of photograph you want.” Google has gotten russian phone number example so good at parsing images that they are now able to match them to perceived query intent. We already know that Google differentiates between images of single products. For example, here’s a search for “engagement rings san francisco”, and do note the images in the local pack: But when I change my query to “diamond necklaces san francisco”, look at how the photo for the business in the top spot changes.
It’s the same company, but a totally different image chosen to match my query: I have yet to find a live example of Google behaving this way inside the 3-pack for shelfie-type queries, but what we do know from Google’ Cloud Vision API is that they are quite capable of distinguishing between multiple objects in a single image: Given Mike Blumenthal’s report from the Product Experts webinar and Google’s ever-increasing ability to parse images, I would highly recommend that you do a photo shoot this spring of your most popular shelves of inventory, because I predict that Google will presently treat shelfies the same way that they are already handling single-object images.