Is LSI Truly the One Basket Where You Should be Placing All Your Eggs?

If you are in the business of marketing your products or services online, you may have recently heard a whole lot of talk about LSI. According to Wikipedia, LSI or Latent Semantic Indexing is based on the principle that words that are used in the same contexts tend to have similar meanings and this method can correlate semantically related terms that are latent in a collection of text. Now, many people believe that LSI is one of the main factors Google uses to rank web content in their searches. One reason many people are leaning towards this method as a way to increase their rankings on Google is that in 2006 there were a number of patents filed by Google that speak about “phrase-based” searching and indexing. However, these patents were not referring to the LSI method; they were referring to phrase or word co-occurrence techniques.

Latent Semantic Indexing is a way to capture relationships between word pairs and also determine the way words interact at the level of meaning. This can be done only through a series of mathematic formulas that take billions of documents into account. It then extrapolates all the related words in each of these documents, determines the number of times they are used and fine tunes the list. Now, this is a very simple way to explain a fairly complicated method, but it should give you an idea of what LSI means.

So, if Google was using LSI to determine its search results there are a number of things that would occur. One expected outcome would be that if LSI is being used you would see very similar search results for the singular form of a word and the plural form. This makes complete sense since both car and cars would have the same related terms. Another expected outcome would be that the synonyms for a word would also create very similar search engine results. Of course, the results in both of these cases would not be 100% identical, but they would be extremely similar. A very easy way to see if Google is utilizing LSI is to simply do a search on the singular form of a word and a search on the plural form and see what happens.

Now, if you just did your searches, you would have seen that the search results for singular vs. plural forms of the same word give you very different results. If you are looking at two synonyms, you will see that the Google results differ by at least a factor of 10, in most cases. What does this prove? Mainly, it proves that Latent Semantic Indexing is definitely not the technique that Google is using to create their search engine results. So, those people that are putting all of their eggs in the LSI basket need to understand that this is definitely not the principle method to follow.

While Latent Semantic Indexing makes sense to use while determining the content structure of your website, it is not the determining factor for high Google rankings. SEO is still the most important aspect of internet marketing. Optimizing your web content for the search engines by using top ranking keywords, 100% unique, well written, search engine-friendly content is still the smartest way to get top results on the search engines. At this time, SEO is still the one basket you should definitely be placing the majority of your eggs.