Posted by 4 Comments Others

Last month, Google significantly revamped the “Top Search Queries” section of Webmaster Tools which put some very interesting data at the fingertips of webmasters and search marketers. It shows the keywords and keyphrases for which a site has been returned in Google search, the number of impressions (the number of times a SERP with a result for your site is displayed), the number of clicks through to your site, and the percentage clickthrough rate (impressions divided by clicks). This data can also be segmented by time period.

The data was actually of limited use and easy to misinterpret, because it was difficult to see how the site ranked for each keyphrase (and ranking drastically effects both Impressions and CTR). It appears that Google listened to feedback regarding this, as on Monday they added an additional column which shows the average position of the site for a keyphrase over the selected period. As before, this can further be broken down to see performance at different positions in search if a page’s ranking has changed over the selected time period:

google webmaster tools data

However it appears there are some serious bugs with this average rank data, particularly for smaller sites. When ordering keyphrases by position, there are many clearly erroneous keyphrases included reporting a high (first page) rank for the site.

One of our new SEO clients gave permission for us to use data for their site to create this blog article.

A significant proportion of the keyphrases reporting a position of “1.0”, “2.0” or “3.0” are erroneous.

The search for “email list cleaner” does not seem even remotely topically related to our client’s site and we found no pages for our client’s site returned for this search:

Similarly with “commercial bathroom cleaning” – this is not a service advertised by our client and there are no pages from the site returned for this search.

Finally, the supposed position one for “nude house cleaners” is particularly unusual as the site does not contain the word “nude” in content (or in backlinks for that matter)

It seems that Google clearly have a few gremlins in the works at the moment. This is not the first time we have found GWT data to be unreliable but previously this has only really been apparent via comparison with analytics data.

The data for average position appears to be somewhat more reliable for sites seeing a lot of traffic, though in these cases useful data can easily be obscured by literally hundreds of obscure 1.0, 2.0 and 3.0 rank keyphrases with less than ten impressions. It would certainly be of use to be able to filter this data before exporting to a spreadsheet.