Danny Sullivan from Google posted another big how does search work piece on the Google blog. This one was named how Google autocomplete predictions are generated. And yes, it explains how those autocomplete search suggestions and predictions work.
Of course, he used a Star Trek example – that is Danny.
To come up with these suggestions, Google said it looks at the “common and trending queries that match what someone starts to enter into the search box.” But not just that, it also looks at the “language of the searcher or where they are searching from” so that the predictions can be more tailored to the specific searcher. Google also looks at “freshness” so in cases of some news spiking, Google may shift to show those predictions.
Google said it has “systems designed to prevent potentially unhelpful and policy-violating predictions from appearing” in the search bar. It also has policies and enforcement teams that remove predictions in accordance with Google’s policies. It likely won’t show violent, sexually-explicit, hateful, disparaging or dangerous autocomplete predictions either.
Recently, Google made policy changes with autocomplete and political responses. In 2018, it made other changes to autocomplete around a lot of spam and manipulation. In 2017, it began a method to report autocomplete issues.
Of course, Google said just because Google is not showing you everything you might want to search for in autocomplete suggestions, it won’t stop you for searching for something.
Nothing above is new, it is just a Google how does this work type of post.
Also, Google didn’t really mention this but in the old days, it was pretty easy to manipulate those autocomplete suggestions. Now, not so much.
One last thing, YouTube works differently:
Autocomplete on Search isn’t pulling from your YouTube searches, to my knowledge. YouTube does its own autocomplete system.
— Danny Sullivan (@dannysullivan) October 8, 2020
Forum discussion at Twitter.