• http://christianbook.com Nate S.

    Hi George,

    thought I’d mention that I had difficulty getting meaningful data from the first couple steps in this process. even with my date range set at a week Google severely sampled my data when I added the landing page dimension, giving me confidence range of +/- 75% or more for almost all the keywords, which is pretty much useless (assuming I’m understanding their confidence rating correctly).

    narrowing the report down to a day gave me clean data, but then there weren’t enough visits/keyword to make the conversion rates meaningful. Plus, making judgements about a keyword’s value on a day’s data is probably not wise anyway.

    In the end this seems like a good tool for smaller, medium sized sites, at least if you’re using GA.


  • http://www.evisionsem.com/ George Aspland

    Hi Nate,

    Just back from Pubcon and seeing this now.

    In most of these steps you should look at data for a number of months in order to get enough data to be meaningful, ideally 3 to 6 months. I usually look at about 5-6 months of data unless I know there has been a fair amount of changes that affected the site URL’s etc