What are direct marketing and interactive marketing? 571
Figure 22.1 displays the result of applying
a statistical model called CHAID (CHi-squared
Automatic Interaction Detector). This is also
sometimes called tree segmentation. Here we
are using CHAID to analyse the results of our
last mailed appeal. We want CHAID to tell us
how to recognize the differences between our
most generous donors and our less generous
donors. In particular, we would like to know
who – if anyone – not to mail next time
around.
CHAID splits the mailing base (all donors)
into two, by picking out the most important of
all the discriminatory variables that distinguish
the best donors from the others. This variable
turns out to be the number of previous gifts.
The 36 per cent who have sent us two or more
previous donations contributed 65 per cent of
the money. This result is shown near the top of
Figure 22.1.
Looking further down, we can see that
CHAID keeps on dividing each segment into
two, like an amoeba in a Petri dish. Each time it
takes the most significant of the remaining
discriminatory variables. What CHAID is
answering each time is:
Of all the differences between the more gen-
erous and less generous donors in this segment,
which is the most important single difference?Looking at the left-hand side of the model, we
see that the least generous 40 per cent of donors
contributed only 11.5 per cent of the cash
received. The model demonstrates that 88.5 per
cent of the cash could have been raised from 60
per cent of the donors by not mailing single gift
donors who had been inactive for over 12
months and who last gave at a different time of
the year.
If we assume it cost £1 in mailing expense
to raise £4 in cash, the overall result from
100 000 donors would be £400 000 raised at a
cost of £100 000. However, the least responsive
40 000 would have cost £40 000 to mail and
brought in only £46 000 cash. We might decide
to send an appeal to these donors only once a
year, at Christmas time, when they are most
likely to respond. Meanwhile, the other 60 000
donors sent back £354 000 – almost £6 for every
£1 of expense.Sending our appeal to these donors only
would improve our income to expense ratio by
almost 50 per cent.The direct marketer looks for solutions by
listening to what the data say. As always in
direct marketing, actions speak loudest. What
people do matters more than their demo-
graphic, socio-economic or lifestyle profile.B2B = Pareto ×Pareto
However strongly Pareto’s principle applies to
B2C marketing, the B2B scene is Pareto
squared, one company having 10 000 times the
purchasing power of another. So companies
differentiate between larger (corporate) cus-
tomers and smaller customers. Frequently call
centres or contact centres (Internet empowered
call centres able to deal with telephone and
e-mail queries and orders from website visitors)
deal with smaller business customers while
field sales teams deal with larger customers.
Dell prefers to deal with all customers
through its websites but differentiates by offer-
ing restricted (extranet) PremierPagesextending
many extra tailored informational services to
large corporate customers on a one-to-one basis
(see Chaffey et al., 2000, p. 156).Principles of direct marketing
Direct marketing and its information systems
focus on what the customer or prospect does.
To put it another way, information about past
behaviour is used to predict future behaviour.
This information is processed on an individual
basis and can be analysed and acted upon on an
individual basis, even if the number of custom-
ers reaches millions. This does not render