2020 has been the year of entity-based search results with Google indicating a clear shift in its ranking strategies.
SEO (Search engine optimization) is slowly shifting its traditional ideology where links and keywords were the top priorities of an organic campaign.
Entities are the new concept that search engines have been investing in. It is, therefore, integral to understand what entities are, and their relationships in order to know how they impact rankings.
The idea is to see your content as Google sees it. And that’s precisely what entity optimization is all about.
What is an entity?
As defined by Google, “an entity is a thing or concept that is singular, unique, well-defined and distinguishable.”
An entity is therefore not just a set of keywords that are restricted by spelling, definition and language. It is an idea or concept that is related to other ideas.
Let us take the example of the following search query to understand what this means:
We note that a search for “Kuala Lumpur” tells us that it is the capital of Malaysia.
Alternatively, a search for “Capital of Malaysia” automatically displays in the knowledge panel as “Kuala Lumpur”.
What’s more, a search by its popular acronym “KL” gives the same results on Google.
Here, “Kuala Lumpur”, “Capital of Malaysia”, and “KL” are different entities sharing a common relationship. Similarly, there could be many more entities that could be linked together.
Check out this graphic:
What you see above are different entities (the rectangles) and the relationships between them (shown with lines).
These Knowledge Panels give us an idea of how different entities, their connections and relationships are used by Search Engines.
Until now, links were Google’s most preferred tools to understand the relationships between the one site and the other.
However, as the algorithms started weighing additional factors like topical relevance, anchor texts, and PageRank Score, the “entity relationship” theory started to take shape.
How do search engines use entities?
Google’s algorithms work by breaking down each entity and assigning a unique ID to each. A search query or an entity is therefore turned into mathematical computations which helps Google to map them irrespective of the language.
In other words, the weightage of each metric is decided based on the search results. The value is determined by evaluating the type of entities and their relevance. Scores (based on metrics and weights) are assigned to the entities and ranks are assigned accordingly.
As per Google’s patent titled “Ranking search results based on entity metrics”, relatedness is one of the biggest factors that search engines use.
Relatedness refers to the strength of entity relationships that is calculated based on how frequently they occur together on the web.
So if Entity number 1 (Kuala Lumpur) and Entity number 2 (Capital of Malaysia) are mentioned frequently together by authoritative documents, a strong relationship is naturally inferred.
So are entities the key factor to rank?
The truth is we don’t really know how much value they carry when it comes to deciding search engine ranks.
However, we do know that traditionally, Content (Keywords) and Links are two very important ranking factors.
Keywords determine the quality and relevance of the content. And entities determine the relationship between the searched keywords.
Similarly, links and references from authoritative sources are considered important to show up on SERP.
So Google’s focus on entities is essentially a lookout for signals that indicate relations between two keywords or links. It might be considered similar to (or a more advanced version) the use of Latent Semantic Indexing (LSI Keywords) for search.
Google patents on entities
Following are some of the most prominent Google patents on entity metrics and their key takeaways.
1. “Ranking Search Results Based On Entity Metrics”
This Google patent on ranking search results was granted in 2015 was one of the first on entities. Following are the key takeaways from it:
- Relatedness is defined as co-occurence of entities. When entities occur frequently together on authority sites, they are considered as closely linked.
- Notability of an entity is determined by factors like links, reviews, relevance and mentions on other sites. Additionally, if the category in which the entity is positioned has low value and competition, the notability is considered high.
- Contribution of an entity to the topic is assessed by the links or external signals that it is providing to other sites. The more relevant the links, the higher their entity contribution.
- Prize metrics are the relevant prizes (like Nobel Prize, Oscar award, etc) that an entity has received. These add to the overall weight and significance of the entity.
Say for example, we type “Best restaurant in Malaysia” on our search bar. This is how Google runs through the process:
- It determines the relatedness of other entities (like different restaurant names) and assigns value to each.
- It determines the notability and assigns weightage to each.
- It checks the contribution by assessing the links that the site has.
- It looks for prizes (like food awards) associated with the search query and assigns value.
- Finally, it calculates a final score for each entity and produces a SERP result.
And this is what the results look like:
2. “Entity references in unstructured data”
This Google patent was granted in the year 2016. Following are the key takeaways:
- Google has an entity database which stores entities and their connections.
- To produce the top results, the search engine runs through this database to check the frequency of appearance of an entity in top 10 results.
- Entities are also ranked for their quality which is determined by the relevance and freshness of the content, the incoming links, the outgoing links, and user intent.
- To determine the relevance of an entity, its relatedness or relationship to the search query is considered. The most related entities are given the highest ranking.
- Google uses methods to assess entities which have the same name but different intent. Example, a search for “apple” could mean “the fruit” or “the multinational technology company”. Google here checks for other related words in the search query. A search for “fresh apple” would therefore produce results for the fruit. This technique is used to define entity relationships when the data is unstructured.
3. “Related entities patent”
Following are some important points to note from this Google patent on related entities:
- Entity databases consist of individual entities and their links with each other.
- Entity relationship priority is determined based on their co-occurrence.
- Entity relationships are also derived from authoritative sites. For example, if Wikipedia mentions “Petronas Twin Towers” in a search for “Kuala Lumpur”, the two entities are considered linked.
How do entities impact SEO?
So now that you know why entities are valued by Google and other search engines, let us understand why they matter to SEO.
Google’s algorithms are now using machine learning and RankBrain to understand search queries as humans do.
It is no longer ranking keywords individually but is treating them as entities that are interlinked to multiple other entities. Signals are therefore being prioritized based on their relevance to the search queries.
Entities have therefore enabled Google to understand the user intent at a deeper level and with greater accuracy. With its ability to analyse not just language but also the tone, it no longer relies only on links.
While links are still important, they are just one of the many signals that a search engine uses to assess an entity.
It has therefore changed the way we write content. The main keyword is not the only focus of the text anymore. It is important to understand various entities and their interrelationship to increase the chances of search engines ranking the content on top pages.
Optimizing content for entities
We’ve so far understood why entity optimization is important to rank on SERP, build brands and create higher domain authority.
Choose your topic and keyword phrase
Don’t just blindly start creating content for the entities on your list. The purpose is to create content that is relevant to your audience. It is, therefore, vital to look for the right topic and keyword phrase before you even start your entity research.
Look for topics/entities that are related to your chosen keyword
Once you know what you want to write for, it’s time to think the way Google thinks.
How do you know which topics and entities Google considers related to your chosen one?
Following are some quick tips that can be used:
- Check Wikipedia and other authority sites to see what Google considers as related entities.
- Use Google image search to check for related keywords.
- Check the subjects covered under “People Also Ask ” to understand the search intent and use them as your possible headings.
- Use Google’s NLP API demo to get a list of entities that your competitors (top rank holders) are using.
Create context through your content
It is important to note that the text that is created should show clear relationships between the chosen keyword phrase and other identified entities. As mentioned above, Google tools like “Related searches”, “Google image search” and “People also ask” help you understand the user intent.
Keep updating the content
Since entity optimization is all about the relatedness of topics, it is important to keep updating your content to keep it relevant in Google’s eyes.
For example, if you’re writing about “beauty creams” and a new ingredient has recently become a popular entity for Google when it comes to this topic, you need to update your content with this information.
The future of entity search
Entities help us identify valuable information that can be added to the content. However, it is important to mention here that they are just guides that tell us what Google considers important. They definitely are not replacements for valuable links, user intent research and quality content.
That said, most SEO experts believe that entities will surpass other ranking factors in the years ahead. There is hardly anything that we can consider more important when talking about search engines and how they work. And we’ve only uncovered the tip of the iceberg.
There’s more to come in the not-so-distant future!