LSI content is becoming an increasingly important factor in search engine optimization (SEO). LSI stands for “latent semantic indexing” and it is a method of analyzing the concepts connected with a document or a collection of documents.
Optimizing for LSI is not replacing traditional SEO, but instead, it works together with the older techniques. LSI content will likely become more important in the future as the major search engines try to deliver web content that provides closer matches for user queries.
What is latent semantic indexing?
The word “semantics” refers to the meaning of words and LSI attempts to discover the concepts associated with web pages by analyzing how words work in combination with other words. Many terms will have different meanings depending on the context of the document.
What LSI attempts to do is to map out the relationships between words in order to help decipher the meaning of the text. While the algorithms used to rank web pages cannot “think” like humans, they can compare the words used in a particular document along with categorizing the structure of that document.
Additionally, search ranking algorithms can analyze collections of documents to help in better determining the related concepts of the collection.
How LSI can change search engine result pages (SERPs)
One of the tangible results of latent semantic indexing is that it can have marked impact in many cases on search engine listings.
Previously, a user query would only return pages that contained all the specific keywords in the query. However, with LSI that no longer is the case. While in most instances, one would find all the included keywords, there can also be results that might miss a keyword or even all the keywords in the query.
Such results would be more common with queries that contain multiple keywords that the search algorithm recognizes as LSI content. The algorithm can make a probabilistic guess on the concepts that the user wants, and then look for pages that best match those concepts even if they do not contain all the query keywords.
For example, if a user searches using the query “fast Italian cars,” a specific page may match the general concept of the query precisely. However, that page may not use any of the keywords contained in the query. For example, it might not use the words “car” or “cars,” but instead, replaces them with other words like “automobile,” “vehicle” and “roadster.” Words like “powerful” and “speedy” might appear instead of “fast.” Rather than the keyword “Italian,” the text would simply mention popular Italian sports car brands like Ferrari and Lamborghini.
Why good writing is important for LSI content
One of important features of latent semantic indexing is its ability to recognize what types of words to expect when dealing with specific concepts. In this way, it can detect, to some extent, the “meaning” of a particular document or collection of documents.
One of the main techniques of traditional SEO is to place a certain density range of keywords and related keywords throughout the text and particularly in certain parts of the text like the first sentence and the last paragraph.
While such techniques may still be important, they are not sufficient to give a document or collection of documents a high semantic score. That is because latent semantic indexing recognizes the types of words used in high quality documents for specific concepts.
An SEO specialist is unlikely to be able to guess these word combinations or to figure them out well using the Google Keyword Tool or other applications. These words along with the desired ratios tend to come out when a knowledgeable writer creates a well-written document.
Site structure and links
LSI does not just analyze single web pages, as was mostly the case with the old algorithms. The methodology places more emphasis on analyzing collections of documents.
Just how search ranking algorithms determine what constitutes a document collection is an important question. Many SEO specialists use the “silo” method to create a file structure on the site server that organizes web pages in “collections.”
For example, on a car site there may be a particular folder or silo that deals only with Italian sports cars. All the important topics on that car site will have their own silos, i.e., British cars, American cars, Japanese cars, etc.
In this way, the SEO team hopes the algorithm will recognize the structure and analyze the documents in the folders as collections. Of course, it is very important to ensure that all web pages and other documents in these folders are directly on topic.
Another way that algorithms may detect document collections is through links. For this reason, it is important to include links to other documents that deal with the same specific sub-topic. If you do have links that diverge from the particular focus of the collection, it is best to create these as no-follow links.
How to create LSI content
Most SEO experts use applications like the Google Keyword Tool to find related semantic terms. Obviously, it would seem best to use keywords that people search for commonly.
At the same time, though, it is good not to overuse related keywords in a way that is unnatural. The algorithms pay close attention to densities and positioning of words and, in most cases, it will be able to detect excessive keyword stuffing.
The best idea is to create content that reads well and covers the important concepts related to the targeted keyword.
Remember that search engines are constantly trying to provide higher quality results to their users. They do not want to deliver results from spammy, poorly written pages. LSI is part of the effort to analyze content that reads like other highly regarded documents for the same concepts.
One thing to ask yourself when analyzing LSI content is: does it explain the topic well or answer a specific question? If the content leaves you better informed after reading it, then it will likely past the LSI test.
However, if the content leaves you confused or unsatisfied, then it probably will need more work.