224 The Marketing Book
law. The stochastic process method is com-
monly used in building brand-choice models of
consumers. In all, there are three basic types of
stochastic process methods – the zero-order,
Markovian and learning models – and each has
its own set of assumptions.
The zero-order model assumes that past
brand choice has no effect on future brand
choice. There are studies on the existence of
brand or store loyalties using the zero-order
model approach which have defined brand
loyalty as a proportion of total purchases
within a product class that a household devotes
to its favourite or most frequently purchased
brand.
The Markov model assumes that only the
most recent purchases affect the brand-choice
decisions. Using the Markovian model, one can
measure the expected number of periods before
an individual would try a particular brand.
Markov models should be used for dynamic
market predictions such as equilibrium market
shares, average time to trial, which is a measure
of the attractive power of the brand, and for
evaluating the success of new product intro-
duction. One other area where Markovian
analysis has been employed in marketing is in
making personal selling decisions, where it is
used in the modelling of sales-effort allocation
to customers.
The third of the stochastic process methods
is the learning model, which postulates that
brand choice is dependent upon a complete
history of past brand purchases, as the effect of
purchasing a brand is cumulative. Therefore,
when applied in the brand-switching complex,
this will mean that purchase of a brand will
ultimately increase the probability of purchas-
ing the same brand again. This model may be
used in monitoring consumer behaviour.
Statistical decision theory
Decision theory is often used to evaluate the
alternative outcomes of different decisions and
to find the best possible decision. Associated
with the statistical decision theory is the deci-
sion tree diagram, which portrays the various
alternative decisions and their consequences.
Game theory, discussed below, is commonly
regarded as an analytical approach to decision
making involving two or more conflicting
individuals, each trying to minimize the max-
imum loss (minimax criterion).
One other application of game theory is for
formulating advertising budget decisions. In
statistical decision theory, probabilities of each
outcome are based upon either past data or
subjective estimates. Pricing decisions in
advertising are another area where decision
theory can be applied. The main disadvantage
of this method is the subjective estimation of
the probability for each decision.
Decision trees can also be used to decide
whether or not to test-market a new product
before launching it. Cadbury-Schweppes Ltd
used this technique to help in deciding the
feasibility of test-marketing a new chocolate
product. By carrying out a test-market pro-
gramme of the new chocolate, the earnings
obtained exceeded those of embarking on a
national launch without prior test-marketing.
This method has been used for making mer-
chandising decisions, such as finding the opti-
mum mix of sizes and widths of fashion shoes
to be ordered, especially when the possible
alternative choices were numerous and carried
high costs.Game theory
Game theory, when compared with decision
theory, has found limited applications in mar-
keting. Nevertheless, it has been applied to
retailing institutions in making product deci-
sions. Game theory helps management to
decide on its advertising budgets without any
prior knowledge of competitors’ budgeting
decisions.
In pricing advertisements, Higgins (1973)
used game theory to provide solutions. The
total reward for all the firms included in the
pricing decision study was considered fixed,
the decision resting on which product price to