At work our Pubs director and Dean and others were thinking about whether our monthly newsletter's really worth the time it takes staff to create and send out. The newsletter covers our various programs, application deadlines, links to specific programs and courses, etc.

They therefore liked my suggestion to look at the log files for our web server and correlate hits recorded there to our pages mentioned in the newsletter. We know when we send the newsletter so if we see a spike in the hits to the web pages it links, then it would be fair to conclude that the extra traffic is a result of the newsletter's link.

To that end, I wrote a short quick Perl script to tell us these things. It knows all the linked pages, counts all the hits in the log files to those pages by day of month. It then generates a list of those pages, a graph of how many times they were hit by day, and shows a total count for each newsletter-reference page.

Now we have data. And something more valid than the "this I believe" approach being used earlier.