Ever wondered how Google determines the ranking of websites in its search results?
Well, the debate over the role of clicks in ranking algorithms has been ongoing for quite some time.
Eric Lehman (Ex. Googler from the search quality and ranking team) shares some insights.
Lehman recently testified in the U.S. vs. Google antitrust trial.
He acknowledged, “Pretty much everyone knows we’re using clicks in rankings. That’s the debate: ‘Why are you trying to obscure this issue if everyone knows?'”
This statement alone has generated quite a buzz, especially considering his insider perspective.
While clicks do play a role in Google’s ranking system, they are not a direct ranking factor.
A website with 100 clicks won’t reach the top of search results if another with 101 gets 101 clicks.
What’s even more intriguing…
…Lehman’s take on the evolving landscape of user data and technology at Google.
According to him…
BERT and MUM, Google’s machine learning systems, are becoming more critical than user data.
This shift is significant because, traditionally, user data has been a cornerstone of rankings.
Google uses machine learning rather than relying on extensive user data to test text quality.
In an email from 2018, he wrote, “Huge amounts of user feedback can be largely replaced by unsupervised learning of raw texts.” This indicates a fundamental shift in the way Google ranks content.
The difference between “user data” and “training data” in the context of Google’s BERT technology is a bit confusing.
Department of Justice (DoJ) questioned Lehman about Google’s advantage in using BERT.
Lehman clarified that “training data” differs from user search data, emphasizing that Google’s advantage lies in inventing BERT.
The trial also discussed sensitive topics, including Lehman’s slide that advised employees not to discuss clicks in searches.
Google wants to avoid the perception that SEO tactics can manipulate search results. While clicks matter, Google might be reluctant to reveal their exact role in rankings.
Interestingly, SEO experts have viewed Lehman’s testimony as proof that Google has been less than transparent in its use of clicks.
In order to get a clearer picture, Gary Illyes, a Google representative, confirmed the use of clicks. Clicks can be difficult to interpret accurately because they are a “noisy signal.” Google mostly uses click data to evaluate experiments and personalize search results.
The debate over whether Google uses clicks in rankings continues, but clicks do play a role, albeit. Machine learning has become a key component of Google’s search algorithm, as noted by Eric Lehman.
Understanding clicks and their impact on rankings is crucial for SEO experts and website owners. So, keep an eye on these developments, as they could shape SEO strategies in the future.
