4 replies to this topic

[websitebeginner]

Hello,

If I missed the answer to this question somewhere else, I apologize in advance.

I have two basic stats programs: AWStats and Webalizer.

Webalizer reports substantially more visitors per month than AWStats. I'm not setting world records by any stretch, believe me!

Which one would you trust more?

I like to think conservatively, and pick the smaller number.

But the numbers are so different, I'm thinking of simply adding them and dividing by two to get an average.

They shouldn't be off by a factor of 10, should they?

[Michael Martinez]

I wouldn't trust either of them very much if you haven't tweaked them. Both packages make assumptions that will mangle your data. But both packages will report traffic (on a busy site) that third-party tools written in Javascript often miss. You really have to get down into the configuration files to make sure you're counting the noses you want to count. And even then, there will be some inconsistencies.

[Jill]

Webalizer is crap.

Get Google Analytics. It's free.

[rolf]

I like both for various reasons, but I wouldn't rely on their numbers too heavily. The main things I use them for are quickly looking at trends that I will then go and dig further into with Google Analytics

GA is much more in depth and really useful in lots of ways (and free) but I don't entirely rely on the numbers from that either.

I think either of the server side ones you mention, configured properly, should give reasonably accurate numbers for page views, bandwidth, file types and some other things, but they have a tendency to over count by counting some actions twice or more when they shouldn't - I have an AJAX page that gets called on every shopping page and according to Awstats is my most popular page, although I know it's not and can ignore this snafu

GA on the other hand has a tendency for under-counting some things as not everyone will trigger their Javascript and it can't count most non-page requests.

Bearing all that in mind, I use stats as a general measure of direction - e.g. over a given period of time x tends to = y, therefore action Z needs taking to increase/diminish that trend.

[Michael Martinez]

Heh! Google Analytics is crap, too. It underreports traffic on large sites by a hefty margin, strips out URL parameters, and otherwise denies you valuable data. But it has pretty report layouts and offers you some features you won't get with the older server-side stats packages.

When it comes to Website analytics, trust nothing in particular and don't rely on just one solution and learn to follow trends.