We have a problem!
Suppose I made a search about apples, and my search showed Apple phones, whereas I wanted to look up for the fruits.
Apple here is a polysemic word meaning the same word has different meanings.
Mouse (the animal or computer device) and Bank (the financial institution or the river banks) are a couple of other examples.
How can a search engine understand the different meanings and possibly offer irrelevant results?
Do you see the problem here?
It doesn’t end here.
Suppose I am looking for information relating to cars. When a query is made about cars, a primitive search engine would find all the information that has the word car in it, but leave out that use the words automobile or vehicle which can be used in place of the word “car”.
Whether it is the difference between big and large, or fall and autumn. These words are called synonyms.
How can a search engine understand other words that mean the same thing?
Now in both the situations a search engine will not know or understand the meanings and differences unless you tell it.
How do you know the difference as a human?
You would look at the context where these words are used. Look at all the words used to describe and then arrive at a conclusion.
This process is called latent semantic analysis (LSA) or latent semantic indexing (LSI) and it is a natural language processing technique developed during the 1980s.
What is the meaning of LSI?
LSI stands for Latent Semantic Indexing and it is a computer program designed to understand synonyms based on context.
This was implemented to help find the relation between words and concepts within a piece of content and provide more relevant results to users.
What are LSI keywords?
LSI keywords are not synonyms but words that are closely connected to and supplementing the main keyword, hence giving the users and search engines a context to understand what the content is about.
It is no longer about focusing on one keyword that matches with the search query but Google is now more ‘intuitive’ about the user intent in searches. When Google crawls a page, it does not look for one frequently used keyword to understand the content, it analyses everything that is offered in the page.
While we are still talking about LSI keywords, did you know what John Mueller from Google said?
See for yourself!
What’s going on here?
Let us rewind a little and start from where it all began.
Why was LSI created?
It was created to solve some problems like the ones I mentioned earlier in my introduction.
The problem was that any person making a search query did not necessarily use the exact same words for which the content was indexed or the words used may mean different things.
How does LSI work?
I cannot explain the technicalities, but I can say that mathematical formulas are used to identify the relationship between words and phrases from a piece of content.
If you are talking about the product Apple, you would use the words phone, electronics, Ipads, Iphones, chargers etc. And for the fruit, you would use the words like fruit, apple tree, green apple, apple pie, apple juice, etc.
If you are talking about the fall season, you would use words like seasons, summer, winter, sweaters and so on.
The search engines would use this information to relate to the search query rather than looking for the exact words in the query and deliver the most relevant results.
Busting the LSI Myth:
Google does not use LSI technology.
How can I say this for sure?
Because, Google representatives said so!
If you are still not convinced, let me offer a few more arguments.
- LSI was created in the 1980s long before the world wide web took over and was still beyond comprehension. LSI cannot work on something on such a large scale.
- It was created to index a known finite set of information. The world wide web is extremely dynamic and expanding which requires analysis with every significant update. This is not practical as it takes a lot of processing power.
- LSI was patented. And even if the patent expired in 2008, Google was smart enough before that to understand and process language.
- Google applies the BERT model to search. It is a model that processes words in relation to all the other words in a sentence rather than left to right and one-by-one order as it is done normally. This enables the model to consider the entire context of a word by looking at the words that come before and after it. This method is useful to understand the user intent behind search queries.
So you see! Google does not need LSI.
I rest my case!
Though Google denies using LSI technology, it still looks for semantically-related words to understand the relation between words.
Let us not rule out the concept altogether.
SEOs continue using the expression “LSI keywords” for related keywords, though it is inaccurate.
Google does depend on other words in the content to identify the context of the words used and evaluates them for ranking by relevance.
That means these words, let us maintain calling them as LSI keywords, are useful and important.
Why should you use LSI keywords (or semantically-related keywords)?
Employing LSI keywords in your content is good for SEO.
Isn’t that what you wanted to hear too?
1. They help boost your website’s ranking:
Incorporating LSI keywords in your content improves the content quality by making it more relevant and valuable. Search engines are bound to choose content with semantically related keywords irrespective of what your focus keyword is. It helps search engines understand your content quicker and rank them.
2. They improve the credibility of your content:
Any content that provides high value attracts more traffic. While you will not be using the same keyword repeatedly or over-stuffing keywords, LSI keywords improve the relevance of your content to search queries because Google understands your page better. This makes your content more credible.
3. Your content is more accessible:
The improved relevancy of your content increases visibility and hence more users can find your content in search.
4. There is a decrease in bounce rates:
Any content driven by LSI is extremely targeted and reaches its targeted audience. It is relevant and useful content leading to better engagement. This decreases the bounce rate.
All these benefits translate into enhanced user experience and hence improved SEO.
How to find LSI keywords?
The most common and probably the biggest mistake made in keyword research is focusing on one keyword. To start with, you need to find keyword clusters around your niche. Then you can find the related keywords around it. You can use the following tools to find keywords.
1. Google Autocomplete:
You must have noticed how Google is always trying to finish your sentences when you type a search query. If you want to find related keywords, you need to do just that.
Type in your target keyword and the search box drops down with a list of query suggestions. The words in bold are the related keywords you need to incorporate in your content.
2. Related Searches:
When you have got your search results, if you scroll down to the bottom page, you find a section titled “Related Searches” containing a list of search queries relating to your search. These are more ideas for your list of LSI keywords.
3. People Also Ask:
This is another section you would normally find on your search results page. They are the questions that people are also asking relating to your search topic. You can incorporate all these questions into your content.
4. LSI Graph:
It is a keyword generator tool available online with both free and paid versions. You can search any keywords and find the list of related LSI keywords.
5. Ahrefs Free Keyword Generator:
This is a great tool to search keywords and all the phrases and questions relating to focus keywords.
You can do your keyword research with these tools to identify the LSI keywords. But when you apply them in your content, you need to use words that make sense. You must include them naturally in your content. You can form keyword clusters to cover a topic in depth so that you include all possible related words.
Conclusion
Search behaviour is evolving on a daily basis. 15% of the billions of searches made everyday are absolutely new and never seen before. People at Google have built a way to find results for queries that cannot be anticipated.
“Search” is an open problem!
I understand that the existence of LSI keywords sounded debatable. But forget the technical words for a moment.
The end goal does not change. Creating content that solves search queries and providing the best possible user experience.
The fact remains that you need to use more keywords to describe the focus keyword, if you want crawlers to understand what your content is about. It is also an efficient way to create in-depth comprehensive content that can rank.
If you are knowledgeable about a topic, you will include all the keywords naturally and sensibly into the content. These keywords can boost rankings and leaving them from your content marketing plan can cost you big time.
Do you believe in LSI keywords? Let me know in the comments.
