I’ve been running this blog for almost 2 years now. In the beginning, it was just for me. Then I started getting a little feedback from readers and decided to start tracking who was coming here and from where. At the time I didn’t have the spare cycles to find and install a web counter, so I found CQ Counter and started using this free web service to count the number of visits to the site. CQ works by having a small icon on the site that is automatically downloaded with every hit to the main page. It’s been nice not having to worry about the mechanics of traffic tracking, but I’ve never really trusted the CQ Counter statistics, especially lately. Several times lately I’ve gone to site myself and the CQ icon hasn’t come up, meaning that at least one hit wasn’t being counted.
In any case, this morning I discovered Advanced Web Statistics. AWS is a perl script that parses the access_log and creates a database from all of the wonderful information contained within. I had to make a few minor changes, like changing the logging to ‘combined’, and there were several permissions issues to overcome, but in general, it was an easy setup. I started seeing data immediately, and was able to make sense out of it in short order. And the first thing that became clear is that CQ had not been doing the job I’d hoped for in quite some time.
Let me give you a hint of the difference between what I’m seeing in CQ vs. AWS. In the last 15 months that I’ve been using CQ, it’s recorded nearly 11,000 unique hits and another 3000 other hits. This averages out to some 25-30 hits per day, with some days hitting twice that. Considering how often I post and the subject matter, I didn’t consider that to be a unreasonable number. Now, on to AWS: Between when I installed it this morning and now, AWS shows 118 unique visitors, 150 visits, 585 page hit and a whole lot of other traffic. Given that AWS is getting this information directly from the access_log file, I’m willing to give it a lot more weight that I am the CQ counter. This means CQ has been under-reporting the hits to my page by a factor of 2 or 3, maybe more.
Part of the issue is how they measure hits, I’m sure. I somehow think that most of the robots that hit the site continually aren’t counted by CQ, as well as a number of the other hits that aren’t from people. And any browser that blocks the gif file used by CQ is probably not counted either. But that still leaves a lot of traffic CQ hasn’t been counting. Which is why I’m making a conservative estimate concerning the differences in programs.
I just wish I’d found AWS earlier. Apparently earlier versions of the program had some vulnerabilities, but the current version doesn’t have any known vulnerabilities. I hope. It will be interesting to see some of the trends that show up because of AWS. I’ll keep CQ around for a while, just because of the historical data, but I love the amount of digging I can do in AWS. Heck, you can look at the data yourself if you’re at all curious.