SEO entities are a relatively new term that are not yet very widespread. Search engine optimization is a continuous improvement, as search engines are always improving their answers to suit the needs of the user . This is where entities appear.
What are SEO entities?
Entities in SEO are the name of an element, a place, a person, an animal, an object, an entity, a thing, etc. This name is associated with other terms or words, such as dates, actions, or other entities.
The goal of the entities is to give better answers to the user based on the search intent . To do this, search engines establish relationships between words or concepts , creating entities.
What types of entities are there?
If we review the official Google documentation for natural language processing, it classifies SEO entities into:
- PERSON :
- ORGANIZATION :
- WORK_OF_ART :
- CONSUMER_GOOD :
- OTHER :
- PHONE_NUMBER :
- ADDRESS :
- DATE :
- NUMBER :
- PRICE :
- UNKNOWN :
Evolution of search engines with the semantic web
Since desktop computers it has always been more common to carry out short searches, but over time longer searches began to emerge that needed an exact answer. With the arrival of mobile phones, this fact intensified to the point of giving rise to more diversity of searches .
Given the problems of giving a more exact response to the user and detecting the search intention behind it, SEO entities were created to offer better results to the user.
These entities are related to be able to respond to the user, so if we look for the name of a singer, we may be looking for photos, biography, discography or songs. And if we’re looking for a city, we may need to see a map, rentals, or restaurants.
How Google understands the relationship between words
First, we must go back in time a little to better understand some terms related to content relationship techniques and what each of them is about.
- Co-citation : when a text cites references from other texts, giving the probability that these references are related to each other. The probability of relationship decreases the greater the distance between the references.
- Co-occurrence : it is the relationship of proximity between terms that appear in a text and the different parts that make up this text. If two terms are close, they are probably semantically related.
- LSA (Latent Semantic Analysis) : in Spanish it would be translated as Latent Semantic Indexing. Try to find patterns of relationships between terms in a set of documents.
- Ontology : tries to provide computers with the ability to assess information by themselves through the use of artificial intelligence .
Thanks to co-citation , Google is able to perform an analysis on what each content is talking about based on proximity. That is why the theme of the articles is so important from the articles that link to us.
The next step is co-occurrence , which has a double effect. On the one hand we have the words that surround the links, whose proximity makes Google understand much better what the linked content is about. On the other hand, we have the relationship of words, phrases and paragraphs, in relation to the distance that exists between them. If we also add the LSI keywords , the Google BERT algorithm reinforces the understanding of the content based on a context, we have an important advance.
Lastly, we would move on to the indexing method relative to the related document set (LSA). This displays related information as we will see below in the knowledge graph .
Google has been working with entities since 2012. One of Google’s first patents on entities is from February 2012 (Related Entities) and that same year Google launched the Knowledge Graph. From that moment on, they no longer focused on text strings but on real-world entities and their relationships with others.
Google pulls all that information from knowledge databases like the CIA World Factbook, Freebase, WikiData, Wikipedia, and many more. Analyzing all this information and organizing it, he has created the Knowledge Graph , a tool to store and relate all the entities in a visual way that is accessible to the user.
All this will be better understood with an example: Michael Jordan. If we treat it as a keyword, it remains as a proper name. You can’t get more information out of it, but what if we treat it as an entity?
In that case we can know many things about him:
- Born in New York, United States in 1963
- Alias «Air Jordan»
- Teams played for: Chicago Bulls and Washington Wizards
- He is in the NBA Hall of Fame
What is this for? To relate the information by itself, without the need for links or external information.
This can be applied for many searches, such as:
- Italian restaurants in Grenada
- Where to eat Italian food in Granada
- What is the best place to eat Italian food in Granada
Before you could make your own page for each keyword because Google didn’t see the relationship between them, but now Google recognizes the entities “Granada” and “Italy” accompanied by the words “eating” or “restaurants”. It doesn’t matter what other keywords you put like “best”, “which”, “where”… it already knows that you are looking for restaurants to eat Italian food and that is what it is going to show you. If you search for those 3 keywords you will see that the results of the SERPs are identical .