I wonder if anyone else sees this, or can explain it? Analytics shows about 2/3 the number of unique visitors as Awstats, which for a while I was blaming on people with javascript disabled. That doesn't really make sense, as Awstats tells me how many people browse w/o client scripting, and it's a lot less than 33 %.

A month ago, someone put a link to my Seattle photos on a popular blog. Awstats says I've had about 45 visitors from that particular domain over two calendar months, while Analytics says 150. If I'm chalking the difference up to js, Google should be underreporting, not the other way around.

I know the absolute numbers don't matter nearly as much as trends, whether the work I'm doing is paying off, and things like that. But suddenly I feel like I'm on shaky ground with my ability to figure this stuff out, if two stats packages are telling me two different things and I can't figure out why. Any thoughts?

No idea, bro. Blogger and IIS both don't use awstats. I keep hearing about that package, not really sure why it's so popular. Still, I'd go with your server logs if there's any doubt.

I find large discrepancies too, between analytics, awstats and webalizer, also when I turn off some google ads, adsense still shows impressions for those ad channels a few days, I trust the stats based on the server logs more than google, but they interpret the raw data a little differently to each other too

Forrest all stats programs will show somewhat different numbers depending on how they collect data. Some of it could be the javascript disable, but not all. awstats doesn't clear out as much automated traffic from spiders/robots which I think is what accounts for most of the difference. awstats catches Googlebot and Slurp and the other biggies, but if you or I sent out a robot into the wild it likely is being recorded by awstats as a real person.

Think spammers who are sending automated traffic.

For me Analytics shows about half the numbers that awstats shows. My view is always to trust the numbers that paint the poorer picture since it makes me work harder.

Also instead of always looking at hard numbers look more at the trends. Both should still be showing increases and decreases the same.

Originally Posted by vangogh

Forrest all stats programs will show somewhat different numbers depending on how they collect data. Some of it could be the javascript disable, but not all. awstats doesn't clear out as much automated traffic from spiders/robots which I think is what accounts for most of the difference. awstats catches Googlebot and Slurp and the other biggies, but if you or I sent out a robot into the wild it likely is being recorded by awstats as a real person.

That makes sense. It's not the answer I was looking for, but it's probably more likely. Although awstats isn't that bad about plucking out the spiders; they look at the UA, hits against robots.txt, and possibly other criteria, and identify about 15 different spiders that crawl my site. Maybe half of them are Gooblebot, Slurp, and the rest, while the other half show up as "Unknown Robot."

But then a lot of scrapers probably cloak themselves as IE, and don't care at all what's in my exclusion file.

Originally Posted by vangogh

Also instead of always looking at hard numbers look more at the trends. Both should still be showing increases and decreases the same.

Yeah, that's what's truly important, and if I pull back to a week or more to smooth out the daily roller coaster, things are generally moving upwards. My background in IT says if I can't trust or understand the numbers that make up the trend I should be wary, but they both show the same general trend, and the way you explained it probably does make sense.

Originally Posted by vangogh

Forrest all stats programs will show somewhat different numbers depending on how they collect data. Some of it could be the javascript disable, but not all. awstats doesn't clear out as much automated traffic from spiders/robots which I think is what accounts for most of the difference. awstats catches Googlebot and Slurp and the other biggies, but if you or I sent out a robot into the wild it likely is being recorded by awstats as a real person.

Think spammers who are sending automated traffic.

For me Analytics shows about half the numbers that awstats shows. My view is always to trust the numbers that paint the poorer picture since it makes me work harder.

Also instead of always looking at hard numbers look more at the trends. Both should still be showing increases and decreases the same.

That's exactly right van gogh.

Forest, In my experience the discrepancy of 2/3 is pretty standard between AwStats and client side stats. Either way, don't worry so much about the actual number of visitors, look at the trends, up or down.

Nice to see you here WebGeek. Yep actual numbers aren't the thing to look at. Ultimately what we want to know are things like conversions and conversion rates and where people are leaving our sites so we can improve those pages.

The actual numbers just aren't going to match up so look at the trends. If things are improving great and if not then see if you can understand why so you can improve things.

So both this services are approximately the same? Or I have missed somthing?

Originally Posted by vangogh

Nice to see you here WebGeek. Yep actual numbers aren't the thing to look at. Ultimately what we want to know are things like conversions and conversion rates and where people are leaving our sites so we can improve those pages.

The actual numbers just aren't going to match up so look at the trends. If things are improving great and if not then see if you can understand why so you can improve things.

Thanks! Nice to see you here too! Well said.

Originally Posted by HowLong

So both this services are approximately the same? Or I have missed somthing?

They're not really the same...AwStats is free and comes on most servers with cPanel. It's a server-side stat program, which measures things differently that client-side stats, which run in your browser through JavaScript. They offer some different perspectives on your traffic, and I find it's good to us a combo to get a better overall picture.

As a cross check, download FunnelWeb and run that against your logs.

You can also download the FunnelWeb Enhancement to identify more bots & browsers or add more when you need to.