Latent semantic indexing

Latent semantic indexing is the process by which Google and other search engines infer what a pages is about based on words other than the page's official keywords. In other words, search engines try to work out what a page is about by considering the latent semantic content of the text. The philosophy of this approach is that you could write a fantastic article about oranges without mentioning that keyword more than once or twice. Latent semantic indexing allows search engines to give that article the credit that it would not get if they only counted the number of occurrences of the keyword; ultimately, it is about improving the search results for the users.

For example, if the title of an article is "Weight" and the main body of the text keeps mentioning weight, search engines have to work out whether the article is about physics, dieting, or something else. To do this they look at the whole text and search for clues that will lead to disambiguation. In this case, if the article also mentions Isaac Newton and gravitational forces, search engines will decide that the article is about weight in the scientific sense of the word. The article will therefore be eligible to appear in search engine queries such as "weight mass and gravity." Conversely, if the article mentions calories, obesity and exercise, search engines will decide that the article is about weight in the health sense of the word.

How does understanding latent semantic indexing help us in SEO? It can help us write excellent copy, which is the heart and soul of SEO. Specifically, latent semantic indexing points us to the old mantra of SEO: write excellent copy. Do not use the same keywords again and again, as this is boring to your users, and search engines can tell that the copy is of low quality. Instead you should use a variety of synonyms of your keywords in addition to the keywords themselves, and you should make the article as comprehensive and deep as possible.

With latent semantic indexing, search engines are able to decide not just what the keywords of your articles are, but also how genuinely useful and authoritative the article is. If you're writing an article about losing weight, mixing many instances of the keywords “lose weight” with a mass of trashy text will not win the approval of Google and other search engines. Conversely, if you make good use of the keywords that are related, such as calories, exercise, cardiovascular health, triglycerides, metabolic rate, micronutrients, micronutrients, etc, search engines are likely to deem your article useful and authoritative, giving it a higher ranking in the search engine results pages.

As you can see, there is a lot more to search engines than blindly accepting the official keywords you feed them. Google and other search engines have become very smart at inferring the objective quality and topic of a page of text, in their continuing effort to weed out run-of-the-mill scraper sites that pollute the Internet with their trashy, pirated content.