Web analytics - where does the data come from?
Web analytics may sound daunting but it simply refers to
looking at who is visiting your website, where they came
from and how they behaved once there. Most importantly though,
it is vital information for any online marketer.
So how do you find all this information. Where does the
data come from that has to be analysed?
There are two methods of data acquisition. Either a small
amount of JavaScript can be added to each page for which
information is required and then the data this JavaScript
generates is analysed or the log files that every website
automatically generates came be used as the raw data.
The javascript method
This is usually effected using a hosted subscription service.
By paying a monthly subscription, the service provider provides
and analyses the JavaScript data and provides an interface
through which it can be viewed. The level of analysis provided
depends on the level of service subscribed to and the particular
provider.
The advantage of the JavaScript method is that you do not
need to access your log files, and the cost of analysis
is spread over a monthly fee. The downside is that that
fee has to be paid in perpetuity, and that your data is
held on the service provider's server, that is you do not
own it or have final control over it.
The log file method
With this method, some log file analysis software is required.
Again, the level of web analysis provided will depend on
the functionality of the software. The log files are downloaded
from the website server and then run through the web analytics
software. The analysis software will then analyse the data
ready for use.
The advantage of log file analysis is that you own the
data (which has security advantages where this is business
critical, such as for banks), and that after a one off payment
for the purchase of the software, there is no further financial
investment necessary. The downside is that log files must
be downloaded and may be lost if not downloaded in a certain
timescale. However, with higher end products, such as ClickTracks
Pro, this is usually automated which avoids the problem.
The differences in the information the two methods provide
We provide training on how to use ClickTracks to analyse
your website data so lets look at how the two flavours of
ClickTracks, the hosted JavaScript version and the log file
version, differ in the information they deliver. Although
very small, there are significant differences.
By the way, you can download a free trial of ClickTracks
@ http://www.clicktracks.com
The robot report
One big and important difference is that the JavaScript
method of data capture cannot monitor robot spider activity.
(This is an intrinsic shortcoming of JavaScript and applies
to all programs and services that use this method of data
capture.)
In order to index and subsequently rank a page, search
engines send out programs, called robots (or spiders) to
grab information on what each site, and page within it,
is about. In order to do this, the robot must be able to
follow the links within a site, otherwise it cannot find
all the pages. Robots are simple beasts and find static
html links the easiest to follow and although robot technology
is improving but it is not keeping pace with the technical
developments of website design and a particular area of
difficulty is the proliferation of content management systems.
Some are spiderable but many are not.
ClickTracks (log file version) has a robot report that
shows clearly which pages the robot has looked at and when.
This report is probably the most useful tool there is if
SERPS, ie organic or free rankings, are your aim. All your
efforts at on and off page optimisation are totally useless
if the search engine robot simply cannot find the page.
The robot report shows exactly what is being found and what
is not.
Returning visitors
The only way in which log files can pick up if a visitor
has visited the site before is if the site has set persistent
cookies. This is done automatically with the JavaScript
version.
Access to data
Sometimes technical issues affect which web analytics method
is used. It must be possible to add the requisite JavaScript
to each page for the JavaScript method. Equally, it must
be possible to access the raw log files in order to analyse
them. Occasionally getting at log files can be problem.
It may simply be a matter of security in that ftp access
to the site is required in order to download the logs. This
can usually be overcome though by the IT department downloading
the log files and then making them available to the marketing
department off line.
A bigger problem is when the hosting company does not provide
raw log files but insists on presenting the website stats
using only their own package, which is usually poor to very
poor. All website generate log files but some hosting company
servers are simply not configured to provide individual
websites with their own log files. When multiple sites are
hosted on a single server, sometimes the configuration does
not separate out the individual sites' data. In this case
the only solution is to move hosting company.
The scope of the analysis
The JavaScript method will give results on pages only to
which the required JavaScript code has been added. This
is usually done automatically but any pages without the
code will not feature in the analysis.
Equally, information about a particular page will only
be available in log files if there has been some activity
on that page, ie someone has visited it.
The web analytics choice
Whether to use the log file method or JavaScript may well
be determined for you. If getting access to your log files
is a problem then JavaScript is the way to go. If robot
activity is vital then the log file approach is essential.
It may come down to a commercial decision based on whether
you prefer a one off payment for a log file analysis program
or subscribing to a monthly service. Or you may base your
choice on whether you prefer to have everything running
on your own machines or whether you would prefer to have
someone else responsible for running the software as happens
with the hosted JavaScript solution. Whichever approach
you take, the most important thing is that you analyse your
visitors' behaviour, anything else is flying blind when
the competition has their eyes wide open!!
Search Engine Workshops Ltd
Rosedale House
Rosedale Road
Richmond
TW9 2SZ
17th August 2005
