- What is Entity and how is it related to SEO?
- Why is Entity important?
- How can optimization according to Entity?
In this article, I will answer the above 3 questions in detail, and guide you from theory to provide detailed solutions so you can practice Building Entity right on your website!
Now then… Let’s get started!
What is Entity SEO?
Entity is an abstract concept. It can be anything: a place, a person, an organization, a concept, or an idea.
Take for example the following text, you will see that Google recognizes many different types of Entity such as people, organizations like Google, timelines like 2007, countries like the USA…
This is how Google gradually perfected the naturalness or ability to understand language and text content.
If I had to define what Entity is, I would say these are semantically related entities, thereby helping the machine understand both the literal and the figurative.
Entity is not only an obvious object, it can also be a concept, idea, or trend.
So why is Google interested in Entity extraction or Entity as a whole?
Surely many of you have heard about link graphs, related to link building as well as spam.
Actually, Google is always trying to prevent this situation.
There used to be a time when a lot of people thought that social signals (the number of times users share your articles on social networks) played an increasingly important role because it told Google this was relevant content – but now it seems to as Entity took its place.
Entity also helps to group content from the same brand or company.
If you think about a company’s app and its website, these two Entities will tell Google which one is from the same brand and which one is from two different brands.
It is also language agnostic and greatly aids in voice search.
Why Entity is important in SEO?
In general, SEO is becoming more and more difficult and complicated, for two reasons:
- Google is using more machine learning.
Many algorithms are released so often that it is difficult to grasp the principle and really understand what Google wants.
- Paid search (paid advertising) is getting more and more expensive.
Running budget-intensive ads is the top reason many companies and individuals are turning to organic search to find customers, with the third reason being that you’re going head-to-head with Google.
- You are in direct competition with Google!
Competing with Google?
Let me explain it to you!
Above is a screenshot of the search results for the query “buy sneaker” on both desktop and mobile versions.
You can see at the top of the results page what users see when they search for this keyword on Google.
All search results show:
Google can instantly provide an answer to your question or satisfy your search intent.
- If your purpose is to buy goods => Google will display the results of pages running ads that you can shop. In addition, Google also gives results in the form of a map where you can find and buy that product.
- If your query is purely informational, Google will try to answer your question directly.
More than 90% of SEOs do not know:
Google has changed the game in distributing traffic to the website!
Google now tries to answer user questions right on the search results page (SERP)
That means you’ll get fewer clicks (because users get answers right on the results page without having to click on your website to see the information).
Therefore, the more you have to rank higher and make sure all web SEO tactics are better optimized.
Entity in On-page SEO: Create content
The question here is:
How to write better content to get traffic and still go to other sites?
A common situation is that more and more content is written to compete for rankings, not to be shared.
Below is a BuzzSumo study that shows the correlation between shares on Twitter and the amount of content being created.
That’s why we need to optimize Content according to Entity, not only to provide the best user experience but also for better presence and ranking on search engines!
Extract Entity SEO & How does Google extract Entity?
Entity extraction is to filter all Entities out of the text to process, analyze, and work with them…
If you are trying to extract Entity yourself then you need the following 4 elements:
- IDs to identify Entities such as addresses and MREIDs (Machine Readable Entity IDs). Simply put, they are like URLs but they are not.
*Note: MREID is a string of characters (usually letters and numbers) written to indicate a single Entity (person, place, landmark). Example: /m/0qs4dyq
- Big data (data) – Google has a huge data warehouse thanks to the corpus or Google index.
- A large knowledge base like Freebase or Wikipedia. Google bought Freebase many years ago, and it’s exactly for this purpose.
- Attributes, basically these attributes are relationships between Entities that help Google understand the concept behind them.
In 2012, Google released some of the data they had – about 500 entities, and 3.5 billion pieces of information in Google’s knowledge base, and all of this data was updated about 7 years ago.
You can imagine how terrible this number is up to now.
Google used an algorithm called Word2Vec to increase language recognition.
The Word2Vec algorithm is based on two other minor algorithms, Skip Grams and Continuous Bag of Words (abbreviated CBOW).
Basically, this algorithm combines words together, and analyzes words that are close together, thereby understanding how words link in the text.
With Word2Vec, Google can draw word diagrams or convert letters to numbers, and then these numbers will be represented as vectors or graphs.
**Note: If you still remember vector knowledge, please read this section carefully. But if you have forgotten it and have never had the concept of vectors, go to the next section!
Google will understand words placed in the same relationship between different languages.
For example, in the first chart, the numbers 1, 2, 3, 4, and 5 in English are one, two, three, four, and five. These numbers in Spanish are also in the same position.
Below, we have the first chart of the words pig, cow, and horse in English are pig, cow, horse and in Spanish, these words are in the same position.
Once you convert a letter to a number, the word will turn into a concept and the number represented will be completely unrelated to the language you are using.
When we join the vectors, we will see the relationship between language and country.
For example, Hanoi is the capital of Vietnam as well as Moscow the capital of Russia, the two relationships are the same.
From there, you can combine letters to derive the basic principles.
Everything becomes extremely logical.
‘THINGS NOT STRINGS’ (Object, not string)
This principle not only helps you know how many words are on a page but also what these words mean, thereby knowing what users are searching for and want to know.
If we consider this algorithm as a graph, then Entity here is the intersection (node) and the relationship obtained is the edges (edges) when connecting the intersecting points.
If you do a Google search for “United States presidential candidates in 2012”, you will get results like Mitt Romney, Ron Paul, Gary Johnson – they are all Republican presidential candidates and still alive.
However, in fact, there are also many Johnsons like Andrew Johnson or Lyndon B Johnson who are also members of the Republican party.
Next Mitt Romney and Johnson are both involved in Medicare (the federal health insurance program) All of the above connections between Entities help Google understand what the nature of these Entities are.
And if considered on a level greater than 1 billion times, Google’s awareness has been raised to a new level, not just the keyword density, but you have to stuff keywords to notify Google. What topic are you writing about?
This principle has been applied in the Knowledge Graph algorithm, in the right-hand column Google has provided various information about the keyword “watermelon”.
This principle is also shown in the carousel (red circle in the picture), it is clear that Google can understand that all iPhones, apple watches, iPod, … belong to Apple products.
And again, this principle is applied in the SERP, when typing the keyword Airbus A380, Google will automatically output search results about civil aircraft.
Google also announced they were changing this way in an announcement from Amit Singhal – Head of Google Search in 2012:
Experts call this a graph – the ability to understand real-world Entities in terms of their relationship to other entities: things, not strings.” – Amit Singhal
You’ve probably heard “things not strings” somewhere and this is truly revolutionary, changing from keyword density and text to what the entity of the text is.
We’ve seen a lot of updates this year and I think these updates are linked rather than just individual updates aimed at Entity and pulling Entity.
These updates are intended to increase natural language understanding, so Google can understand what language and content the content is about.
10+ Suggestions to help you implement Entity for your website
1. Not only building a website but also building a brand
That means your brand must provide Google with all the criteria that other famous big brands have, including:
- create a social network account
- topic content
- create an AdWords account
- link your social accounts to the website…
In which domain, logo, and maps are important signals; so don’t forget to add the company address on the website. And surely you also want to become an expert on many different topics. Then read on to tip 2.
2. Create in-depth content
So to create in-depth content, you need to first create content around a single topic but still cover all the information like a miniature library on that topic. That’s the Topic cluster method that I introduced in early April!
It is advisable to avoid writing a series of articles on mixed topics. So once on this page, you will find all the information you need on any given topic.
3. Use Google natural language API
You can go to here to log into the API and test it directly by copying a piece of text here and seeing what Entities Google finds in that piece of content.
Here is a sample Google article illustration:
Google can recognize concepts, places, people, brands, and other small pieces of information.
You can also find:
- Sentiment about each Entity – that is, are you writing that Entity in a negative or positive context?
- Salience is a value between 0 and 1 that indicates how important that Entity is in the context of the article you are writing.
In the article, I mentioned important concepts such as backlink, linking, and Cheirank … so I want these Entities to appear at the top with the highest salience score possible.
Sometimes Google also inserts links to Wikipedia articles on a particular topic. This is a good sign that Google relies heavily on Wikipedia to understand Entities.
And now Google says your article contains Entities for which Wikipedia already has its own article about these Entities. From there, you’ll be sure that Google knows the trustworthiness of your site or the credibility of articles on a given topic.
Another great but often overlooked feature is categories. With the same landing page, and the same API, Google understands what your content is in and what is not.
If you are writing a piece of content and want to know if it is well optimized, just copy and paste it here and change it as suggested by Google. Note that the most important concepts must correspond to the most important Entity in the output.
4. Build a content structure
This seems to be the basics of SEO, but I want to emphasize it again.
Your content should be easy to read, and understand, and use a lot of headlines, and tables. Google likes these. Using bulleted or numbered lists will help you better communicate to users and web crawling systems.
The better the content structure, the better Google understands your content, how different sections, paragraphs, and words fit together, and how important it is.
Don’t forget to include the source and the author’s name. There is a lot of debate surrounding E-A-T (short for expertise and authority), authority and trust pages are becoming an important factor in SEO.
5. Change standard SEO content based on SERP
It sounds ridiculous, but this is proven to be an effective way to reproduce quality content. That is, you will be based on the results of Google on the SERP to understand what information Google thinks is the most important.
Because this section has not been implemented by Google for the Vietnamese market yet, you can switch from “how to lose weight” to English as “how to lose weight” and refer to the “People also ask” section.
You can tap expand and Google will give you more questions. Use these questions as paragraphs, subheadings, and make sure the article covers all the important information here.
Also, don’t forget to check out the Knowledge Graph section if you have one. This is extremely valuable information to help you answer the heart of that topic.
Also, you can get ideas from ‘Searches related to’ at the bottom of the page or ‘People also search’, ‘People also look for.
In general, you will easily get an outline for the content just by looking closely at the SERP.
As you know, you should use schema and schema json-ld but a nice analogy is using itemprop tag “same as”.
This code will help inform Google something like “This organization is like the one mentioned in this article on Wikipedia” or “This is the social profile of this brand”…
So add this tag when inserting structured data into your page to help Google understand the direct relationship between Entities.
7. Optimizing according to the search journey
When writing content, you need to know where users come from and where they will go in the search journey.
Every query, every keyword a user types to find your site has a past and a future. You need to ask yourself what steps the user searched for before and what the next step is.
From there, you can put other articles as related articles below or insert them directly into the content.
In addition, adding breadcrumbs, and navigation to the page will help users know which query they are looking for in which topic or field.
That’s why I previously advised you to create a library of information about the topic, not just expand different articles on different topics.
8. Satisfy the search intent
Today, if you don’t directly satisfy user search intent, you won’t be in the rankings.
Google divides search intent into 4 basic categories as I mentioned earlier, including information, navigation, product review, and purchase.
When looking at the search results, we can determine what the search intent here is, from the perspective of Google. For example, with the query “buy sneakers”, it is clear that the user wants to buy sneakers.
If a user asks a question, it is clear that the search intent here is information. In other words, we can completely know the different search intent of users based on the search results displayed.
It sounds simple, but it’s not always the case. For example with the keyword “Independence Day”. Most of the year, when you search for this keyword on Google, you will see a series of results related to the movie “Independence Day”.
But what if the 4th of July is coming, the results shown are about the holiday? It can be said that search intent and Entity change over time. User search intent also changes over time.
For example, when it comes to Apple. How to distinguish whether Apple is the name of a famous brand or the name of a fruit? If Apple wasn’t as famous as it is today, maybe things would have been different.
Keep this point in mind and constantly monitor search results to understand what Google wants and how it’s changing.
9. Refer to the top 5 search results
Don’t forget to refer to the top 5 articles to see what your competitors are writing and try to write better than them.
You don’t have to watch exactly 5 articles. The number is not important, but the key is that you have to observe what your competitors are writing, what their subtopics are and how the structure of the content is organized.
Then you can do the same, or better. You can find out what information your competitor’s articles are missing, thereby supplementing your own for a better chance of ranking higher on Google.
10. Avoid having a poor click-through rate (CTR)
We all recognize that user signals (signals of user behavior that Google relies on to determine website rankings) are very important. CTR or time on site are both considered user signals.
If you were also affected by major updates this year, you should go to Search Console and check what the CTR was before the site lost traffic and see if CTR was the cause of you being affected by the update.
Therefore, you should regularly go to Search Console to see which queries have low CTR to optimize the snippet and content.
Check if your article has satisfied the search intent, or if the content is old, the content is poor, and re-optimize the new post.
At this point, I have explained in detail for you to better understand Entities in SEO and how Google perceives the relationships between Entities!
And most importantly, you already know 10 ways to apply Entity to your website. How will you apply? Let me know the results!