Why are there sometimes very few terms in a semantic index?

In most cases, Wordmetrics will return a semantic index with 50 terms. But the index may have fewer terms depending on which search term is used. Generally speaking, a smaller number of semantic index terms is an indication that the search term is too broad.

Because broad search terms may return a wide variety of contextually unrelated search results, a statistical analysis of those pages will return fewer statistical similarities. When topics are contextually unrelated (or have a limited relationship) not only will a the list be smaller, but the quality of the semantic index may degrade.

For example: Suppose a user enters a New Search Term with multiple possible meanings, like "big fan". (You can enter this term in Google yourself to visualize the problem). The search results will be taxonomically unrelated because search intent is unclear. The multiple definitions of the term "fan" end up producing a contextually dissimilar corpora of page results, and the statistical similarity between those pages is limited.

In the event that your semantic index is too short to be usable, Wordmetrics will automatically credit your account for that optimization. Try again with a more specific search term.

Did this knowledgebase entry answer your question?