Practical Ecommerce

Determine Analytics Questions In Advance

Once a web analytics solution is installed, the most common initial use is to validate some past event.  User testing shows that 80 percent of marketers focus on charts that go up and to the right, thus validating the success of the website or an advertising campaign.  I call this “web analytics for entertainment.”  

While this is gratifying and fun to do, it’s not very productive. 
The truth is that good web analysis can give you a competitive advantage across many aspects of your online business, such as search engine optimization, site design, advertising and on-site conversion. But the key is to ask the right questions, in advance, of your web analytics tool and not to let the tool determine what your questions should be.  

Here are the three requirements for forming good questions:

  • The question must be related to your business goals.
  • The question must be answerable by web analytics.
  • The question must have actions related to the possible answers.

Miss any one of these and the magic doesn’t work.  So let’s try one out together:  “Which pay-per-click campaigns shall I cut this week?”

The question is related to a goal of acceptable return on advertising investment, which is clearly a valid business goal, as in A above.  Requirement B is covered, too: A web analytics tool can show the investment return on pay-per-click campaigns. So far, so good.  You can cut or adjust the poorly performing campaigns based on the data from your analytics program, which is a valid action for the possible answers, as in requirement C. With all three of the requirements met, you are ready to go.   

I strongly recommend that you have your questions ready before starting your web analytics sessions.  When you apply this practice, you’ll find your web analytics sessions to be shorter, you’ll get your answers faster, and the answers will apply to your business goals.

Michael Stebbins

Bio   •   RSS Feed


Sign up for our email newsletter

  1. Legacy User January 24, 2008 Reply

    Good advice Michael. As someone who regularly reviews these analytics I can tell you the biggest thing I'm concerned about is the ROI. How much are we spending and what are we earning as a result?

    That's the bottomline I like to start with.

    Next, I want to know what's working and what's not? What are our top 5 or 10 keywords in our PPC campaigns that are working? Those keywords tell me a lot about what more we could do to capitalize on those. I like to see those on a week-to-week basis to see if there are any changes or trends.

    I also like to see what our worst 10 keywords and see if we should remove them or tweak them.

    I like to see these on a week-to-week basis and not be too quick to react. It can take months to see a positive ROI so it's important to keep that in perspective.

    And you're absolutely right, the most important thing is to make sure your analytics are giving you the information you need to make business decisions. Without that, you're just guessing.


    Dennis Sullivan
    "The Publishers' Guru"

    — *Dennis Sullivan*

  2. Legacy User January 24, 2008 Reply

    Another problem we often see is the use of sophisticated web analytics tools as nothing more than another reporting tool. In one recent client case, for example, all kinds of web analytics reports were being looked at on a regular basis — but just as a sort of "nice to know" piece of information.

    Using metrics to understand how you're doing is nice, but it's missing the point. The real power in measurement comes from using the numbers and ratios to significantly improve your performance. To do that, you need to get beyond the "nice-to-know" factor and ask powerful diagnostic questions like:

    Is this level of performance good or bad?
    Why is this metric at this level of performance?
    What is causing this metric to change over time?
    How can I profitably influence this metric in a sustainable way?

    Once you understand the root-causes behind the metrics and the various elements that can have an effect on the numbers themselves, you have the keys to identifying profitable, systemic improvements.

    Rafe VanDenBerg
    Principal Consultant

    — *Rafe VanDenBerg*