Google's Gary Illyes described in the latest podcast how Google ranks features or universal results in the overall Google search results user interface. In short, each feature (i.e. news, videos, images, featured snippets, etc) bids for a desired position in the result, they request position X and bid for it. Google's overall search engine that decides where it should go based on several factors. I covered this at Search Engine Land but it is super interesting and wanted to cover it here too.
I like how John Mueller summarized Gary's description of it in the podcast. John said "it's almost like all of these different indexes, or kind of content have their own search engine and basically, they're saying, like my result is like super relevant, or kind of relevant. And then, there's like a super search engine on top of all of these search engines that mixes them all together?" Gary responded, "technically, yes."
In short, Google assigns scores to every result in its index. It then can pass the scores to individual vertical indexes. So image result scores go to the image index, video result scores go to the video index, and so on. Then Google lets the features or vertical search engines request a desired position by bidding on the desired position like an auction. Again, super important to understand that Google is not ranking the individual results (i.e. a specific image or specific video and so on) with this, but rather deciding if an image carousel or top stories box or list of videos should be displayed in the main web results. What images or videos show up in that box, is another thing.
These vertical indexes or features can "also say that I don't want second position, or third position, or fourth position, or so on. I only want the first position," Gary explained. Some can request specific locations, Gary said "we have preferred positions for somethings like, for example, the video results." Related searches often requests the bottom of the page, as another example.
Then John asked Gary how does Google decide when to show what index in the search results. And Gary said clicks can help Google decide that. John asked "how do you recognize if we should show images or videos? Or that? Is it just like video search thing?"
Gary said "we learn it." How does Google learn it? Based on what searchers want and click on. Gary said "so, like when you search for something, something that normally doesn't have images or videos, and you tap the images tab on the result page. Then, you are essentially teaching Google that there was this random person who wanted images for this particular query. And if there are enough users doing that, then you are essentially teaching Google that, that query might deserve images, or videos, or whatever."
You can listen to the podcast here:
Super interesting. And I am sure this will spark the Google uses click data and CTR for ranking debate again. :)
Forum discussion at Twitter.
