Let’s get visual: Shopping from within images

Columnist Christi Olson outlines exciting developments that bring image search and shopping together to enhance the search experience for users.

A little over a year ago, I published an article on the importance of data feeds as influencers of future search:

Structured data feeds front-load the search engine results pages (SERPs) with user-rich information, creating a new search experience for more personalized, localized and actionable results.

Fast forward to today, and we’re seeing innovation in visual shopping experiences using AI to identify an element within an image and show you similar images, similar products and where you can buy said products. From movies to local business listings and voice search results, structured data — both on websites and in data feeds — is playing a central role in moving search forward.

The search engines and retailers continue to pursue newer, better — more structured — ways to communicate and for consumers to take action. This includes the transformation of image search into a visual search which takes shopping to a new level, as searchers can now use images as query inputs. It’s a win-win — searchers can easily shop related images, and retailers can reach even more customers through the ease of shopping feeds and structured data.

How visual search works for shopping

Have you ever looked at a picture, magazine ad or celebrity photo and wondered how you could find out more about a particular item? With visual image search, you can search, browse, and then discover where to purchase products within a few short clicks. Here’s how:

In the example below, a Bing search has been conducted for “Oscar dresses,” and the searcher has clicked on the “Images” tab. There, the searcher can browse through the images; if the store icon appears in the top left corner of the image, it’s signaling that a retailer sells this particular item. (Note: this functionality is specific to Bing, my employer, at the moment.)

After clicking on an image, the searcher is able to view an assortment of related products or images. The product details are pulled from a variety of sources, ranging from the structured HTML on the page to shopping feeds. From here, users can shop similar items without leaving the image results page.

Searchers can then click on the store listing, which will direct them to retailer sites where they can make purchases.

Searchers can also zoom in on pictures to search particular sections of a photo. For instance, let’s say I fall in love with the belt on Brie Larsen’s Oscar dress. I click on the photo in search results, then click the search icon within the photo. This allows me to adjust the search box to zoom in on the desired area. From there, it refines the results of similar products and images.

The results return a wide array of related images, which I can then further browse or shop.

My examples above are clothing-based because I love to peruse the web for what to wear; however, the ability to home in on an element within an image to find that product or similar products is endless. There are numerous opportunities for retailers to have their products appear and shine within image search results, whether in fashion, babies and children, automotive, gardening, home decor, outdoor activities or sports.

I can also envision how similar images could help drive awareness and traffic for things like travel in finding similar destinations or tours/experiences. The opportunities to get more visibility into your products and brand through related images are endless.

Your ‘Go Do’ list: Make sure your products are showing

As a marketer, my first thought was, “What do I need to do to make sure that the products I’m selling are showing up in related products?”

The answer is relatively simple. Think of the best practices for image search optimization, and then layer that together with structuring your data to give the search engines additional context for what products and features that the image contains.

• Structured data markup (where applicable). Structured data markup is code to help search engines return more relevant results by allowing them to more accurately parse your page content. Use the proper schemas to outline a specific product, not a category of products. Make sure you use the schema.org Product type and include at least the name, image, price and priceCurrency, and availability properties.
• Image naming. Use descriptive names for images, and include the target keywords. If you have multi-word names, use a hyphen to separate the words. For instance, the image name for the blue Oscars dress from above could be: Brie-Larson-Blue-Gucci-Dress-Beaded-Belt.jpg.
• Alt attributes. Be descriptive — just don’t keyword stuff! Use the Alt attribute on your image tag to include a description of the product or products within the image. If you are selling products that have model numbers or serial numbers, include that in the alt tag as well.
• Image size and quality. In order to search within the image itself, the image quality needs to be good enough to see individual components without being overly pixelated. This poses a challenge, however, because of the impact of image size on page load speed. You may need to experiment a bit to find the balance that works best for your site.
• Product feed optimization. Similar to optimizing your website, if you are advertising on Shopping Campaigns through either Bing or Google, you should optimize your data within the product feeds. Keep your product feed data fresh and relevant, and make sure that you optimize the titles, descriptions and images. While there aren’t ads appearing on visual search today, shopping campaign feeds and data give another signal to the search engines related to products and images of products. If you start optimizing your feeds today, you can be a first mover and early adopter if and when ads become available through visual search.

Visual search is more than shopping

Visual search doesn’t just significantly enhance the shopping experience, it’s being used in a wide variety of ways:

• Upload your own images to find similar products or images.
• Locate image sources.
• Locate higher resolution photos.
• Use reverse image search to investigate duplicate/copyright issues.

So what are you waiting for? Jump onto Bing’s Image tab and find out how visual search will enhance your search experience.