It’s hard to imagine the internet without being able to search for images, but it really wasn’t that long ago when all search results were just plain text. Now, less than two decades after those humble text SERPs, technology is quickly moving beyond the world of merely looking for images, but rather leaping into the organic cosmos of the visual search with the help of artificial intelligence (AI).
Image search beginnings
It all started back in 2000, when the buzz around Jennifer Lopez’s cut-to-her-navel dress that she wore to the Grammys literally overloaded the then text-only Google search engine. This, more than anything else, prompted Google less than a year later to roll out image search. Not only does it allow people to find the visual content they’re looking for, but also enables content providers to drive traffic to their website—something that is absolutely vital for ecommerce sites, stock photography sites, and everyone else who depends on images to bring people to their sites.
For over a decade, content providers could effectively drive traffic to their sites by managing their image SEO in Google. Until 2013, when Google decided to add “View Image” to its image search. High resolution pictures could be downloaded from the search result page — no further navigation needed. Soon afterwards Define Media Group did a study and found that image traffic collectively decreased a whopping 63% among the 87 sites they studied. Image-based websites in fashion & lifestyle, entertainment and news had even higher losses, often losing over 70% of image-driven traffic. After an anti-trust complaint with the European Union, Google removed the “view image” button, restoring image search to its previous desirable state — driving traffic to websites. This all just proves how important visual search SEO has become in driving traffic and sales.
Image search — it matters more than we think
The number of searches being conducted through Google image search makes it well worth our while to pay attention. Recently Moz’s co-founder Rand Fishkin—using data compiled by the behemoth clickstream data provider Jumpshot — has managed to shed some light on just where search clicks are going.
After analyzing the clickstream data of millions of active U.S. searchers (both mobile and desktop) over the course of a month, Jumpshot found that image search accounts for more than 26% of all searchesdone on the internet in the U.S.
According to the pie chart above, that means that Google image searches were engaged
- 5X more than YouTube;
- 10X more than Amazon or Bing;
- Almost 20X more than Facebook.
And while this data was compiled from Jumpshot’s October 2016 clickstream data, recent upgrades to Googles image search and the increased integration of its fledgling child, visual search, is likely to keep use high. More and more we’re becoming an internet landscape based on the visual, and image search will probably be staying relevant for a long, long time.
AI and the advent of the visual search
These days AI is as ubiquitous in the world of image recognition as seagulls on a beach. All the major search engines are investing thousands of dollars in moving from the (figuratively-speaking) two-dimensional world of image search to the more immersive three-dimensional world of visual search. In image search the user types or speaks words that describe the image they want to find. Visual search, on the other hand, is more like a Shazam for images. Instead of songs, the user takes a photo or uploads an image and POOF, the desired information comes up in the search results as images, videos, links to relevant websites, or suggestions for further exploration. It’s based on image recognition algorithms and wouldn’t be possible without today’s machine learning (i.e. AI).
AI now allows search engines to realistically interpret textures, shapes, colors, and even actual words within images. That’s why photo editing software companies like Skylum are using it for everything from noise reduction and HDR-merging to upscaling low resolution shots.
Want to know the history of a landmark you’re standing in front of? No problem. Simply take a photo of it and run it through an app like Google Lens (now integrated into Google’s Image search for mobile as well). Want to know what breed of dog your neighbor is walking? Take a photo of it and run it through Amazon Rekognition. Looking for style ideas to go with your favorite shirt or tie? Take a photo and load it into Pinterest’s Lens Your Look feature. It’ll come back with multiple outfit ideas that match. It’ll even tell you how much the different clothing items cost and where you can buy them. As Bing puts it at the end of their video announcing their new AI-powered visual search app, “A picture is [now] worth a thousand text-searches.”
The continuing evolution
Google Image Search, as well as the other major search engines, is really just in the beginning of its transformation. While the simplistic text-query-to-image-answer that we’re all familiar with will continue apace, the impetus to move to a multi-dimensional immersive experience that literally amounts to a conversation with the user is where things are headed. Soon, search engines will be able to assess an image of various food ingredients and suggest recipes or understand the interior design theme of your living room and recommend complementary products. You may even be able to take a pic of your bookshelf (if you still have books) and receive a number of reading recommendations.
The takeaway for ecommerce marketers?
Visual search will soon become a major player in both transactional searches and, more excitingly, the world of product discovery. And we need to be ready for it. Image SEO might have been a low priority while the view image button was still active, but today it’s back in the game as a main player.
Start with visual analytics
Marketers should get on board, if they want to get the maximum out of visual search. Any optimization starts with the gathering of analytic data. Track the performance on the web of all the brand photography on your website, in your correspondence and social media. Catalog all of your photos with the help of such visual search tools like Curalate or Tailwind, then match the data with images found on user’s blogs, Facebook, Pinterest, Instagram. Even if users are sharing images of your products without mentioning the brand, you will still find relevant data, the usernames of the people who shared those photos, and what was the impact of the post. There are many optimization strategies you can use with such data.
Plan and create visuals based on data
Start using analytical data to outline the strategy in your content plans. Let’s suppose that you are selling t-shirts. Which prints are outperforming in organic search? Which colors are most trendy? Should you switch to V-necks? Maybe an unbranded image of one of your t-shirts has really captured the attention of potential customers. Based on these questions and answers, you will be able to plan and create visuals that users find most appealing. Note the most popular features, the trendiest colors, the most attractive approach and implement all of this in your visual content strategy.
Constant improvement is what keeps any product alive. Be up to date with the evolving needs of your customers. A way to do that is to search for the same category on websites that are full of visual content and do not support any particular brand: Google, Amazon, Pinterest, Instagram. There you are often prompted to jump from one product to the next in the same category with handy “You will also like this” popups. You will see what features are lacking from your products and what are the trends in the target category of your products. This will highlight the gap in your product line up and give you suggestions to patch up those holes where you might be losing market share and find upselling opportunities.
Turn fans into advocates
Visual analytics can help you catch the audience that is interested in your brand. You may be surprised that some popular influencers already feature your line of products in their posts, but sadly, without direct mentions. Visual content search can help you find these people, offer them paid promotions, and let them become your advocates, in turn upscaling your revenue and increasing the amount of your audience. Not to mention, that you will be able to analyze the main characteristics of that particular audience: maybe people of a particular age group or with particular hobbies are among the potential sources of revenue that you haven’t targeted yet.
Brand protection on the free web
You cannot be sure that your logo or brand imagery is always being used lawfully. Positive instances of use are an opportunity for revenue, while negative can decrease sales and harm your brand’s reputation. There are tools specifically designed to catch instances of unsolicited use of your visuals, for example BrandWatch and Logo Grab. Thanks to them you can easily identify uses of images that weren’t sourced from your website, including catching counterfeit retailers that use otherwise untrackable content. In order not to fall in the same trap, make sure that your images are high quality and don’t infringe anyone else’s copyright: hire skilled designers to create new imagery, shoot your own photographs or buy the rights to stock images.
Detailed Content Optimization
A webpage with images, videos and infographic earns 94% more views than a text-only page. Visual messaging in all your channels, from your main website, to email newsletters to banners all should keep your visuals in mind, creating a strong, recognizable brand and conveying a quick visual message, more effective than paragraphs of descriptive text. However, this doesn’t mean that text is unimportant. Build a strong SEO strategy for the texts that surround your image. Think about a menu in a restaurant: the name and description of the food is essential – even if the food already looks delicious on photos. The content to back your images allows users to better commit it to memory, increasing chances of repurchase, shares, recommendations.
Remove multiple steps in finding products
Some people have money, but don’t have time. Searching for an image and having the ability to quickly buy removes unnecessary multiple steps in finding products and creates a gateway towards purchases. Let’s suppose, you see someone on the subway wearing a pair of great shoes – you can snap a photo and quickly buy the same shoes, without even knowing the name of the brand. In this way, your camera transforms into the keyboard, opening the gates to online products and quick checkouts in just a few taps.
Guest Contributor: Bonnie Skott is a freelance journalist focused on emerging trends and digital insights.