[thelist] click-through statistics

[Joe Crawford]

Stan Haskins wrote:

> The organizers of a project I'm working on are interested in tracking the
> amount of users who link to other sites from within our site. We are
> currently using thecounter.com to track users of our site and how they got
> there. Does anyone know how we can find out what links they're following
> from within our site?

If you watch the URLs on yahoo - you can see how *yahoo* does it -- check it
out:
Go to:
    http://www.yahoo.com/
Search for "evolt" (of course!)
    http://search.yahoo.com/bin/search?p=evolt

See that link for the browser archive? It looks like this:
    http://srd.yahoo.com/srst/22924617/evolt/1/1/*http://browsers.evolt.org/

Pretty fancy - when I click it I get filtered through a CGI redirect to
browsers.evolt.org.

Yahoo can now track that in a fancy way (say, with each cgi redirect it gets
plopped in a database, or to a flat file) - or heck, they can just look at
the logs for srd.yahoo.com and see what people are clicking through to just
by parsing the standards HTTP logs.

You could do the same deal with any server parsed language - then have some
conversion of all your pages so anything that's an external link gets some
kind of filter.

Now, one caveat is that I'm remembering that having the :// combination as
part of a url that's not the protocol may not be legal ... but I'm not 100%
sure. But you could do the same thing and just pass the URL like so:

http://www.stanhaskins.dom/external/site=browsers.evolt.org - then have site
be the thing that gets redirected to.

Make sense?

    - Joe
--
...........  Joe Crawford : thinking and design about the web
.... enigmatic narcissism and miscellany : http://artlung.com
.... community instigator : http://WebSanDiego.org
.... San Diego, California, USA .....................AAAFNRAA