Following on from Emily’s post about Google’s webmaster tools dishing yet more info that may feed our SEO needs I thought I’d throw in some analysis.

After some rooting around it has become apparent that the “significant” words that the Significance list presents are pretty shocking. Here’s why…

Across a number of sites I found instances where tag names are listed in GWT as being significant keywords. These are not chance occurrences where a word like “div”, “span” or “class” are mentioned in any title, meta or content.

It has even given “blogger” some significance when that site uses the Blogger platform for posts- again this word only appears in meta “generator” tags or is part of link that calls java from Blogger itself.

Here are some samples:

…all of this leads me to believe it is just scanning mark-up and reporting all instances of a word, and doesn’t appear to take into account where the words appear.

The problem is that this data can be interpreted as an exact measure of the keyword weighting/relevance of a site, which is misleading, and annoying.

Posted in Blog | Comments Off