There are many questions related to Entity: What is Entity and how do they relate to SEO? Why are they important? How can you optimize the Entity for your website?
Kevin Indig solved all these questions in a great talk that he shared at Optimisey.
And Kevin gives a warning when you get involved in some pretty difficult and complicated topics like machine learning and Natural Language Processing, but you should stick with it.
He completely turns all that theory into advice and that advice works for what you can really do for your website and content plan by 2020 to help your SEO efforts achieve success in the future.
Content:
Why is ENTITY important in SEO?
Entity in SEO: CREATE CONTENT
What is ENTITY in SEO?
How GOOGLE performs ENTITY extraction
THINGS is not STRING
10 Tips to extract ENTITY
- 1 # Branding and more than just a website
- 2 # Create expert-level Content
- 3 # Use Google’s natural language API
- 4 # Structure your Content
- 5 # Content inversion optimization technique in SERPs
- 6 # Schema
- 7 # Optimize your search journey (journey search)
- 8 # Satisfying user intent (user search intent)
- 9 # Look at the top 5 ranked results
- 10 # Avoid low clickthrough rates
Why is ENTITY important in SEO?
Let’s start with the question: why is this really important?
And there are three big reasons for this question.
Overall, SEO is becoming much harder than it was ten years before I started and when I went up. @Kevin_Inig speaks at a #Optimizely event
SEO becomes much more complicated, for many reasons.
So, the first reason, the complexity of SEO has increased because Google is using more machine learning. The more smoothly the ranking elements are deployed, the more you know over time, so it is hard to identify a model and really know and understand what is going on.
And so, the second reason here is that paid search is generally more expensive.
This is a slide from the internet trend report from Kleiner Perkins that I really can, really recommend reading, and it shows that the price for paid search is generally on the rise.
Therefore, many companies and many people will return to organic search to get customers.
And then, the third reason is that you compete more directly with Google.
So here is a screenshot from the query ‘buy sneakers’ on desktop and mobile devices – you see the fold right here – which means that this top is what people see. when they search on Google.
And what I am highlighting here is that all the SERP features that Google shows you immediately satisfy and answer your question or satisfy your intentions.
So, if you have any transactional intent – you are trying to buy something – Google will show you certain ads where you can buy the product. It will show you the map where you can buy certain products and if you have an information query or information search (try to find something) then Google will try to answer your question directly.
So, Google has changed in terms of sending traffic to websites and now they are trying to answer the questions people have – right in search results. @Kevin_Inig
It means you will get fewer clicks than average, so you have to rank higher and make sure you know all your tactics better.
So, the bottom line: we need to be more efficient. And as we all know Content is one of the great pillars of SEO or has become one of the pillars of SEO. @Kevin_Inig
Entity in SEO: CREATE CONTENT
So, the question is: how can we create better content to get traffic still going through other websites?
And we also have to consider creating more Content to have more competition and less content to be shared.
So, this is from a BuzzSumo study that basically shows Twitter’s share over time compared to the amount of content generated. In this case for machine learning topics (but you will see the same thing happening in all different types of topics) it is even more competitive when it comes to content.
And that is why I am here to talk about Entity optimization.
I will give a few examples.
So, we will start with a theory section and then we’ll move on to more practical and tactical advice.
What is ENTITY in SEO?
Let’s find out. What is an entity?
Entity is a slightly abstract concept but I will try my best to explain it to you.
Entities can be different things that they can be places, people, organizations, but they can also be concepts or ideas. So, for example when you have a text, you will see Google recognizing different types of entities like people, an organization like Google, a timeline like 2007 or a country like the US, etc.
And this is how Google improves its natural or linguistic understanding and basically its content understanding.
If I try to identify which entities, I would say they are semantic objects that help machines understand the language clearly. @Kevin_Inig
And here is the crucial part right here to understand: it is not just about what you are, saying clearly, but about implicit concepts, ideas, and trends.
The obvious question is, why does Google focus on entity or entire entities?
And there are a few good reasons but here are a few of them.
So, first of all we know that a lot of you know ‘link chart manipulation’, build links with stuffing and spam.
So, Google is getting smarter and you should try to stay away from that. For a while – you have seen that people think social signals are becoming more important because it is a better indicator of content related to Google – but right now it seems these are the Entities.
It also helps group content from the same brand or the same property. So, if you think about an app from a company and website – and a few other things we will talk about in just a second – then entities that allow Google to really understand what comes from the same brand.
It is also agnostic and it supports voice search. And I will go deeper into all these reasons.
How GOOGLE performs ENTITY extraction
So first of all, if you try to do entity extraction yourself – and perform the extraction basically try to get these entities out of the text to handle them to work with them and so on. – There are basically four different parts that you need to know:
- You need an ID to identify an entity as an address and have online MREIDs that are basically like URLs but not the same concept.
- We need Big Data – which Google has because of its data store or its index.
- We need a Knowledge Base like Freebase or Wikipedia and remember that Google bought Freebase a few years ago and is probably correct for this purpose.
- And then we need Attributes – and attributes are basically relationships between entities that help Google understand the concepts behind them.
And in 2012, Google announced some of the data they had – this number is thought to be 500 million objects in their knowledge base and 3.5 billion facts (facts) and it already exists seven years ago.
So, you can imagine how high that number is today and how good it is.
And Google uses a set of algorithms called Word2Vec to increase their knowledge of the language. It is based on two sub algorithms.
The first one is Skip Grams, the other is Continuous Bag of Words – it is the word ‘CBOW’ – and it basically combines all these words.
It looks at you knowing that it basically looks at which words are next to each other – and on a large scale. And it can understand how words are related in the text.
For example, subject, object, etc. And then it can also cluster your content into categories.
And by the way, this is publicly accessible information, this is from the natural language processing API, which I will talk more about in this presentation.
What it shows is that Google has a lot of different categories there – I think about a hundred – to cluster content into and understand what the broader scope and subject of that content is. And with that algorithm, it can basically map words or convert words into numbers, and these numbers can be mapped onto a vector or graph.
That is not fun but this is! And it can basically understand words that have the same relationship in different languages.
So here one, two, three, four, five and these are the same numbers in Spanish. And you see them almost in the same place? They almost have the same position.
And the same goes here: you have a pig, cow, horse and in Spanish, they almost have the same place.
When you translate words into numbers, they become concepts and what they represent, becoming very clear independent of the language you use. It is very tall, very strong. @Kevin_Inig
At the same time, when you connect these vectors, you can understand the relationship between them. And this works in different formats.
So, do we really have comparative languages and countries? So, these are all capitals of different countries.
For example, Ottawa and Canada are like Moscow in Russia from a conceptual perspective. Both are the capitals of a country at the same time, Verb tense and their relationship between men and women.
And here are just a few examples of languages.
And what is allowed is that you can use words to make basic calculations. So, if you want to say:
King – a man + a woman = a Queen.
That suddenly became extremely reasonable, right?
THINGS is not STRING
It really helps you not only understand how many words on a page but it also helps you understand what they mean. And you can recompile to understand what users are really trying to figure out and what users really are trying to know.
So, can you imagine if you mentioned a diagram you could think of entities like nodes and relationships as edges?
And so, when you joined the United States presidential candidate in 2012, you had for example: Mitt Romney, Ron Paul, Gary Johnson and of course they were all Presidential candidates.
But then they are all living, aren’t they? They are not dead, like Abraham Lincoln or George Washington, they are all members of the Republican Party.
And then, there are others who also have the surname Johnson – like Andrew Johnson or Lyndon B Johnson, also part of the Republican Party.
But then, for example, Rom Romney and Johnson were both involved in Medicare?
So, through all connections between entities helps Google understand all that they are saying.
And if you increase the scale by a billion times, that is really a really strong understanding different from the keyword density we used to look at a few years ago – five to ten years ago, when I was not able to write – you only need to write text with keywords.
And so, this obviously has been applied to the Knowledge Graph example where Google shows you lots of different information about all kinds of words.
It is also shown in the carousel here, where Google clearly understands that all these brands have created products for project management software.
And it is also being applied to other SERP features – suddenly starting to pop out of nowhere – meaning those features are displayed in search results pages like these passenger planes.
I searched for Airbus A380 and Google could understand that all this is also a passenger plane.
But Google also publicly announced that it would make that change and that was in 2012 a notice from Amit Singhal, Head of Search at the time.
I will not read it all for you but basically, he said – and here is the highlighted part at the bottom:
That is what we did on a smart model – in a “graph”, understanding real-world entities in their relationship to each other: things are not chains. (string).
– Amit Singhal
And you have probably heard this ‘Thing is not a String’ a few times in various blog posts and that is exactly the change I am trying to convey here. It is a shift from just the number of keywords and text and more to a deeper understanding of what the text really is.
So, we have seen a few algorithm updates this year – with a few things I mean almost an update every month – a lot has changed. I have worked with a lot of big brands and websites that lost a lot of traffic.
And I and a few others in the industry think they are all connected. These are not updates that occur in isolation but they all play a bigger role, which is the type of shifting to the entity and extracting the entity.
And if you want to read a really good article about that, take a look at AJ Kohn’s algorithmic analysis
She has a series of five stunning articles on Entity and Entity Understanding that really show the various applications of the concept, and voice search, and organic search, and mobile search.
But Kevin noted: these updates are natural language understanding updates which means Google has refined what language and Content knowledge really is.
10 Tips to extract ENTITY
So, with all that theory, you should think about what to do with all the things that I have just given you?
That is why I compiled 10 tips to apply the theory of entity extraction.
And the first one is really:
1 # Branding and not just a website
Now that sounds abstract, but what it really means is that when you have a brand, make sure you give Google all the indicators that popular and popular brands have in search and on the Internet.
Create social accounts; you can spend on AdWords; ensure that there are people associated with your brand and company as authors or employees if you have an application that links it to your website.
Obviously, the domain name is a strong index. Logo is a strong indicator. Position is a strong indicator. So, do not forget to have your company’s address somewhere on the website. I have been forgotten in the past.
And obviously you also want to be an expert for various topics. So, let me talk about the topic.
Tip number two is:
2 # Create expert-level Content
Now we have heard the idea of”10x Content” but I really want to solve that problem because I think it is really important.
At Atlassian, what we did was we created so-called micro websites. And a micro website is a collection of pages that revolve around a single topic – and by the way, we have these pages in subdirectories not on subdomains.) as they are commonly used.
Make sure they are in a sub folder because Google easily recognizes that they belong to the same domain and therefore belong to the same brand. @Kevin_Inig plays
And these micro websites, as I said, all revolve around a single topic. Here you will find The Agile Coach ‘, all related to the Agile theme and you will find everything from Agile, how to apply it, practical rituals, all the small details and everything that you know you really need to become a real agile pro.
We also have a micro git site and that’s all you need to know about git’s topic.
And you know this is an example of how traffic will look over time. From the end of 2015 until recently, we had a nice boost along the way as Google’s language understanding got better and as we continued to build our website.
So, think about this expert-level content, first of all the content revolves around a single topic but also content that looks like a library, right? Not a bunch of blog posts mixed with a lot of different topics but a type of library you know.
That if, when you land there, you can find out everything on the subject.
And the third advice is:
3 # Use Google’s natural language API
So, I talked about this earlier. If you go to https://cloud.google.com/natural-language, you can not only register for the API and check it, but you can also check it directly. On that landing page, you can copy and paste your text and see the Google entities found in your Content.
And I did that here with an article I wrote and it found 461 entities. That is what it all looks like.
This means looking at how many entities Google can recognize. Are these all the concepts, places, people, brands, and things that Google understands in the text?
So, this is huge. You can also learn about the emotions of each entity – that is, do you write about it in the context of positive or negative concepts? And it is very important that the salinity is here.
So, salience is a value from 0 to 1 that shows how important you really are in the larger context of the article you are writing.
So obviously you know something like: backlinks, links, Cheirank, etc. These are important points and concepts in the text that I wrote. And I want them to appear here as much as possible.
Later, Google will sometimes link to Wikipedia articles on that specific topic. That is a good sign because Google relies heavily on Wikipedia to understand entities best, when they show you you have lots of entities, there is also an article on Wikipedia. You can be sure that Google understands more about your importance or about the importance of the text that I wrote for that specific topic.
A well-overlooked feature is categories. This is still the same landing page and still the same API – so here you can see if Google really understands what is right for your content.
So, this is about web services and business services. That is the highlight and Google shows very high reliability for these two categories as well because it is really good.
And then if you use what you know – if you create content on a smaller scale but you want to make sure it is optimized – just simply copy and paste your content into that NLP API and simply edit a bit.
Just look at what appears and you know and make sure that the most important concepts you are trying to convey are showing up as the most important entities in the report.
4 # Structure your Content
And this is a relatively basic thing that you learn when starting SEO but I just want to highlight it again.
Make sure your Content presentation is easy to read, understand, use many headlines, use things like tables. Google likes those things.
Lists – you know bullet points or numbering do not matter – just make sure you make it easy to communicate not only to readers but also to the machine.
And the more structured your content is, the easier it will be for Google to understand how different sections and paragraphs relate to each other and how important they are.
Do not forget your author.
Having lots of conversations about EAT means Expertise and Authority and Reliability is an important factor nowadays in SEO.
I am not saying this is wrong but I am saying that if you are an author writing on many different websites and creating a lot of content Google will understand that this author is an expert in that particular field and can – and Here is an assumption here – maybe you know will give you a reward if you have an author writing your content.
And let me emphasize this again, this is an assumption I have no evidence for that – but it makes sense when you think about the relationship of entities.
Yeah: titles, descriptions, related posts, all interesting stuff!
5 # Content inversion optimization technique in SERPs
And this may seem a bit pointless when you think about it but has proven to be an effective tactic to create good content many times.
I mean, use the features that Google shows up in search results pages to understand what Google thinks is important, right?
So, when you have the ‘People Also Ask: People also ask’ box and you write an article about ‘How to get upgrades with miles’ looking at the question people see, on what Google shows here.
You can – when you open some of these, Google will expand and show you more questions – do you use that? Use it for paragraphs for subheadings and make sure you include all the information that seems important here.
Also consider integrating the Knowledge Graph, if available, and see the data that Google displays. A Knowledge Graph here is all super valuable and can only help you become more important for that particular topic.
And then there are a few other things you can use as Google suggests. Obviously, it is a no-brainer but still very valuable. There is ‘Search related to your query’, this is at the bottom of the page or of the search and it is called ‘People Also Search’ or ‘People also look for’. Now ‘Related Searches’ become very interesting. Make sure you get that on your radar.
And then in the Knowledge Graph, now you have ‘People are also looking’ to get inspiration from that.
It is easy to combine a summary of content by just looking at what the actual search results pages look like.
And then of course Schema.
6 # Schema
So maybe you all know that you should use structured data and json-ld schema. But there is one really striking data and that is the ‘item as’ itemprop tag.
It helps you say, basically you know the code, ‘Hey this organization is like the organization mentioned in this Wikipedia article right here’ or ‘These are the social profiles for this brand’, so make sure you tag carefully when you add structured data to your site because it really helps Google understand the direct relationship between entities.
7 # Optimize your search journey (journey search)
And what I want to say is that when you create Content, try to grab where the searcher goes and where he is going.
So, my point is:
Every query, every keyword that takes visitors to your website has a past and a future – as in the steps that searchers might have searched for before? And what is the next one? @Kevin_Inig
And you can use it to display other posts at the end of your article or even mention it in the content – you know more breadcrumbs, navigate to your page so that users can receive them. Get complete information of the topic in which the query area is located in which he is searching.
And this is what I want to say earlier with creating a theme library of yours and not just you know how to smash out different blog posts on different topics.
8 # Satisfying user intent (user search intent)
So today if you do not satisfy the direct intentions of users, you will not rank. @Kevin_Inig
This is a fact and it seems that Google has provided us with four basic intentions: information search; find places; looking for a brand; and buy something or you know how to buy something.
And when we look at the search results pages, we can determine what Google wants to see right now is an intention. So, for example, ‘buy sneakers’ it is clear that someone wants to buy something or you know it clearly.
But when people ask questions, this is clearly a very informative intention. And when we put all these together, we can identify all these different user intent depending on which search feature is available.
I wrote a small article on the so-called ‘User Intention of Steroids’ where you know that you easily identify what searchers are doing right now.
It sounds very simple but in many cases, it is not true. Imagine something like “Independence Day”. During most of the year when you Google ‘Independence Day’, you will see a lot of results about the movie Independence Day.
But nearing the 4th of July, you will see everything about the holiday right? Therefore, these intentions and entities are seasonal. Their intentions change over time.
I thought of something like Apple right? Which is now clearly a brand but also a fruit. So, when it is not imagined today it is not known so you will see something very different right?
So, keep that in mind and use the search results feature to understand what Google is trying to see and what Google is trying to get here.
9 # Look at the top 5 ranked results
And understand what they are writing and try to be better, right? You should not only include and summarize topics in your content.
And this is just an arbitrary number – you know the five results that I want to show here – but you should see what they are writing; what is their topic; how their content is structured.
Do it the same and make it better right?
Think about information that may be missing in these articles. Because obviously there is a reason Google ranks these things so highly.
And the last advice is
10 # Avoid low clickthrough rates
So, for a long time we said, or a lot of people including myself said, user signals are really important.
User signals are something like clickthrough rates or time on site but now we really have proof.
So, this is from a website that I have been working on in the health sector – the huge website – that has been affected by one of these bigger updates. This is especially in March and when I went to the clickthrough rate in Search Console I saw that almost all of the keywords or queries that the site lost in traffic had really weak click rates.
So, if you have been affected by one of the major updates this year, go back to Search Console and check the clickthrough rate before your website loses traffic and see if there is a possible clickthrough rate. Low may have enabled the update or not.
I bet this is the case.
So, you can often go to Search Console and see which queries show a low clickthrough rate and then adjust accordingly, optimize your snippets, optimize Content, see if you may not hit the user intent or the content is just old or poorly designed and make sure you optimize them.
So, this little rundown – I will not go through all this again – but give you a quick second to get an overview.
This article is really great, do not forget to share and leave a comment.
Good luck!
Bizzvn, Source: optimisey