[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[cobalt-users] WebLog setup on Raq3
- Subject: [cobalt-users] WebLog setup on Raq3
- From: "Clint Decker" <nim1998@xxxxxxx>
- Date: Sat Dec 30 11:56:01 2000
- List-id: Mailing list for users to share thoughts on Cobalt products. <cobalt-users.list.cobalt.com>
I am trying to get instructions on ho to setup the WebLog script on a RAQ3.
I tried it before and I couldnt get it to work. Does anyone know of a
program that can automatically setup WebLog for each site when I create a
new virtual site. I installed WebAlizer and it does everything
automatically.. but I dont like the output. It doesnt show me all the
referres, and it cuts off a lot of days.
Im looking for a Log Generator that outputs ALL the referrers, and hopefully
fullyautomatic on a RAQ3.
Please let me know if anyone has any suggestions. Im all of a sudden
getting MASS hits and I dont know where their coming from.
Thanks,
Clint Decker
Jim Euliano<jime@xxxxxxxxxx> wrote:
> The stock cobalt web interface log analyzer is one
> step away from useless. <snip>
> are there any other better free (key word) log
> analyzers that are good?!?!?
As a specialist in web search placement who relies on good stat
reports (as well as a previous Webalizer user), I'd like to
highly recommend the WebLog Perl script. It's free, fast and
easy to set-up: figure about a half-hour for the first site and
about five minutes per additional site. Once it's set-up, your
customers can view their updated stats using their browser on a
daily basis without any intervention on your part.
You can exclude/filter any files (like jpeg/gif's) selectively
from your "top" files, agents or referrers reports; and include
or exclude any domains (like your own) from the top referrer
counts. It will also group referrer domains like yahoo.com
together as one domain as opposed to Webalizer which will count
s1.yahoo.com as a different referrer than s2.yahoo.com.
My customers love the detail they get from their WebLog reports
and rave about their new found ability to see the exact keywords
that people are using to search for their web pages. The only
thing I miss is Webalizer's detailed monthly reports that could
be accessed anytime vs WebLog's simple monthly summary:
Hits Bytes Visits PViews Month
19,220 134,548,418 1,798 3,527 Sep 1999 <bar>
If you'd like a Cobalt "How-To" for installing WebLog on your
server, drop me a line... or for more info and documentation, you
can visit the WebLog site at:
http://awsd.com/scripts/weblog/
Bill Carlson<wcarlson@xxxxxx> wrote:
> A hit is defined as a request for the server from a
> client, it doesn't matter if that request is a CGI,
> an image or an html page. A hit is the unit everyone
> can relate to, after there are way too many different
> ways to interpet the stats (a hit is an html access,
> oops, what about cgis?)
IMHO, hits are the most abused, misused and easily manipulated
numbers used for web stats. Many deceptive designers will say
"you've had 19,220 hits!" and most customers will believe that
means 19,220 visitors to their page, when the true number of
users / pageviews is just a small fraction of the total number of
"hits"... (in the example above) about 1,798 actual users
generated 3,527 pageviews and 19,220 hits.
These same "hit-focused" designers may inflate their "hit" stats
by adding lines of "bullets" or other fast-loading repetitive
graphic images to make a single page generate 20 or 100 hits per
view! My first job with a new customer is to explain this "hit"
factor and allow them to understand that "hits" are really a
useless gauge of actual web traffic. As seen above, even page
views are misleading, as a single user will often view the same
page multiple times.
Many stat report programs (including WebLog and Webalizer) will
try to estimate the actual number of individual users viewing the
site, based on the URL of the visiting agent. This means if the
same URL visits the website multiple times within a short time
period, they are counted as a single user. While this is a less
than perfect approach, I feel that it is currently the most
accurate measure of actual website visits. It may not be as much
fun as saying "you've had 19,220 hits!" -- but it's a lot more
honest.
Have fun,
Paul