-file caching: if a returning visitor has file caching enabled on his web browser (usually it is the case) the request for the file he already viewed will not be sent to the web server, and he will view a local copy, so awstat will not detect multiple visits some times.
-Page tagging relies on the client (visitor) browser voluntarily providing the analytical information requested. Some browsers disable so me data collection for security reasons.
The main advantages of logfile analysis are the fact that the data is already logged by your webserver. every transaction is recorded by the webserver reguardless of the visitors browser. and this data is on your server directly in standard format so you don't rely on a third party server. Anoher advantage is the fact that logfiles provides information about visits from search engine spiders and bots, failed requests which is really important if you want to optimize your website.
i guess you can continue using both in complementary aproach!
hope this reply was useful
I recently installed the google analytics code and also run awstats. The difference in reported traffic is huge. Google analytics reports about 1/3 the traffic that awstats does. Google analytics breaks down the traffic in useful ways that awstats does not, but the overall amount of traffic reported is way off.
I have also recently installed G analytics in selected pages on my site. Compared to the log based stats tools I use, I also experience a huge difference in the session numbers.
Ahh, yes, also..........
We just installed awstats and will be adding google analytics later.
If the server and connections are slow... the visitor bails out before page loads.... is there any record of this in awstats?
Spiders and bots cannot execute Google Analytics code -- this alone excludes a lot of traffic especially if you have Adsense on your pages.
Google Analytics script sets a persistent cookie to track user's navigation and visits regardless of the IP address -- this is not possible by digging the server logs alone.
I think Google Analytics gives you more accurate stats. The down side is that there's a noticeable latency every once in a while when loading pages seeded with Google Analytics tracking code.
|Spiders and bots cannot execute Google Analytics code -- this alone excludes a lot of traffic especially if you have Adsense on your pages. |
|Google Analytics script sets a persistent cookie to track user's navigation and visits regardless of the IP address -- this is not possible by digging the server logs alone. |
|The down side is that there's a noticeable latency every once in a while when loading pages seeded with Google Analytics tracking code |
Several people on this site have mentioned that, for this reason, they do not put the code in the head, but at the foot of the page.
|If the server and connections are slow... the visitor bails out before page loads.... is there any record of this in awstats? |
If the request at least made it to the server, then it will be logged right away by the server and AWStats will report it even if the user bailed out early.
One of the fields in your raw server logs is the number of bytes transferred. If this is less than the file size, it was aborted for some reason.
Another field is the result code, which is normally 200 (success). If only a partial page is sent, the result is 206 (partial), but I'm not clear exactly what circumstances trigger a 206. It might be when the user presses ESC and their browser explicitly terminates the transaction.
Or to put it another way, I think it is possible for a full page to be sent out by your server to a user who has already bailed out without cancelling, and you likely have no indication that's what happened. ... [Edit: Come to think of it, in this situation your server will log the request but the Google Analytics code won't run, which could provide a useful statistic, but only if it happens many many times.]
[edited by: SteveWh at 12:26 pm (utc) on Feb. 14, 2007]
1. Are you looking at "Hits" in awstats rather than "Pages"? "Hits" count all requests to the server, including images. Be sure to look at "Pages" - the number of page views - when you compare with analytics.
2. Are you sure that you have the (full) analytics script in all of your pages? Do you for example have an admin section where you don't use analytics?
3. Did you place the analytics script in either the HEAD section (not recommended) or BODY section (recommended - just before the closing tag)?
I'll bet the solutions to this problem is really simple and obvious... once you figure it out :-)
Hey guys - this is my first post yet a good opportunity for me to jump in. I've used Urchin for years, even before it became Google Analytics. Excellent product. I've user AWStats for some time too, but gave up a few years ago.
I am still not sure if I read an explanation that makes sense. We use Google and AWstats as well, and in terms of pageviews there is a factor 3 more pageviews in AWstats than Google. In terms of visitors it's 50 % more.
It can therefore never be the disabled cookes or JS since that can make a difference of 10 % maybe 15-20 % but certainly not more (95 % of users have js switched on and our average visitor is not a geek but more a layperson who I doubt will have cookies switched off).
Our hosting company also splits traffic on AWstats between normal users and spiders, bots, and more.
We have the script running at the bottom of the page, so perhaps early terminated pages may be an origin, but to such an extent?
I read one interesting comment on Norton. Does anyone know if anti-virus software can block google analytics? If you need to somehow set Norton/McAfee etc to allow google than I can see a large group of people maybe not allowing for this.
We do use a lot of JS, and sometimes do have errors in IE (which is predominantly used by our visitors still), so perhaps interaction with other JS may be a reason. How would we test this?
Is there any other origin.. spam attacks for example (we notice a lot of porn keywords in awstats whereas google analytics does not show any)...
Although a lay person I get the distinct feeling some spam attacks may create a lot more traffic which somehow would not be measured by Google..
Log spam would show up in your logs but not run the Analytics JS. Probably using a common user agent so not showing up as bots in AWStats
People with JS turned off would show up in your logs but not run the analytics JS.
People with certain security products may have the JS stripped out of the page and therefore not run it. Especially if they are non-techy and just accept whatever the defaults are.
Scraper bots would show up in your logs but not run the Analytics JS. These in particular are likely to be trying to pass as a normal user and AWStats probably doesn't know they are bots.
Copyright checking bots would show up in your logs but not run the Analytics JS. Some of these do their dubious best to pass as a normal user as well.
Is AWStats separating out the 'hack' attempts as well i.e. the page requests for common scripts such as bulletin boards, or email scripts.
What sort of volume of page requests and users do you have in Analytics? How many pages does your site have? Most of the various bots will be trying to request them all.
Since we are used as a reference site by different large sites, we have had several spam attacks and I have seen scraper pages with our material. So there are definitely scraper bots. I will also check the hacks and the log spam. Thank you!
We do not have many visitors (yet).
Our site has around 7000-8000 pages, and where Google analytics says we have 6296 visits and 25.250 page views in February so far, AWstats says we have 10.248 visits and 101.238 page views.
Quite a difference, but especially your comment on those particular bots trying to request all pages could explain the enormous difference in page views, and the much smaller difference in visits.
I think this will really help and will check about the different bots (my knowledge is too basic).
How does AWstats count a visit?
GA uses a cookie, but so, without it how does it do it?
Yes.. IP address. And visits (at least on our site) expire after 1 hour, so a returning visitor would be counted as a "new visit" after that hour.