Natural Search

The Importance of Entities in Natural Search

The Importance of Entities in Natural Search

Despite what seems like regular commentary on the gradual weakening of the power of the backlink, recent studies have suggested that backlinks are in fact becoming even more important to Google’s ranking algorithm. Whether backlinks are becoming more or less powerful (a debate for another time), a legitimate question remains: will Google’s ranking algorithm move beyond the hyperlink? And if so, what is it most likely to move towards?

The scope of this question is huge and one tackled by PhD mathematicians in Mountain View on a daily basis. We won’t attempt to answer it in the blog, but we will attempt to shed some light on a recent patent publication titled ‘Ranking search results based on entity metrics’ published by Google that offers insights on changes to Google’s algorithm. Namely, entity search.

Google’s Knowledge Graph, a system that connects information about things and concepts together, represents an increasingly important aspect of Google’s search ranking algorithm. An integral part of the Knowledge Graph is the entity - ‘a thing or concept that is singular, unique, well-defined and distinguishable’ - and yet the entity and both how to become one, and how to utilise your power as one, is still somewhat unknown to many marketers.

What is an entity?

Google defines an entity as ‘a thing or concept that is singular, unique, well-defined and distinguishable.’ Tom Hanks is an entity, as is the Eiffel Tower. Google links all its known entities together in the Knowledge Graph and this fundamental understanding of the web helps to rank search results.

A figure from a recent Google Patent filing gives a rough overview of some of the links Google makes between entities:  

How does this affect rankings?
There are four broad elements to entities’ effects on search rankings:

  1. Relatedness: How ‘related’ entities are to one another. This is determined by co-occurrence, effectively how often entities appear together.
  2. Notability: This is one of the more complicated metrics, but it essentially looks at the popularity of an entity through variables such as links or social mentions.

  3. Contribution: This considers contributing influence factors such as reviews. Positive reviews from particularly ’influential’ sources give greater weighting than poor reviews or positive reviews from less influential sources.

  4. Prize: This gives weighting to ‘prizes’ associated to entities, such as a Golden Globe given to a film. 

There are some widely understood takeaways here. Positive reviews help more than negative reviews, more links are better than fewer links etc. But an important note for marketers is that in trying to stitch together relevance and rank search results Google is, at a basic level, looking to understand entities and how they fit within groups (and is using data from trustworthy sources to better understand these entities and groups). 

Remember, Google cares about the groups you are in and what other entities say about you.

What are the key takeaways for marketers?
  • Don’t ignore the importance of citations: Links remain the single most important ranking factor and fundamentally connect entities together, but brand citations can play a part in Google’s wider ranking algorithm. If you secure a feature but don’t secure a link, there may still be natural search value in the placement as citations in the right groups alongside the right entities may grow in importance in the coming years. 

  • Get on Wikipedia: We know that Google used Freebase data to understand entities and feed its knowledge graph and that Google is using Wikipedia to better understand entities and relevance. Try wherever possible to be referenced alongside the right entities in the right groups on Wikipedia and similarly authoritative informational sites. 

  • Get the most famous people in your industry to talk about you: If you can’t beat them, join them. Identify the most famous, trustworthy people/websites in your industry and try to make connections with them. Be referenced by them and alongside them wherever possible.

comments powered by Disqus