Watch: User Behaviour Data as a Ranking Signal
How search engines collect and process user behaviour data as a ranking signal: click-through rate, position and attractiveness bias, session duration, pogo-sticking, the reasonable surfer model, and Chrome's MetricsService.
Transcript
Search engines do not just look at keywords and links; they also pay close attention to how we interact with search results. By processing user behavior data, search engines can interpret the overall user experience and use it to refine their rankings.
One of the most important signals is click-through rate, or CTR. Search engines assume that if users frequently select a particular result, it is likely relevant. However, this data is heavily influenced by position bias, since people naturally click on top results. To combat this, search engines use sophisticated models to adjust for position and isolate the true attractiveness of a result.
But clicks are only part of the story. Search engines also monitor what happens after the click. They look at navigational paths, session duration, and active engagement. If a user quickly bounces back to the search results after visiting a page—a behavior known as pogo-sticking—it tells the search engine that the page did not satisfy their needs. On the flip side, if users spend high-value active time on a page, that engagement sends a strong positive signal. Interestingly, if you link out to highly engaging pages, search engines may even credit some of that positive engagement back to your own site.
Through browser tools and active monitoring, search engines can track subtle interactions like mouse hovering, scrolling, and tab activity. The lesson is clear. While traditional signals still matter, long-term search visibility increasingly relies on creating a genuinely satisfying user experience that keeps people engaged.