Website Stats

Andrej Marjan andrej-igvx78u1SeH3fQ9qLvQP4Q at public.gmane.org
Sun May 1 17:19:07 UTC 2011


On Sun, May 1, 2011 at 12:05 AM, Jason Shaw <grazer-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org> wrote:

> My guess is that they each define "visitors" in a different manner. Notice
> that Google's number is the sum of the other two.  I'd read the manual and
> see how they define visitors to see where the discrepancy lies instead of
> adding another analyzer, and thus another different number.
>
> -jason
>
>
There's no such thing as a "visit" in HTTP -- only "hits", which are
individual requests for resources. The notion of a "visit" is essentially a
grouping of HTTP requests into something approximating a single user's
session on your site. HTTP is a stateless protocol, so unless you have
application-specific session tracking logic, the best you can do is guess at
user sessions from the individual request data.

Different programs use different methods for gathering usage data
(in-browser javascript, server logs, both), and different algorithms for
analyzing the raw data and reconstructing the "visit" information.

I don't know which of these is best. I know a lot of people are satisfied
with Google's approach and I presume that their in-browser data gathering
would give them a better chance of tracking visits than a purely server-log
approach, but they also miss entirely visits from people with JS disabled.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gtalug.org/pipermail/legacy/attachments/20110501/0ddfc78d/attachment.html>


More information about the Legacy mailing list