It’s a known fact that Google’s algorithms are “a secret”. Were the search giant ever to release a book entitled “Google’s Complete Algorithm for Ranking Websites”, it would sell out in moments.
Alas, access to the complete algorithms is a way out of reach of even the most senior of Google employees.
Or is it?
What Google tells us through its patent applications
Whilst Google would indeed be foolish to release all of its ranking ‘rules’ (cue a world where search results are occupied by only the most wealthy businesses who can afford to invest in the now incredibly expensive SEO consultants who rule the search land…), it does let us know something about its inner workings.
It does this through its patent applications.
This is where Google lays out, in as much details as they can, the workings of its ranking systems.
In his latest post for Search Engine Land, Dave Davies explores what the latest Google patent means. And Dave, to thank you for your awesome overview, we give you this…
“Dave Davies is an SEO professional and CEO os Beanstalk Internet Marketing”
(not that guy from The Kinks ;-).
The entity patent
Google has laid out in the entity patent a series of important points which shape the way it understands ‘entities’. For ease, an entity is a ‘thing’ such as a person, a place, a business name, a band, a sport – basically, any noun.
Dave explores each of these points in much more detail, so it’s well worth reading his post, but to summarise:
- To answer the ‘who is’ question, Google looks for the name most commonly used in the top 10 search results
- Frequency of use and topicality are also important factors in understanding the entity
- Different entity types are assessed in different ways, e.g. ‘freshness’ is essential in ranking answers for queries about ‘weather’ but ‘quality’ is most important when providing a definition
- Google doesn’t do all of this in real time; its knowledge graph incorporates entity information so it can retrieve that information quickly when needed (the alternative would be a lot more processing power required)
- Google likely has trusted sources of information to call on when immediacy is required, e.g. in weather queries
- Context is an important assessment factor in understanding an entity
- Inbound and outbound links are considered quality signals in entity understanding
- Google doesn’t understand entities by name, but by unique identifiers, likely to be numerical code
- Each entity can be related to a ‘node’ and an ‘edge’, the ‘node’ being data and the ‘edge’ being a relationship, e.g. the entity Impression is related to the node ‘SEO’ by the edge ‘provides services in’
- Information is grouped into ‘domains’, where the entity ‘Impression’ is related to ‘SEO’ by the edge ‘provides services in’ and both are contained with the domain ‘digital marketing’
- The context of the question is also important in providing entity based search results; the entity ‘Impression’ may relate to everyone’s favourite digital marketing agency, but could also refer to a ‘view’ within the context of advertising or a perception within the context of ‘making an impression’, so Google uses the question format to understand the best search result to serve
- Nodes and edges define the relationship between an entity type and its properties, thus defining a schema (even when traditional schema markup is missing)
All of this is described in much more detail by Dave Davies, so I won’t bother to reinvent the wheel here, and instead suggest you check out his post in full.
That said, I do want to pick up on one element of this patent…
The importance of context and relationships
For me, one of the most interesting parts of this patent as described by Davies is the focus on context. The context of an entity is directly related to the quality score and thus ranking of that entity for a given search query.
Context is therefore essential in enabling Google to build out its entity knowledge base and to serve answers appropriately.
I described the idea of brand associations in my personal blog many years ago now, and this is along similar lines. Context being as important as it is, the relationship our brands have with other brands and other ideas online are essential.
This means, as marketers, we need to be aware of the associations our brands have more than ever. When our brand is mentioned, it needs to be mentioned in relation to our services, our location, our people – all the things you’d expect as standard, but things we might have given little thought to – until now.
This is where a discipline like PR really comes into its own. By pitching your brand and your people for features which better align you to your core products/services/location etc, you can help to strengthen the associations between your brand and those things, thus aiding contextual understanding.
Furthermore, those of us involved in the content production process will need to be more aware than ever of the importance of context within copy. We can use this to create contextual associations – e.g. ‘Impression, a digital marketing agency’ as opposed to ‘Impression, an agency’.
Using context as a tool for improved on page optimisation
Context is really important in helping Google to understand your business and offering. It’s also very powerful in helping you to improve your own search rankings.
For example, if you wanted your SEO page to rank for ‘SEO’ related queries, you’d want to ensure your first cover all the expected elements of what SEO is, and that you supplement this with additional information that answers questions the user is likely to ask as a follow up. As Dave suggests, Google is looking to provide its users with answers as quickly as possible and, where the ranking page can also benefit that user by answering related questions, even better.
What to do now
Essentially, this patent doesn’t really tell us anything brand new. The idea of ‘entities’ has existed for years and the concept of Google using logic to better understand those entities is reflective of any other ranking-related patent.
What is does do is to highlight the importance of content structure and placement. By investing time and thought into strategic content and PR, you can strengthen the associations of your business and improve your online visibility along the way.
Google Plus as an entity development tool?
Here’s a slightly more out there idea, but see what you think…
We know that Google introduced ‘collections’ to Google Plus last year. And we know that Google Plus information is used to feed the knowledge graph, as least in part, already.
It goes to follow then that Google Plus collections contribute to the contextual understanding of an entity. I’d suggest building out those collections, even if Google Plus isn’t a platform you actively use.