In the past, SEO professionals did Search Engine Optimization by stacking as many keywords as possible in your website's content. Now, Google uses Knowledge Graph, and SOE is much flatter, with over 200 different ranking factors.
One of the most important ranking factors is Google entities. Google entities take ideas like keywords and topic relevance and bundle them up into a package that's much more fine-tuned.
Taking advantage of Google entities requires talented writers and expert SEO, but your content quality and Domain Authority will thank you.
By focusing on your Google entity, you deliver complete information and allow Google to make more suggestions that point traffic to your content.
A Google entity is any noun. We conceive a person, place, thing, or concept as a container that can act upon and take internal modifiers.
In Google's example, Da Vinci is an entity that has modifiers like date of birth and date of death. He also links to Italy, Michelangelo, and Mona Lisa.
When you google COVID-19, you might also be thinking about "face masks," "social distancing," "quarantine," and "vaccines."
Google will link these entities and can weigh them based on your location, search history, and any other information it might guess about you.
Good content thoroughly examines an entity and uses natural keywords with multiple links to related entities.
Google defines a Google entity as a thing or concept that is
That's a lot to unpack, but you can think of a Google entity as similar to an object in programming. It's a conceptual box that encapsulates everything having to do with a single idea.
Thinking of an idea as an entity rather than a keyword is powerful and makes Google results relevant. In addition, entities are closer to how humans think about the world, and they're helpful for machine learning.
Ok. So. Take the concept "running," for example.
When I think about the word "running," I picture a jogger running in the morning, sweat dripping down their body. I also think about "exercise," "track and field," "shoes," and "Nike." So, let's google "running and see how my brain stacks up against Google.
Sure enough, Google fills the image results with runners, some of them zoomed in on the shoes. I also get a list of related concepts that show people, cartoons, and animals running.
Running has internal modifiers, which are also entities. Ideas like "speed," "human," and "time-of-day" can be independently learned and then applied to "running."
The difference between a keyword and an entity is the difference between a string data type and an object data type. It's a shift in thinking that allows engineers to work with search terms as conceptual boxes.
If Google counted the number of times that the string "running" appeared on a page, spam sites would dominate the first page. Discounting spam sites, a page about how cheetahs run might rank highly.
By putting a conceptual box around running and then ranking the linked concepts, Google showed me exactly what I was expecting.
Google weighs running humans higher than animals and shows me related entities like "exercise'' and "trail."
This Google patent discusses some of the thinking behind entities. It's very dense and, strangely, it's not human readable. Way to go, Google!
A plurality of metrics is determined associated with a search result obtained from a knowledge graph, wherein the plurality of metrics is indicative of the relevance of the search result, and the plurality of metrics is determined at least in part from the knowledge graph.
Basically, google makes conceptual boxes called entities, labels the boxes, and then has data points within the boxes. Those data points are individually manipulatable and can be weighted to predict the best result based on various factors.
Think of this page from Google ads API. Google conceptualizes the ad as an object, called an entity, while various data points are attached to that box.
To improve Google's understanding of the entities in the real world, it looks for known and trusted entities. It then analyzes the keywords associated with those entities to look for new entities.
Google gets a lot of its entity data from curated and trusted content like Wikipedia and IMDB. The problem is that this data might not be timely and certainly isn't complete.
Each entity exists only once in the world. This data might not be timely and certainly isn't complete. So how can Google correctly build the knowledge graph to point out the actual relationships between entities?
The name "Leonardo" is used by Leonardo Dicaprio, a ninja turtle, and a renaissance painter. The "Leonardo" name entity connects to multiple people, characters, and concepts, but the entity "Leonardo" still exists precisely once.
Leonardo Dicaprio and the ninja turtle are both named after the renaissance painter. Leonardo Dicaprio is also linked to his movies and co-stars, while the ninja turtle is linked to fan art and toys.
Google often uses "entity" as a synonym for "object," but it also uses them for SERP suggestions and knowledge panels. It also gives Google a better understanding of the meaning of content rather than needing to rely on keyword string counts.
SEO Power Suite gives some great advice.
Google uses entities to prioritize content that is complete and thoroughly answers a search question. Content that stays on topic while thoroughly exploring it will rank better than keyword stuffed content.
Pages are optimized if they return the results that humans expected when they typed in the search terms. That's a good thing because it prevents pogo-sticking and increases our conversions to page views ratio.
Classic SEO techniques still work, but Google is returning more and more relevant results all the time. We're getting closer to the time when "good content is SEO content."
Rankbrain uses entities as manipulatable objects. "Christmas" becomes an object with time, location, gifts, and "merry" attached. Christmas results might be heavily time-weighted, and so retrieve different results on November 11th vs. March 15th.
Google can make intelligent guesses about whether I'm talking about this Christmas or Christmas as a general concept based on the date and location of my search. It can give weight to the concepts within the entity and decide appropriate search results.
In my Christmas example, November 11th might return Black Friday shopping deals, whereas March 15th might return the Wikipedia page and generic information.
At this point, it’s getting harder and harder to fake your way into first position. Build your content to stay on topic but be topic complete.
As Google classifies more of the world as entities, it builds a picture that is closer to the one that humans see. Entities are exciting for SEO, but ultimately will be very useful to Google in self-driving cars and machine learning.
Entities are a deceptively powerful way of programming for the world around us. Entities are essential for Google, and they're here to stay.
You should be.
If you're not number one in the Google search rankings, you're leaving a lot of traffic and money on the table.
If you're not number one, you need to find out why and fix it fast.
Google entity optimization is one of the top ten tools SEOaudit uses to maximize organic traffic. We use every tool possible and leave nothing on the table.
See what's possible. Discover your website's true potential.