Posted by Lee Jackson on Google+ & filed under SEO Copywriting.

One of the most important aspects of SEO is thinking carefully about Latent Semantic Indexing, or LSI. But unfortunately if there’s one aspect of SEO that is constantly misunderstood, it’s LSI. Just this week I’ve seen three in depth blog posts by people who really should know better, informing people confidently to use plenty of synonyms in order to boost their LSI rating.

This is wrong on two counts. Firstly there is no such thing as an LSI ‘rating’, and secondly, if you start using synonyms then Google will almost certainly thump your content into the back of beyond where no one will ever find it. Let’s understand why.

First of all, let’s understand why Google decided to introduce LSI algorithms, and then we can understand a little better what LSI is, by seeing what it’s for. The problem Google was facing was that a great deal of content was (and still is) being churned out that was utterly hopeless. It was either stuffed full of keywords, or stuffed full of waffle, interrupted occasionally by another keyword being sneezed onto the page. Google wanted a way of identifying content which was subject relevant, and likely to be more reliable.

Latent Semantic Indexing is a mathematical way of analysing a document to see how likely it is to be relevant to the subject, and how reliable it’s likely to be. This is done by cross referencing the vocabulary used against known trusted sources of information on the same subject.

So an article about china tea pots is likely to include words such as drink, tea, cup, pour, brew, tea leaves and so on. The broader the vocabulary, and the more words used which relate to the subject, the more likely it is that the content is highly relevant to the subject.

However, the use of synonyms tends to suggest that the author is either trying to force keywords in without Google noticing, or that the content has been computer generated. So if the word ‘cup’ is used, as well as the word ‘mug’, ‘teacup’, ‘vessel’, ‘goblet’ and cupful’ Google will consider the content to be keyword stuffed (since all those words mean the same thing) or that it’s computer generated, and therefore not to be trusted.

So when creating optimised content, be aware of what LSI really is, and what it’s for. When a writer creates content for people he or she tends to stick with calling an item by the same word each time. When writing informative content they’ll naturally tend to use a wide selection of subject relevant words.

So try not to be vague by using words such as ‘it’, or ‘thing’, try to be consistent, using the same word each time when referring to the same item or concept, and try to include a broad range of subject relevant vocabulary.

Need help with your SEO copywriting? Get in touch and we’ll be happy to help!

If you enjoyed this post, make sure you subscribe to my RSS feed!

  • Lisa Blackler

    Thank you for this clear, simple explanation of a concept which I firmly believe in. If google’s job is to return relevant pages for any search, then it follows that you should do better if you content is truly relevant and useful to the end user. I get frustrated at people who see google as their ultimate target rather than the site visitor.