Information Technology Reference
In-Depth Information
•
a
ffiliate Marketing (division)
C
• ommission Junction (Business Unit)
c
•
j.com (primary domain)
u
•
k.cj.com (domain for the UK)
d
•
e.cj.com (domain for Germany)
m
•
embers.cj.com (domains for affiliate application)
E
•
tc.
and i am sure they have hundreds of domains to complement these and hun-
dreds of vanity domains on top of that as well. as you can see, we are left with a
decision as to how to group these domains (what tracking script to place on which
domains) and the future reporting implications of those choices. My best advice is to
replicate company structures and group not by web domains but by matching goals
and thus data similarity. in our ValueClick example, that could look like this:
P
•
roject 1: ValueClick Corporate Marketing Websites
v
15
•
alueclick.com
v
•
alueclick.co.uk
P
•
roject 2: Pricerunner Shopping Comparison Websites
p
•
ricerunner.com
p
•
ricerunner.co.uk
p
•
ricerunner.dk
p
•
ricerunner.se
P
•
roject 3: Commission Junction Public Websites
c
•
j.com
(primary domain)
u
•
k.cj.com
(domain for the UK)
d
•
e.cj.com
(domain for Germany)
P
•
roject 4: Commission Junction affiliate applications
m
•
embers.cj.com
By grouping some 10 Pricerunner countries into one big—and at first sight,
messy—pool of data, we create the opportunity to do two types of data segmentation:
vertical and horizontal segmentation.
Vertical segmentation means that we can, at any point and on any of the col-
lected data, create a segment based on, say, the dimension entry domain (such as
entry domain =
pricerunner.co.uk
for the UK country manager). or it could be, and
probably more correctly so, a segment based on the dimension of Visiting Countries
