Similar Items in Google Image Search
Update: A week after publishing this post, Google officially announced their launch of “Similar Items” on mobile web and the Android search app.
There may be a new benefit to adding product schema markup to your product pages. A few months ago, Google started displaying product rich snippets in the Google Image Search results on mobile devices.
Before that, Google announced the inclusion of Product Listing Ads (PLAs) in Google Image Search, which appears on both mobile and desktop devices.
Now, Google seems to be testing a new, organic feature in the image search results that displays similar items. Not similar photos, but items. These similar items appear in a carousel after clicking on an image result. It doesn’t appear for every image, but it seems that the similar item carousel is more likely to appear if the image also includes product rich snippets (however I spotted a few images that presented similar items that did not display a product rich snippet).
However, there appears to be more at play here than just structured data markup. It looks like Google is making these recommendations based on both product schema markup AND image recognition. In the image below, you can see that even though the search is for women’s skinny jeans (as is the subject of the picture), the similar items being displayed are based on the third model’s shoes.
Since the page, this image is sourced from does not contain any schema markup (and is not even a product page), the only other assumption one can make is that Google is surfacing these similar items based on what it can recognize in the image, similar to the related photos feature.
Sign up for weekly articles & resources.
It is unknown whether this feature is a test or if it will become a permanent feature as Google tries to mimic many of the features found on Pinterest (such as the image filter buttons). Or a likely scenario is that Google is testing this feature in organic image search so that perhaps they can swap it out with PLAs if they see enough user engagement.
Posted by Kara Eccleston