Web site statistics
| Published date | 01 September 2001 |
| Date | 01 September 2001 |
| DOI | https://doi.org/10.1108/03055720010804168 |
| Pages | 55-60 |
| Author | Jane Yeadon |
| Subject Matter | Information & knowledge management |
VINE 124 — 55
Web site statistics
by Jane Yeadon, Web Site Co-ordinator,
Learning and Information Services,
South Bank University
There are three levels of Web site statistic s:
page counters, third party services and log
analysis packag es. Log analysis can tell the
Web master about c hanges in site usa ge over
time, the popularity of certain pages, guide
design and uncover navigational proble ms. A
case study shows how Learning and
Information Services South Bank Unive rsity
have used log analysis to fine-tune the design
of their site, LISA.
Introduction
There is a lot more to Web site statistics than the
simple Web counter that displays the number of
hits on that page. A bewildering range of informa-
tion is available to Web masters on how their site
is being used, by whom and where they were. We
are now in the position to know so much about the
usage of our Web sites that we are in danger of
being overwhelmed. This article looks at how to
collect this data, how to analyse it and most of all
how to make it work for practical benefit.
What statistics are available?
Statistics can be gathered at three main levels:
1. Basic – Page based counters, displaying the
number of hits on a pag e;
2. Intermediate – Third party services that
provide reports on site usage;
3. Advanc ed – Analyser packages that the W eb
master can use to interpret server produced
log files .
Although a vast amount of potential information is
available, it is not 100% reliable or complete. A
major problem in most cases is caching. Most
browsers cache pages so that when a user looks for
a second time at a page - the request will be served
from th e cache not the Web s erver. As a cons e-
quence there is no record on the server for the
request. Many ISPs and other institutions run
caches to speed up connections times. A popular
site’s hits will be underestimated. Some of the
information could in theory be incorrect; for
example users can use anonymisers that deliber-
ately send false information about the browser they
are usin g.1
Basic
Adding a Java script driven counter to a Web page
is an extremely simple operation. There are many
places on the Web where you can find the code to
cut and paste into your own page.2 The coun ter is
a simple numerical display indicating the number
of times your page has been accessed since it was
installed.
Though this is undoubtedly simple, quick and
cheap, its value as a source of statistics is limited
for example a counter will usually only record
visits to a single Web page. This method is also
only for those confident of high volume traffic for
their Web site. It would do little to enhance the
credibility of your Web site if there was a Web
counter on the opening page indicating that it had
been visited only a handful of times. Nevertheless
a Web counter is a simple way to give users of
your site an indication of how well it is being used.
Intermediate
A step up from the simple Web counter is a third
party Web site tracking service. Again there are
many services available offering a variety of
statistical information.3 These are usually fr ee for
sites with low to medium hit rates (typically less
than 20,000 hits per month). You will usually have
to put the logo of the service on your Web page -
this is the mechanism by which the third party is
able to gather the information about your users.
The graphic is situated on the server of the site
tracking service. When your page is accessed the
graphic will be downloaded along with the HTML
and other page elements. This download is tracked
providing you with the usage statistics for that
page.
To continue reading
Request your trial