It’s super useful to know what queries bring audiences to your site. You can better understand your customers, make sure that you’re solving their problems, and can know how well your content is performing.
Unfortunately, most query data (from Google organic search) is no longer provided to web analytics programs. Most visits show up as “not provided”. This leads to two issues with query data:
- Many queries don’t appear at all
- The queries that do appear generally show lower than accurate visit counts.
Google Webmaster Tools mostly eliminates the second issue: it lists accurate visit counts for all queries it lists. However, it doesn’t show 100% of queries. The reason is different than the “not provided” problem though. Google Webmaster Tools doesn’t show all queries because it shows only top queries. (How many qualify as “top” vary widely depending on the traffic patterns of the site.)
Keylime Tools can mostly get around that issue. Here’s how it does it:
- GWT reports top queries for each site added, and it considers every subfolder to be a separate site. So the more subfolders you add, the more queries GWT shows.
- But that’s not the whole story. All subfolder data is going to include some duplication with the folder above it.For instance, www.example.com will include some query data for the root of the site, as well as some query data for www.example.com/folder1 and www.example.com/folder2. And www.example.com/folder1 will include some query data that’s also listed in www.example.com/folder1/subfolder2.
Keylime Tools solves that by aggregating all the data together and then removing all of the duplication. The end result is the maximum number of queries possible!
- Keylime Tools then matches the queries that show up in both GA and GWT data. It generates an Excel file that shows you the visit data for each query from each source and that adds up the totals using the higher count.
- After all that, some visit data is still “unknown” (that is, not associated with a query). Keylime Tools estimates how this traffic likely is distributed and calculates “estimated totals” for each query segment.