Latent Semantic Indexing Matters

Latent Semantic Indexing (LSI) has unexpectedly become a point of contention in the world of search marketing. LSI is a mathematical method used by search engines to analyse the raw data of a document and establish relations between ideas, keywords, and concepts.

Some marketers argue that LSI doesn’t impact a website’s page rankings in a meaningful way, whereas others stress its importance in search engine optimisation.

Despite having some detractors, you should pay attention to latent semantic indexing because it plays an important role in optimising websites for search.

In this blog, we’ll take a closer look at all the different ways that LSI can complement your overall SEO strategy.

What is LSI and Why is it Important

If we boil down the term to its essence, LSI simply means more relevant keywords. While this concept has only recently entered the digital marketing world – SEO in particular – it has a long history in academia.

Latent Semantic Indexing is a mathematical model developed in the 1980’s designed to optimise data retrieval. It searches for the underlying relations between words to make informational retrieval more accurate.

LSI analyses unstructured data found in documents in order to learn how different words and concepts are related.

It was used primarily for scanning a small number of documents, but search engines have developed the method and used it to scan web content.

To get a better understanding of the role LSI plays, we need to know how search engines function.

In essence, a search engines rely on algorithms to identify two things:

  • The content and its context
  • A searcher’s intent and how it relates to particular keywords

Algorithms that do a good job at understanding these two things are the ones that give us the information that we need.

So, LSI makes search engines run more effectively by finding related keywords and synonyms.

Therefore, LSI takes into account the role context plays in natural language. Earlier search engines struggled with this aspect of data retrieval. In particular, they had problems with understanding how words had different meanings depending on various circumstances.

For instance, the words “hot” and “dog” are easy enough to understand, but they get new a new meaning when used together.

Or, let’s think about the word “cell”. The results page targeting this word could be anything from websites about the 2000 American science fiction film starring Jennifer Lopez, to pages about the basic biological unit of all organisms.

If the page is about the film, then the LSI keywords that indicate this could be:

  • “film”, “science fiction”, “Jennifer Lopez”, “reviews”

On the other hand, if the page is about biology, then the LSI keywords could be:

  • “cell biology”, “cell structure”, “animal”, “plant”

Search engines today rely on LSI keywords to provide much needed nuance in order to produce relevant search results.

One important note before we continue. Don’t confuse LSI keywords with synonyms. LSI keywords are terms that are closely related to a particular keyword:

For instance, take the word “hiking”. Its synonym is “walking”. While you can use that synonym in your content, it isn’t an LSI keyword for “hiking”.

Those would be terms like: “boots”, “nature”, “backpack” and the like.

Problems Search Engines Overcame

Earlier search engines had blinders on when they were scanning the internet – the webpage either had the keyword or it didn’t. There was no nuance. And the only way search engines gauged the relevancy of a page was through keyword density.

This brought about the infamous era of “keyword stuffing”, where unscrupulous content creators would load their pages with targeted keywords as much as possible while paying little attention to actual content value:

Image Source: THE MANIFEST

Other notorious SEO ranking practices included listing phone numbers and pasting chunks of text with the names of cities and states they were targeting.

As a result, this “strategy” was penalized by search engines by giving priority to LSI keywords instead of keyword density. The shift towards understanding natural language search queries happened with algorithm updates like BERT, Hummingbird and RankBrain. The Penguin update gave keyword stuffing the final blow.

This change in algorithm is best illustrated by the conversational nature of our queries and the fact that voice search is becoming increasingly popular.

To sum up, Google and others use LSI to understand both what the searcher is looking for and the webpage content.  Search engines like Google aim to understand a website’s overall topic.

Therefore, people find what they are looking for in the results page.

Benefits of LSI

If we stop to consider just how massive the internet is (and it’s growing exponentially), it becomes apparent that information retrieval is not a straightforward task.

Search engines are always looking for the most relevant and insightful content for a particular query. Therefore, they reward pages that use appropriate LSI keywords with higher ranking, which in turn drives more traffic.

However, there are advantages that go beyond ranking on search engine results page. Publishers benefit because their content is shown to their target audience. Searchers also benefit because they seamlessly find the information they need.

It’s a win-win for both parties.

LSI also helps you avoid being labeled as spam by search engines. This can happen when a user lands on a website via a search engine but doesn’t find the content they need. The searcher can report your website as spam to Google, which will harm your ranking.

This is also means that with LSI, you will lower the bounce rate of your website if people are finding what they need through their searches.

For instance, if you sell Ray-Ban’s Aviator glasses, you don’t want to attract traffic from people who are interested in reading about the film “The Aviator”.

Finally, one of the great changes LSI has helped bring about is the quality of content. As we mentioned in the previous section, long gone are the days of keyword stuffing. Publishers can focus on generating content that is made for people and not bots. Therefore, the overall flow of texts is smooth and user-centric as opposed to keyword stuffing monstrosities of previous years.

Generating LSI Keywords: What to Tools to Use

Let’s now look at the best ways of finding LSI keywords to use in your content in order to get a strong SEO boost.

Good Old Google Search

The simplest way to find LSI keywords is through Google. It’s also free. As you type your keyword in the search box, Google displays other terms that are related to your query.

Here’s an example of how Google autocompletes our query with related terms for “aviator”:

Description: https://blog.hubspot.com/hubfs/%5BAgency_Post%5D/lsi-1.png
Image Source: HubSpot

 The keyword “aviator” can refer to many things: it could be anything from a type of glasses from Ray-Ban to the film starring Leonardo DiCaprio from 2004 called “The Aviator”.

As you can see, the offered suggestions take us to vastly different websites, so it’s necessary that we provide Google with more keywords so that we get the information we need.

Notice that at the bottom of the results page is a “Searches related to” section which displays common terms Google associates with your query:

Image Source: HubSpot

This list is a good place to start your LSI keyword research for your query.

Google’s Keyword Planner

If you have a Google AdWords account, then you can use their planner to research LSI keywords.  

Enter your target keyword in the search bar. Then click on “Get Ideas” and you will see a list of associated keywords along with how strong the competition is for that keyword – low, medium, or high:

Image Source: HubSpot

In our example, we again searched for the word “aviator”. We are shown the average month search volume along with the competitive ranking and suggested bid price for each keyword.

Pro Tip: Instead of entering a keyword, you can include a landing page. Use this feature if the competition is ranking well and you want to find a way to leapfrog them in the results.

Google will thoroughly examine the website and list out keywords it identifies to be relevant to that site’s subject matter. 

LSI Graph

LSI Graph is a free tool designed to generate LSI keywords. Just like in Google Keyword Planner, you only need to enter your target keyword in the search bar to get suggested LSI keywords.

Image Source: HubSpot

We typed the keyword “aviator” yet again, and the listed results are easy to read but don’t come with the additional information like competitive ranking.

Ubersuggest

One of the most popular free tools for finding LSI keywords is Ubersuggest. Once you’ve typed the target keyword, Ubersuggest spits out a list of related queries along with their search volume, cost per click and difficulty.

We entered the phrase “SEO Audit” and these were the results:

Image Source: BACKLINKO

 The neat feature with Ubersuggest is that it gives you the likelihood for ranking in top 20 for that specific keyword.

Picking the Winners and Discarding the Losers

Once you’ve put together a list of LSI keywords, you need to select the best keywords to use in your content. You need to look for terms that are closely related to your content, but also enjoy a healthy search volume.

Different Ways to Think of Intent

When looking for LSI keywords, it’s always helpful to identify those that directly address questions searchers have about the target keyword. This means putting yourself in the shoes of your audience and thinking about the different search queries they’re using to find relevant content.

There are 3 main types of intent we use when we search for things online. To illustrate the importance of this point, let’s go back to the “aviator” example.

Let’s assume that the keyword is on a webpage about the film “The Aviator”. These are the different types of intent a searcher could have to use the term:

  • Informational: This refers to a general query. For instance, “What is the plot of the film?”
  • Navigational: This is a specific query. For example, “Who is the main character in the film?”
  • Transactional: This query is about purchasing a product/service. In our context, it could be about streaming the film online.

Keep these different types of intent in mind when you’re thinking about the motivation of your audience.

Use Terms that Provide Context

As you now know, LSI keywords must be closely related to your content. So, for example, if you were to write a review of the film on your website, you must only include LSI keywords that relate to the film only. That means avoiding any keywords that are related to the Ray-Ban glasses, for instance.

How to Use LSI Keywords

Now that you’ve picked the best LSI keywords, where should you use them? There is no rule when it comes to this question. However, as long as Google can find them on your page then you’re set.

So, you should include LSI keywords in:

  • The Title Tag
  • In the Image Alt Text
  • Your H1 tag
  • H2 and H3 as Subheaders
  • The first paragraph and last paragraph
  • Meta-description
  • Content

As you’re adding LSI keywords into your text make sure that it sounds natural and includes relevant content. Try to limit each LSI keyword to one per page. Also make sure that your LSI keywords are spread out and avoid keyword stuffing – you’ll be penalized!

Remember that user-experience must be a priority, because it’s a crucial ranking factor.

If your content is all fluff and no substance, then users are more likely to leave your page too quickly. This will cause your bounce rate to spike and your ranking will suffer.

LSI Keywords Are Important

While important, LSI keywords are only one part of an overall SEO strategy. You still need to add valuable backlinks, relevant alt tags, and most importantly, you must produce superb content. There’s no way around this!

Nevertheless, invest some time and energy into researching LSI keywords that you can include into your content. While LSI keywords are not an SEO magic bullet, they will improve your ranking.