"Useful content" and the Navboost system

[arafatenzo]

One of the most important is the Navboost system, which collects data on how users respond to search results. It uses various signals, including clicks, to determine the most relevant results. It remembers clicks for queries up to 13 months old and divides results by location and device type (mobile or desktop). According to court testimony, Google has been using Navboost since about 2005. Roger Montti found that around this time, a patent was filed for Systems and methods for correlating document topicality and popularity 2004 . This patent describes a process in which a user selects documents (which can be understood as clicking on them) from sri lanka email list search results, and based on this, the popularity of the documents is then determined, see the quote from the patent: “(…) a document that users have visited more often than another document may have a higher popularity score.”

Navboost stores each query along with information about user interactions . What did they click on after searching for that query? Did they click on a specific web page? On the search function? Did they hover over a specific element in the SERP? Or did they click on the first result and never return to the search?

It’s not hard to imagine that this type of information could be used to determine what content is useful to people. However, Marie Haynes believes that the way Google uses the information collected by Navboost has changed in 2024. She believes that Google’s systems are learning to predict usefulness without needing a lot of user data to do so.