Last night I gave a short talk on web traffic analysis to a data mining class taught by my friend, Dr. Aleks Jakulin, at Columbia University. Sharing the podium with Dr. Hilary Mason from Bit.ly and Blaz Fortuna from the Josef Stefan Institute in Slovenia, I decided to present a primer on the web analytics industry and leave the science to the experts.

It is a rare 400 level course with no prerequisites, so some students come from a statistics background while others study mathematics or business. What could I say that would be useful to all three disciplines? The main point I wanted to get across was simply the importance of acting upon your insights, regardless of whether you pursue such interests for academic or financial reasons.

I still find this to be the largest problem within web analytics today. At HitTail, we championed the idea of actionable analytics in 2006 (and even before that in Connors’ client offerings). Now, suddenly other companies keep using that word. I do not think it means what they think it means.

Many researchers in the space seem keen on marrying predictive modeling to analytics. Of course that is a cool concept, but it is not always actionable. What do you do with the knowledge that traffic patterns will adhere to the same seasonality as before? Perhaps you can increase ad spend during the same time every year. That is something you should already be receiving from your market research.

Meanwhile, the patent-pending process we created in HitTail was as much about creating a workflow as about its underlying SEO analysis algorithm. How can you begin to take advantage of the long tail when it amounts to a daunting number of keywords in the thousands? Identifying suggested terms to focus on is a key first step, but is not enough on its own to affect change.

An analyst also needs an ability to manage that massive amount of data and, critically, be able to show some progress. Therefore, the HitTail user interface was designed with two minor revolutions in the field of web analytics: an ability to check off items from a To Do list and filter out irrelvant data without having to constantly redo queries.

Imagine going into an enterprise application such as Omniture, Webtrends, or Google Analytics and having the audacity to delete data. People would have heart attacks. Yet it is precisely this overflow of information that can partly lead to analysis paralysis (especially in SEO).

The concept of paring down lists can literally guide you on a path to success. How you pare down lists is up to you. For its part, HitTail provides options to directly add your action items (in this case, long tail keywords) to your ad campaigns or blog about them. The suggestion algorithm helps you be strategic so you can be free to just complete the task at hand and move on to the next idea. That is how you make things happen.

I believe a good analyst is defined by his or her ability to become a catalyst for change. I have only limited interest in insights for their own sake. I love to learn and knowledge is power, but actions still speak louder than words. The cynic in me, for instance, wonders why schools bother teaching history when students never seem to learn from it! Is it the fault of the teacher or the student in that case? Perhaps it is the parent, for not showing how the past can relate to modern life.

The change I am advocating does not just have to be for a social cause, though. This concept fits just as well into building business cases or developing marketing campaigns.

Analyze and then catalyze.

That idea may seem obvious to some, but certainly not all. I want to reach anyone who still isolates themselves in silos of data. Before digging deep into a research project, make sure you are looking to find results that will have a practical application. Come down from those ivory towers and make a difference!