It's a Perl5 script that analyses WWW logfiles for visitor information like the :|
- most commonly used ENTRY points into the website
The reports are generated in HTML format for easy viewing (sample reports).
aWebVisit 0.1.6b by itself can be downloaded from here (120 KB). It was released on 18/Feb/1999. (version 0.1.6b contains a minor bug fix for version 0.1.6, which was released on 17/Feb/1999.)
Last changes are the configurable exclusion/inclusion of URLs, hosts, entry points (e.g. for robots), partial removal or replacement of URLs, ... And last but not least, the adaptation of the statistics file for use with aWebVisit-Map 0.1.6. You can download the CGI script separately from here (100 KB).
Note : if you have problems downloading this software, please try again first. If the problem remains, please post a message at the new Support Forum for aWebVisit with the symptoms and your browser version.
In order to get the graphical web maps, you'll also need the fly program (free). It's just three little steps :
If you want to use the companion program aWebVisit-Map, you'll need to configure at least some parameters since it's a CGI program (hint : ask your webmaster to install it for you). Don't worry, it's not very difficult : everything is explained at the beginning of the script. And the results are worth it...
Getting reliable visitor information has always been tricky. Some tools rely on
a specific SSI (server-side include) or CGI for each page to keep track of visits, others use
the referer log information, while others work with cookies to keep track of sessions.
All of this is fine, but what of the logfiles you collected in the last few months/years ? And do you even want to change your pages or server for this ? The only non-intrusive way is to use referrer information, which is often standard on webservers, but this does not really work for frames inside your website. And some browsers don't send the referrer, while others might send anything...
So, here we are with aWebVisit and aWebVisit-Map. It's easy, it's free, and it's yours for a try-out.
It may not be perfect (especially with AOL), but it'll give you a good idea of what people might do on your website. Just don't forget to send some feedback !
If you're also interested in getting the 'standard' web statistics, try Analog. If you want to find out which search engines led to your website, try Relax. They're free too...
So, if aWebVisit tells you that the average time spent on a particular page is 75.1 seconds, this does NOT mean that people actually spent 75.1 (or even 70) seconds on that page on average.
What it does mean is that - for pages where aWebVisit counted a significant number of hits as entry pages and/or transit pages - those people who can be identified by a unique IP address during a visit are probably spending more time on a page with an average of 75.1 seconds than on a page with an average of 49.3 seconds. If they are not all taking a coffee break at that page, that is...
And if aWebVisit tells you that the link from /abc.html to /def.html was followed 14 times, this does NOT mean that 14 people actually followed that link. The link might not even exist, except as random noise in the sequence of webserver hits.
But if the link from /abc.html to /def.html was followed 12514 times, and the link from /abc.html to /ghi.html was followed only 3501 times according to aWebVisit, well, you might consider making that link on page /abc.html a bit more attractive to people. Maybe you'll see a difference next month...
You can have a sneak preview of some of the features that may become part of the next release (if there is sufficient feedback).
Project Homepage : http://sourceforge.net/projects/awebvisit/