One of the issues we wrestle with in the digital media world is how we analyze traffic on our sites. There are many software packages out there to do this sort of basic data mining and I’m fortunate that I don’t have to worry about which one to pick…I’m very happy with the one we use here at CHNM and so I don’t have to look around at others. What kinds of lessons can the digital historian/educator learn from simple analysis of logs? Here are a couple of examples:
Example #1
This semester, as every semester, I have my students writing in class blogs. Those who have taught more than one section of the same course in a given semester know that there are often significant differences between the two sections in terms of student engagement, the quality of student work, etc. But such things are generally impressions only–my sense for how the class is going and why it is going that way. This semester I find myself in such a situation–two sections of Western Civ–one of which is going better than the other. For a change, I can offer some objective evidence of the difference in engagement between the two classes–courtesy of our server logs.
Each class has 25 students, each receives the same assignments from me each week–they’re even taught in the same classroom on the same day, one after the other. The 25 students (plus me makes 26) in one class have visited the class blog 849 times in the month of September (an average of 32.6 times), but the students in the other section have visited only 559 times (an average of 21.5 visits). That’s a 50% difference in blog visits between the two sections–evidence that my impressions of the level of engagement are not off the mark.
Example #2
Regular readers of this blog know that I’m the Executive Producer (meaning chief minion) of our Hurricane Digital Memory Bank project. One of the many ways we’ve promoted traffic on this site this year has been through Google AdWords. Fortunately for us, Google has a pro bono program for non-profits like CHNM so we don’t have to pay for the clicks. We just ran a report on the results of the AdWords thus far and it tells us that between August 24 and September 26 we got 2,225 visits from Google’s AdWord links. These visits constituted 12.8% of our traffic (17,331 visitors) during that time span. If we had been paying, Google would have charged us $1,746 for those visits, or about $.78 per visit.
AdWords also makes it possible to find out what the most motivating search terms were. By “most motivating” I mean those searches that generated the most clicks to our site. So, for instance, the search “Hurricane Katrina” brought the link to our site to screens 164,388 times (Google calls these “impressions”), generating 1,165 visits to HDMB (1 visit for every 141 times our link showed up on the user’s screen). By contrast, the search “Katrina victims” resulted in 2,603 impressions, but those produced 26 visits (1 in 10).
These are just two simple examples of some of the data I’ve been looking at as I’ve been working on various digital projects. Obviously, much more sophisticated data mining is possible–a good place to start to learn more is Digital History Hacks–but I don’t need more sophisticated data for my purposes.