What web stats package do you have? Do you trust it? My web hosting package comes with three pre-installed: Awstats, Webalizer and Analog. I was always pretty content to use Awstats as it produced comprehensive, easy-to-read statistics.
However, it wasn’t until I happened to look at another of my web stats packages that I noticed some serious discrepancies between the three.
Take a look at these figures for October, 2004:
| Awstats | Webalizer | Analog | |
|---|---|---|---|
| Visits | 10,017 | 15,521 | N/A |
| Page Views | 32,113 | 36,053 | 20,912 |
| Bandwidth | 530.11 MB | 710.67 MB | 692.69 MB |
I’m pretty surprised by these differences, especially when all three packages are running on the same server. I can understand how things like visits and page views might be calculated slightly differently, but bandwidth?
Checking the HTTP bandwidth using the tool that my web host provides in control panel, I see that according to it, my bandwidth usage for October was 765.36 MB (of course, it would be the highest).
So, as they are higher, perhaps I should go off Webalizer stats rather than Awstats. But what if a potential advertiser who wants to put a link on my site asks me for my traffic stats? Which ones do I provide?
Should conscience dictate that I provide an average of them, or the lowest, to be safe?
I wonder if anyone else has noticed anything similar?
Hmm, weird. My stats don’t match either. I also use Awstats most of the time, since it’s easiest to read and has a lot of good information. What’s weird is that it looks like Webalizer matches my cpanel bandwidth amount, Awstats is somewhat less, and Analog is way over. Beats me.
It’s a mystery to me too. Which is one of the reasons I never trust web traffic figures that people throw out, other than in the most general terms. There are just too many unknowns to take into account.
At least services like Alexa help to level the playing field somewhat.