The implications of Google using click data in ranking
The implications of Google using click data in ranking
So, it’s out: leaked documents in Google’s antitrust lawsuit confirm Google uses clicks in ranking.
For a long time, the narrative was “we’re not using clicks in ranking.” Now, it turns out Google has all along.
We try to avoid confirming that we use user data in the ranking of search results. (source)
But what does that mean? And how does it change the way we work?
Where and how does Google use clicks?
Using clicks in ranking can mean a lot of things: CTR, long clicks, pogo-sticking (clicking on lots of search results until you find what you want), etc.
But just using such signals would be overly simplistic.
Ranking signals don’t exist in isolation. For example, even if CTR is a ranking factor, the question still remains how much weight it has when it’s applied, e.g. The issue is complex.
When the DOJ questioned Eric Lehman, who worked on the Google search quality team for 17 years, in the most recent Google antitrust case, we learned there's a difference between user and training data.
Training data = Google uses click data to train systems like BERT / Rankbrain, etc. and launches changes during (core) algo updates.
User data = Google measures who clicks on what in real time and adjusts ranks for all or just some users (personalization).
We've known for years that Google is using clicks in some capacity.
Gary Illyes confirmed that Google uses clicks to display SERP Features at the SEO meetup in SF.
It's also likely that Google uses clicks to detect user intent changes (on top of query rephrasing).
Google also mentions using click data for experiments.
Even though the leak is just a few documents that mostly try to hide the fact Google is using clicks in search in fear of SEOs, the leak is significant. It also shouldn't come as a surprise since all major content platforms use engagement signals. TikTok uses swipes. YouTube uses watch time. Why wouldn't Google use clicks?
How does click training data change the way we operate?
Google using clicks to train ranking systems doesn't mean we should change the way we think but rather reinforce it.
Let's say Google is using CTR, pogo-sticking and long clicks the way we've suspected for years. How would that change the way we work?
We would:
Pay more attention to design and user experience since it impacts whether users come back.
Keep Optimizing snippets to draw more clicks (even though Google rewrites titles).
Invest in brand awareness so users recognize domains and click our domains in search.
Keep people on our sites with meaningful engagement.
Provide outstanding user experiences.
We can see when Google tests our results against others - maybe based on click-trained systems. When analyzing rank positions for keywords, we can see that Google flips the results for keywords for a few days and then makes a call on whether to change the order or not - presumably based on click data.
3 examples:
I assume it goes beyond CTR: when a result gets more clicks than one ranking above it, it might trigger Google to test the two against each other and decide which one should rank at the top based on meaningful engagement (e.g., users don't continue their search journey).
What this is really about
While the fact that Google uses click data in ranking is interesting, it doesn't lead to quick tactics. Instead, it highlights the need for broader collaboration between SEO, design, product, and content to optimize for brand and user experience.
Working closely with other orgs has been critical for a long time, but we still have more articles about about ranking factors than effective cross-collaboration. Consider this a promise to write more about that topic.
Until then, here are some articles to warm you up: