Google’s long memory in content indexing challenges
I have encountered the unique challenge of Google’s long memory when it comes to indexing content, says Ted Kubaitis.
Through the use of the Volatility Tool within the Cora SEO software, it’s been revealed that Google can recall and test versions of a webpage that are several years old, with evidence of split testing on a page from four years prior.
This discovery underscores the complexity of SEO work, especially within large organizations where different teams manage various segments of a website.
For instance, a brand update might lead to a change in logos, but Google could continue displaying the old favicon.
The concept of “protocol buffers” at Google is central to understanding this phenomenon.
These buffers act as a comprehensive record for a website, storing the last handful of updates.
If a page isn’t frequently updated, an iteration from four years ago might still be in the last five to ten updates that Google remembers.
This knowledge is crucial when diagnosing why Google might not reflect recent changes.
It could be due to an oversight in updating, Google’s retention of old data, or internal conflicts within the website itself.
To address these issues, SEOs must collaborate with various departments, including operations, to ensure that all parts of the website are aligned and that any redirects or changes are recognized by Google.
SEO is not just about optimizing content but also about understanding and navigating the intricate web of how search engines like Google store and retrieve information.
Information about split testing versions of your pages by Google in the SERP has already been covered in this post.