How Latent Semantic Indexing has Changed Search Engine Optimization

Latent Semantic Indexing
  • Home
  • SEO
  • How Latent Semantic Indexing has Changed Search Engine Optimization

The Rise of Semantic Search (and Semantic SEO)

The way that people are using the Internet is in a state of constant flux – it’s mutable and ever-changing, which is a good thing. It pushes search engines to get better at their game all the time.

Today, search is driven by user intent. We’re no longer satisfied with typing up a keyword in the search engine and sifting through tens or even more results before we get to the information we’re looking for.

When was the last time you typed in ‘restaurant Palo Alto’-or wherever – and hit the search button? Probably years ago, right?

Gadgets we’re using today make it easier for us to search for stuff on the fly. Plus we’ve gotten accustomed to search engines answering our questions – essentially reading our minds.

So today we will probably type in:

  • ‘best sushi place Palo Alto’
  • ‘great Chinese takeaway Palo Alto’

See, none of those contain the word ‘restaurant but search Google will invariably recognize our intent and serve up the most relevant results.

But search engines were not always that smart – they were, and still are, grappling to keep up with users, even though the algorithms are getting better every day.

Semantic search (and by extension semantic SEO) didn’t always exist. So what changed? What happened that enabled search engines to understand our queries on a level of an intelligent 6-year old? Well, it’s all thanks to something called Latent Semantic Indexing.

Latent Semantic Indexing – What the Heck is That?

That’s what changed. Search engines started to use latent semantic search to order information into understandable and connected chunks – text, ideas, and topics, everything you can find on the web is now interconnected and helps search engines serve up better results.

So what exactly does semantic in LSI stand for?

Basically, semantics is an old, old linguistics term. It is a branch of linguistics that studies the relationship between words and sentences and their actual meaning.

When applied to Internet search, semantics takes on a slightly different yet basically same meaning.

Semantic search is a technique that takes into account searchers intent and the contextual meaning of words or phrases in an effort to improve accuracy and display the most relevant results.

In the world of the LSI, the semantic web and searches, it’s only natural that SEO practices are going to shift as well. That shift is already happening, and I encourage you to think of long tail keywords as LSI keywords. This is because search engines look at the keywords entered, complete with supporting and related terms, in order to gauge user intent.

Right now, semantic SEO is just a buzzword. However, at some point in the future, the term semantic SEO will disappear – because the only people left in the industry will be those who understand how to thrive on the semantic web.

In practice, what this means is that Google knows that when you search for ‘highway’, in many cases, that means the same thing as ‘freeway’. When you search for ‘pop’, in some contexts that can mean ‘soda’, sometimes it has something to do with music, and sometimes it’s to do with balloons.

Search Then Versus Search Now

Back in the day, the algorithm would spew out pages based on keywords entered and the overall ranking of the page. Links equaled better positioning. Back-linking and shoddy gray hat methods were the webmasters best friends – back-linking is still up there, although the quality version of it.

This is because crawlers depend on something called co-citation to verify the importance of a page. Co-citation is one of the building blocks of semantic search and it happens like this.

When two different pages mention a page and they share a similar topic (remember, linking and shared keywords are not really required) they send a signal to the crawlers that the mentioned page is important. Thus, all those pages become interlinked and boost each other’s ranking.  Chances are that they will appear on the same page during a search because they evolve something similar to a symbiotic relationship. Naturally, one page is ‘heavier’ than the other – usually the one that gets the mentions – and it will probably be located on an authority site.

Another important signal for the algorithm is something called co-occurrence. Co-occurrence is a term that describes topically similar keywords that are not the same but naturally appear in the text containing the keyword as supporting words and terms.

For example, go to Google and type in ‘omega 3’. Open the first result (it’s an article on and do some keyword digging. You will notice that omega-3 shows up nearly 50 times on the page but it’s not lonely. Words ‘health’, ‘fish oil’, ‘krill oil’, and more show up almost the same amount of time.

Thanks to the fact that the crawlers use co-occurrence and co-citation as signals, Google search will be full of results about the ‘correct’ Omega meaning – you will not see a single mention of Omega nebula, Omega watches, or the Omega constant (a math thing).

As I’ve already mentioned, this new semantic search engine is changing how we do SEO – slowly right now but those who adapt to it immediately will reap some substantial benefits in the future.

Latent Semantic Indexing

SEO in the Age of LSI

Although SEO professionals shouldn’t panic when it comes to latent semantic indexing and semantic web search they definitely need to prepare for the inevitable – a shift in how SEO works and on what it focuses on.

That focus will be on user intent and context, not on keywords. You will have to understand how search engines ‘think’ – for the lack of a better term – and adapt your strategy accordingly.

Of course, keywords will still be important, especially long-tail keywords as they will serve as a ubiquitous signal to the search engines – they will let them know what your page is about. But the interplay between co-occurrence and co-citation is what will determine whether you succeed or fail in the future.

  1. Mentioning specific keywords will matter less and less

This is something that you might have already noticed. Type anything into the search engine and look at how many results show up not containing one or more terms you’ve used in your search.

That’s because search engines care less about providing an exact match then they do about getting the results they perceive will be most valuable to you.

Instead of trying to shoehorn your exact keyword (e.g ‘best brand of pop’) into an article 4 or 5 times, focus on finding related terms that you can incorporate into your content so you can harness the power of co-occurrence. Increasingly, ‘top soda brands’ means the same thing to the search engine as ‘best brand of pop’. This will be the most important signal search engines will use in the future – along with user intent and co-citation.

  1. Stop focusing on anchor text

LSI is not focusing on anchor text as the most important signal nowadays. Co-citation and co-occurrence are pushing it out slowly. Instead, focus on how to spread your influence and benefit from co-citation. When selecting anchor text, link naturally – there’s no need to force your exact keyword matches into anchor text.

  1. Harness the power of Google Knowledge Graph

Google knowledge graph is all about answering users’ questions on the search page and presenting inter-related terms and concepts in a way that goes beyond typical search.

It’s all about presence and while popular brands and famous persons benefit most from it, ignore it at your own peril.

Google uses and Google + profile to create Google knowledge graph panels.

To appear in it:

  • Maintain a strong Google + presence – it’s a pain but might help you in the long-run
  • Create your own Wikipedia page
  • Create a strong content marketing strategy and get people searching for your brand – it signals to Google that you’re important enough to arouse interest
  • Create valuable content that people will share

This feature might, in the future, be of immense help to marketers looking to build presence and grab people’s attention while they’re still doing top-of-funnel searches.

  1. Maintain presence and get exposure

Remember co-citation? It’s all about getting mentions, likes, and shares, and getting picked up by authority sites as a good source of information. Whenever this happens, your ratings in the SERPs will get a slight boost – this is because those mentions signal to the search engines that you’re valuable enough to be picked up by others.

Even who you mention might affect your exposure. Don’t be shy about mentioning your competition – even if you have to do it with a positive spin. SEO is increasingly becoming a PR game and building your brand is more than simply dissing your competitors. Find other ways of differentiating and use co-citation to drive the value of your page up.

  1. It’s all about semantics

Structure and create content so it references the terms that are related to your keyword. Use supporting jargon and language and make sure that you know what you’re talking about. Search engines will award you if they can figure out that you’re an expert on the topic – so make sure you actually are.

Also, build your site so everything supports each other. Crawlers will look at the whole picture, not just one article, to determine whether or not you are a match for the searcher’s intent. Semantic web search means just that – searching the web for relevant info and you would be wise to remember that all of your pages are a part of the web.

LSI Keyword Research

Since we’re moving towards LSI web it’s helpful to keep note of how to best find LSI keywords that you will center your content on, making sure you’re helping out crawlers identify your strong points easily.

LongTailPro – Essentially, most long tail keywords can be considered LSI words that correlate to a seed keyword. LongTailPro is an excellent tool that spits out a lot of suggestions that you can then using in getting your page to rank better in the SERPs.

LSI Keyword Generator – This is a free tool that allows you to enter your seed keyword and spits back a list of related terms that have a close co-relation to your search. Use those keywords to build valuable content by sprinkling them around your page – remember to keep it focused and strategically wise.

Ubersuggest – Another great tool that you can use to get a glimpse of what search engines deem to be related and important supporting terms for your keywords.

LSI in a Nutshell

Latent Semantic Indexing proved to be a driving force behind how we now do SEO and more importantly how we’ll be doing it in the future. But remember; it was prompted by and was a direct result of how people use the Internet to search for things.

If anything, I want you to take the following from this article:

  • Stop stressing about keywords and anchor text – search engines know more than you think so quit over-optimizing or you will end up hurting your efforts. Look for LSI keywords and make sure to harness the power of co-occurrence.
  • Make sure your content is valuable and engaging –so much that other sites won’t think twice about linking to it or mentioning you as an authority site.
  • Build your brand up – make sure you’re recognizable and differentiate yourself from the crowd.

And that’s it. Remember, traditional SEO signals (heavy anchor text, keywords that are shoehorned into articles) are rapidly losing effectiveness. If you’re in the SEO industry, it’s time to start focusing on user intent – its sounds like magic but think hard about how you search for stuff and how you would like to find it and adapt yourself to that pattern. And make sure you understand LSI and its development and implications. If you understand how search is developing and improving, you’ll be able to stay ahead of the curve.

Try Long Tail Pro with a 7 day FREE trial.


When to Switch Long Tail Pro