest released news about the new form of its visual search engine. Complete the Look, a new search tool created specifically for the Home Decor and Fashion categories, makes style recommendations for multiple items in a photo; clothing and accessories or paints and soft furnishings. It’s recommendations consider the full suite of surrounding objects and details in a photo such as; season and weather, subjects body and measurements, simple aesthetic elements like colour and cut, and even images locations relative to available light.
Unfortunately computer models for fashion design and suitability can be very subjective, in the end it’s the person who wears the clothes or decorates the room who knows what style they like best and its impossible to negotiate the entire spectrum of taste but early usage and testing by Pinterest has apparently been very promising, and more accurate with its recommendations than previous versions.
Pinterest’s tools compete with the likes of Google Assistant’s Lens, which can identify certain objects, translate and detect text, find similarly styled clothes and much more. Earlier, Amazon introduced StyleSnap for ideas based on user submitted pictures from social media, camera shots or even online magazine articles.
As you can imagine this sort of unbroken ground has sparked quite a lot of interest in the marketing sphere. Below you can find some important information about the efficacy of Pinterest and its advertising: There are over 600 million visual searches on Pinterest every month. Pinterest are on course to make over $1 billion in ad revenue per annum by 2020. Brands can target 5,000+ categories via visual search advertising on Pinterest. Pictorial Pinterest Ads have around an 8.5% conversion rate; 21% of Pinterest users use text searching less when they can instead use visual search.
Pinterest so far hasn’t shared when Complete the Look searches will be made available for normal Pinterest devotees but I’m sure we all agree, looking at the figures, it’s an amazing step that could have some interesting ramifications in the world of digital advertising should it become a more mainstream process.