Almost everyone agrees LSI is here to stay, and will keep on increasing in importance in the search engines. LSI or Latent Semantic Indexing is a way to evaluate a document based on semantically related terms using a sparse matrix with rows corresponding to terms and columns corresponding to documents. A weightage is assigned to these terms based on the number of times they occur in each document.
It involves some complex mathematical equations, but in simple pie terms it just means the way search engines will form an opinion of your website or page based on groups of related terms, phrases, and keywords.
If you wrote about apples, then a word like Gala could also be related, as would terms like:
- Fuji
- apple pie
- orchards
- cultivars
Unlike the term, “comparing apples to oranges” though, this model tries to guess your concept based on all direct and indirectly related keywords. If there is a lot of fruity terms and names in the site, then apples are related to oranges as far as the search engine is concerned (it must be a fruit site), but definitely not an Apple computer. You won’t find anything related to fruits on the Apple site either…
A further example, your site is about spiders, but how are the search engines to know what kind of spiders these are? If there’s a lot of arachnid stuff and terms on your site, then it must be about real spiders, (not web spiders).
Being a search engine puts Google in the ultimate position to retrieve from it’s database limitless related words and phrases, including rare ones. This is what you wouldn’t usually know, unless you were someone really passionate about your topic to research everything about it, which should reward the genuine Joes from the scammers, spammers, and GRQs.
How is this LSI knowledge important to a webmaster?
LSI is being used by search engines to determine how relevent your site or page is in relation to a term. The more relevent term gets the higher rank, of course. So when writing your article, a tip is to have a broad keyword in mind, and build your article around that keyword, with well researched, related terms aplenty. Actually, you don’t really have an option anymore.