Contents
Introduction
Providers of Web content are interested in answers to the
following questions:
- How often is my site visited?
- Which pages are visited how often?
- What is the "location" of the visitor (IP-addresses
or computer names, domain ...)?
- What was the referring page (search engine, link page
...)?
- Which Webbrowsers are used?
- ...
Similar questions can arise in an Intranet:
- Which pages are viewed, how often?
- When is the use of Intranet particularly high/low ...?
- ...
Using (personalized HTML-) mailings, the following questions
arise:
- How many recipients actually open the mail?
- How long does it last, until a recipient opens the mail?
- How quickly does the recipient react (e.g. with viewing
a web page)?
- ...
There are two common ways to collect such access data:
- "Server-Logs": A webserver can write
all client requests to logfiles, which can be (e.g. periodically)
processed with an analyse-tool.
The advantage of this method is the large amount of detailed
data, which can be the base for different statistical analyses.
A lot of commercial and freeware/shareware tools are available
for those analyses. In order to run those tools, the server
logs must be directly accessible.
As the method of webserver logs was created for "technical
information" about the server, it often does not cover
the information needed for marketing purposes.
- "Access counters" are provided as services
from companies (e.g. Sitemeter, Showstat, ...), they are
partially free or financed by advertising. This method usually
works with CGI-programs, which write the data to a database
system. Statistical analyses are generated e.g. automatically
and are provided via email and/or web interfaces, the raw
data is usually not available for the user. The advantage
of such services is the relatively simple handling (integration
into existing web pages) for the user of the access counter.
The disadvantage: those services are hardly suitable for
usage in an intranet or for analysis of mailings.
Based on the experiences with those two methods, a concept
was developed by creativebytes.net, which summarizes
the advantages of webserver logs (large amount of detailed
data) and access counters (simple integration into existing
web pages) and provides additional advantages: cb_WET
- Web-Tracking.
Mode of operation of cb_WET
- The software cb_WET-Server is installed on a computer
in the internet (or the intranet).
- All web pages, which should be monitored, contain a hyperlink
to a image file at the cb_WET-Server. The pages can
be located at different Web servers.
- When a webbrowser loads a page, it also requests the image
file from the cb_WET-Server, which returns the file
and logs the data of the request. This data can also be
displayed online at the user-interface of the cb_WET-Server.
The log files generated by cb_WET-Server contain the data
in a standardized format, the data can be analyzed with all
common tools ("logfile analyzers").
Advantages of cb_WET
cb_WET offers the following advantages:
- Server-spreading: cb_WET can monitor access
to arbitrary pages (all or only selected), even if they
are distributed on different webservers or websites.
- Independence from other systems: cb_WET
can be used independently of or parallel to other logging
or tracking methods. The logfiles are available at any time
for analyses. It is not necessary to access the logfiles
from the webserver(s).
- Standardized and extended logging: The data logged
by cb_WET-Server contains all information, which
is logged by a "normal" webserver and - depending
on the configuration - much more.
- High flexibility: The system allows the definition
of "virtual URLs" - the requested URL can contain
arbitrary paths and filenames. This means, that the structure
of the logged URLs can be completely different than the
directory structure of the "real" website. Thus,
regarding later statistical analyses, the raw data can be
improved (structured, grouped ...).
- Simple handling: To define the pages to be monitored,
it is only necessary to insert hyperlinks into the HTML
code (as with traditional access counters). The software
for the cb_WET-Server is easy to install and requires
no complex configuration.
- More accurate results: All parameters, which are
necessary for "switching off" http caching mechanisms,
can be configured easily.
- cb_WET can be used for further functions.
All scenarios, where a client requests files via http are
possible. Clients could be webbrowsers, HTML enabled email
programs, webenabled TV settop boxes, mobile phones with
WAP, ...). Example: Monitoring of the response time and
feedback from personalized HTML mailings.
Prerequisites
The program cb_WET-Server must run on a computer,
which is permanently connected to the "target network"
(Internet and/or Intranet). This computer could be an already
existing webserver:
- cb_WET in the Internet: Computer (Windows
NT/2000/XP) with permanent Internet connection, fixed IP-address
or hostname via DNS.
- cb_WET in the Intranet: Computer (Windows
NT/2000/XP) inside the LAN, fixed IP-address or hostname.
To adapt existing web pages for the use with cb_WET
the following sourcecode must be available:
- Static pages: the HTML code
- Dynamically generated pages: the code of the according
scripts
Further information see: Detailed
description
Comments, suggestions and ideas? office@creativebytes.net
|