For post hoc segmentation
purposes we use characteristics or variables to describe consumers.
Characteristics such as demographics, media viewing, web browsing habits and
shopping patterns, which may be used to identify and reach consumers are called
descriptors or descriptor variables. Characteristics that distinguish
the segments, typically relating to the needs, preferences, attitudes and
usage, are called basis variables. Post hoc models use the basis
variables for crafting the segments, and the descriptors for targeting them.
The methods may be classified according to whether
interdependence or dependence analytic techniques are employed to segment the
market.
Interdependence or descriptive methods are
suited for strategy development studies where the objective is to understand
the market as a whole, and segment it according to the diverse needs, beliefs,
attitudes and behaviours of customers, within their distinct need-states. The basis variables for these studies include needs and desired benefits, product
usage behaviour, brand loyalty or a combination of these variables.
Dependence or predictive methods are
appropriate for taking decisions relating to new products, pricing, advertising
or distribution, with the objective of determining which customers to target
and how to target them. The basis variables used would depend on the
application of the segmentation.
For decisions pertaining to a new product concept,
in addition to product related benefits (attributes), the appropriate basis
variables may include product choice, i.e., the product a respondent
chooses or intends to buy.
Store patronage (behavioural store loyalty) and
store attributes are the relevant basis variables for studies relating to
distribution.
Price sensitivity (price elasticity of
demand) forms a good basis variable for studies relating to pricing.
Media habits, customer needs, benefits and attitudes
are relevant basis variables for studies pertaining to advertising decisions.
