Urchin / Webalizer Statistics Show Different Pageviews - Other Statistics

[achabot]

I've got a good question for you guys out there, do you have any idea how this could happen?

As viewed on December 10, 2004:

Statistics for domain.com are as follows:

Webalizer - December 9, 2004 Total pageviews = 83
Urchin - December 9, 2004 Total pageviews = 282

In Summary:
Same domain name, same day, same log file being analyzed. Different Numbers.

On December 8, there were 264 on Urchin and 58 on Webalizer. What gives?

How do I explain this to my customers?

If I'm missing something here, please bring me up to date.

Any help greatly appreciated, :Please:

[jmbeach]

you know what? I asked tech support a couple of months ago why my HSphere CP was showing over a gig of traffic on one site, while webalizer was showing only 600megs. The extra 400 certainly wasn't email, and it makes me nervous thinking the either the stats are completely wrong, or that HSphere is overcounting somewhere...

I know it's not the same issue you're having, but it goes to show that sometimes numbers can be wrong, or miscalculated...

[dpyers]

I don't think that any two stat packages count things the same way - e.g. pageviews - is the count incremented when the page starts to load or when it finishes? Is a refresh a new view?, etc.

The only one that really counts is hspere as the billing is based upon that.

[cambodia]

so anybody in psoft forum found any thread complain about this ? or that software is have bug ?

[DVHost]

Originally Posted by dpyers

I don't think that any two stat packages count things the same way - e.g. pageviews - is the count incremented when the page starts to load or when it finishes? Is a refresh a new view?, etc.

The only one that really counts is hspere as the billing is based upon that.

That's disconcerting.

[dpyers]

I like awstats. At least it breaks out search engine/bot views of pages from "regular" views and I find than clients tend to understand it without to much handholding.

I do have a standard speech I give them about how stats are useful for general trends only, and should always be taken with a large grain of salt.

The alt.www.webmaster faq used to have a good link for explaining the futility of stats but it appears to be broken.

[cambodia]

Originally Posted by dpyers

I like awstats. At least it breaks out search engine/bot views of pages from "regular" views and I find than clients tend to understand it without to much handholding.

I do have a standard speech I give them about how stats are useful for general trends only, and should always be taken with a large grain of salt.

The alt.www.webmaster faq used to have a good link for explaining the futility of stats but it appears to be broken.

but that awstats log is bigger than webalizer , but webalizer is stuck now no new developing look like that developer is stop do their work

[PeterD]

The issue I have with this is slightly different. Webalizer and Urchin do show slightly varying results which is just the result of different methods they use to determine what a hit is etc.

All well and good. Where I get confused is when, over the course of the month, Webalizer's hit count increases as one would expect. But once the month rolls over to the next, the Webalizer stats for the old month drop WAY down.

For example, in Feb my service domain got just under 50,000 hits according to Urchin. Weblizer sort of agreed with Urchin DURING Feb, but now that we're into March, Webalizer reports February's hits at under 2,000. That's about a 96 percent discrepancy. So where'd my hits go!?

This isn't a problem if you have faith in Urchin and are prepared to pay the monthly cost for it (which isn't a big deal). But in my reseller plans, I only offer Urchin to my top-end customers; the middle order gets Webalizer. So I'm waiting for the day to come when a customer picks up on the Weirdalizer's way of accountig for things.

[doublez]

Wow. I'm glad I did a search before posting. I have this exact same issue. When I first got Urchin working... Webalizer stats were a lot higher than Urchin. Now Urchin is much higher than Webalizer. June Webalizer hits: 800K - Urchin hits: 1.25 Million

That is a serious difference regardless of the methods used to count hits.

Is it possible that Webalizer has an issue with accuracy once the hits/sessions get too high?

Or could it be the other way around? Which set of stats do I trust? Webalizer or Urchin?

Thanks,
Ron Dunn Jr

[dpyers]

I tend to think Urchin is the more accurate, but that's just a hunch. No real basis for the opinion.

You might want to take a look at the logs for both for a partcular session and see what the differences are. Could be something off the wall like domain forwarding, url masking/rewriting, ssl, etc., that's skewing the stats for one or the other.

[doublez]

Originally Posted by dpyers

I tend to think Urchin is the more accurate, but that's just a hunch. No real basis for the opinion.

You might want to take a look at the logs for both for a partcular session and see what the differences are. Could be something off the wall like domain forwarding, url masking/rewriting, ssl, etc., that's skewing the stats for one or the other.

How do I go about looking at the logs?

- Ron

[mresell]

You can view them via cp or download via ftp.