Until recently, search engines did not take into consideration the semantics i.e. the synonyms and related keywords. Latent Semantic Indexing (LSI) adds a new approach to keywords, because now a document does not even need to include the search term to rank in the SERPs as long as keywords very related to the search term exist in the document.
“Latent semantic indexing adds an important step to the document indexing process. In addition to recording which keywords a document contains, the method examines the document collection as a whole, to see which other documents contain some of those same words. LSI considers documents that have many words in common to be semantically close, and ones with few words in common to be semantically distant.” (Yu, Cuadrado, Ceglowski, Payne 2002.)
Based on the findings of the webmaster community, Google is using the LSI element in its algorithm, but at this time, there is no confirmation about the other search engines (WebmasterWorld.com b). Words that Google considers related can be found with a tool provided by goRank (http://www.gorank.com/seotools/ontology/).
For optimizers this means that they should use related keywords in their content and links when they are optimizing for a specific keyword, instead of repeating that specific keyword. For example, it may be helpful if a page that is being optimized for “nike tennis shoes” also included terms like “nike tennis sneakers” and “nike tennis footwear”.