There’s talk about related information within the SEO community. Maybe it’s all the “What is…” searches that lie behind it. Maybe it’s old signals resurfacing. Or perhaps it’s wishful thinking about RankBrain?
Before we dive into the article, we want to be clear that what follows includes both reported discussions and some of our own speculation. We’re usually careful to separate speculation from fact in SEO (someone has to be), and we’re not making an exception today.
There’s been quite a bit of discussion around information that is related to – but not directly matching – the query Google is currently trying to answer. (One person who’s touched on the subject is Bill Slawski, whose blog is always worth reading, albeit often quite speculative.) We’re going to look at the phenomenon from the angle of a possible solution rather than a problem within SEO. That perspective might help us better understand what’s happening in Mountain View right now.
Imagine RankBrain
Picture a self-learning algorithm whose goal is to present the best possible result for your search – often one that’s never been made before. If you were to apply user data, you’d have to pull it from other, similar searches, since there’s little to no data available for this particular one.
Traditionally, Google has interpreted signals from on-page content, site structure, and backlinks. But in the era of self-learning algorithms, we might imagine using data from related searches instead. Which site has performed best in a similar result? Maybe that one could be promoted in this new query.
We’re not claiming this is true – there’s little concrete evidence that user data is being used in this way, even if there are hints. What Google has disclosed in connection with algorithms like Panda and Penguin – both previously self-learning – is that they use user-like signals rather than actual user data. These algorithms attempt to mimic what a user might prefer.
Your track record could boost your rankings
A model for ranking content on completely new search terms could therefore be based on how users have reacted to your content in related areas. For SEOs, that could mean fairly positive effects.
Generally speaking, well-optimized sites tend to perform well on related keywords. Keyword research has likely been done, and content has probably been created to match users’ search intent. That should mean the site may be prioritized even for keywords it hasn’t directly targeted. This property already exists to a large extent – whether or not RankBrain uses these signals – because you’re building authority (through backlinks) that spreads throughout the site.
So in short, your past performance could help you gain rankings – perhaps even more so than before.
Magnus Bråth
Consultant & Adviser
Magnus is one of the world's most prominent search marketing specialists and primarily works with management and strategy at his agency Brath AB.