Monday, July 26, 2010

Search Engine Optimization and LSI

Many people are confused by search engine optimization and LSI. What is LSI and what has it to do with SEO? In fact nothing. That’s right – nothing. Zilch. Absolutely nothing whatsoever. So why all the fuss? Through ignorance and possible subterfuge.

LSI stands for Latent Semantic Indexing, a term that has no meaning whatsoever. There is a term called ‘latent semantic analysis’ that is the analysis of the hidden meaning of text and its explanation by means of an analysis of the way that other words in the passage are used. The code for each letter in a word is analyzed, and the word thereby identified. There are certain known juxtapositions of certain words that provide a meaning to these words, although all may not be what it seems.

For example, is I ‘bought an apple’, did I buy a computer or a piece of fruit? That only becomes apparent when the rest of the semantics of the passage is analyzed. The word semantics means the meaning of words in the way that they are used. For example: “I bought a dog lead for my German shepherd” indicates that the German shepherd is a dog, and not a Teutonic gentleman looking after sheep. It is the rest of the semantics in the text that makes it clear what the term means.

Take another example: “spiders crawl the web looking for ….”. What does this actually mean? Is the passage referring to arachnids seeking flies or search engine spiders looking for specific keywords? Only the remainder of the text will make this clear. What does the term ‘the history of locks’ refer to? If you came across a website with that title, what topic do you think it would cover? The history of canal locks? The history of door locks? Only an analysis of the rest of the text in the page would tell you that, and that is what LSI is. It is not ‘latent semantic indexing’ but ‘latent semantic analysis’.

What the term relates to is the ability of search engine algorithms to determine the relevance of your content to the search term used by the search engine user. Someone searching for the “history of English canal locks” dos not want to be presented with a website entitled “the history of locks” that is an authority on jail locks. The algorithm analyses the text on the page and decides the topic from the word used.

It also ensures that the content of the page is relevant to the keyword, and does not only consist of a continuous repetition of the keyword, as many sites did prior to Google’s algorithm change. In fact, this type of analysis put paid to the software that generated thousands upon thousands of websites and page content from a single template containing no content at all to speak of but a form of words into which any keywords could be entered. Thus, “information on dogs is available all over the internet, and dogs are a popular topic with many websites providing information on dogs”.

Any keyword you can think of could be use in place of ‘dogs’ and the content provides no information whatsoever. That is typical of the reason why Google introduced the LSI algorithm, even though the term is meaningless to the webmaster. Apart from the fact it should really be termed latent semantic analysis, LSA, and not LSI, it is an analytical technique, and analytical techniques cannot be used by anybody to improve their website. You cannot make a website LSI compliant, but you can write a web page with good content that is relevant to your page title. You can include varied information and good English, rather than endless repetitions of the keyword, and so allow the search engine spiders to understand what your web page is about.

No comments:

Post a Comment