Recently I participated in a web analytic competition. This is the second time I have participated in this competition. Yet during both experiences I learned a lot about web analytics.
The first time I participated I wanted to organize a diverse team so that we could each look at the problem from a different perspective. In that team we had a computer science major, a supply chain major, and a information systems major. This team worked well together, yet we focused on aspects of web analytics that were not esteemed very highly by the judges. Although we did some good research and functioned as a well designed team, our main focus point was not what the judges were interested in.
The second time I participated in the competition I played more of a team member role, rather than a leadership role. My team had some good objectives and strategies in this competition in that we let the data guide our course. We identified several of the top exit pages and then dove into the data to determine why they are the top exit pages. The information we found highlighted some major problems that the company has with both its web site design, and its persistence of bookmarks.
One lead we pursued dealt with the clearance items. We found that a majority of the people who go to their clearance section leave nearly immediately. we were going to recommend that they change that page. Yet when we were gathering up our last bit of evidence we found that the site changed that page to what we were going to recommend. This was somewhat of a blow to our research because on the one hand we had spent a lot of time and effort making this recommendation which was now obsolete. Yet on the other hand our team decided to turn this tragedy into a success, and we did that by using the little bit of data we had for the last few days to prove that idea was successful.
Originally I wanted to research into correlation between user reviews and orders. I had formed a hypothesis that if a product has more user reviews, it would be more likely to sell and get a higher conversion rate. Yet after running into problem after problem, our team leader suggested that I work on one of the other ideas that we had developed. In some ways I was glad to let that go, but in other ways I wished I would of found some data backing the claim. Yet in the end it was probably wise to drop that and move on to something else.
Overall we found quite a bit of good information, and most of the information we had collected we were able to make solid recommendations. I think we may have done better had we tied financial information directly into the recommendations that we made in order to prove to the judges that the data we found was quite valid and relevant.
Even though my team did not win, this experience has helped me gain a greater appreciation of taking data and turning it into an executive summary report. Lesson learned: Sometimes the path to the data and recommendation for the data is not as important as the results from the recommendations and its expected benefit.
The only question I have remaining actually deals with the hypothesis I formed in the beginning. I would like to know definitively if there is any positive or negative correlation between reviews and orders on the site we examined. I think if I get a chance to talk to one of the professionals who do analytics on that site I will ask them how to go about finding out that information.
That is my take on the web analytic competition.