Segmentation

The aim in segmentation is to find groups of people who are sufficiently alike in some way to make them a target market. There are broadly two ways of doing this. The first way is die a priori approach that splits a market by some predetermined criterion like age, sex or social class. This method works when products closely align with the a priori segments, but often they do not. A priori segmentation sometimes works in new markets, hut it becomes inadequate as the market becomes more sophisticated. For example, when the paper-nappy market scarfed, it had one design aimed at one a priori segment - people with babies. The market became more complicated as products were made to fit different-si zed bottoms -P & G Pampers Phases have six sizes for babies as they grow from infant to waiker. P & G also discovered that the different plumbing of babies mattered and developed special versions - it seems that baby girls squirt downward, while baby boys squirt upwards and have bits sticking out. Then came products of different quality, compact versions and environmentally friendly ones. P & G markets up-market Luvs and Kimberly-Clark has Huggies with elasticated waist and Pull-Up trainer pants. The market was once segmented by a single n priori criterion, but became complicated.

Automatic Interaction Detection

Automatic Interaction Detection (AID) breaks a market down according to a series of a priori

Marketing Highlight

criteria, taking one after another. The First Direct telephone banking service used AID to identify 12 segments. It started with the knowledge of the revenue from customers and discriminating variables including residential status (owning or renting a home), occupational status, age, income and frequency of income payments (weekly, monthly, etc.).

AID analyzes the discriminating variables one at a time (see Figure 1) in order to find out which has most impact on the dependent variable - in this case, customer revenue:

1. It starts with all customers and then searches to find the discriminating variable with the most impact on revenue. If it finds out whether a customer is part of a family or not, the first division that AID performs is to split customers into families and non-families.

2. In the second stage AID first looks at families to see what factor most discriminates between them in terms of revenue. In this case it is the distinction between low- and high-income families. In the same way it splits non-families into young and grey.

,V In the third stage it determines what discriminates between high-income families (social class) and so on.

Division after division occurs until there are too few people in each segment or no further discrimination. Looking at the 12 segments (see Table 1), First Direct could see that they were very 'skewed towards young, upscale and

Figure 1 The AID process
TABLE 1 FIRST DIRECT'S TWELVE SEGMENTS FROM AID

SEGMENT

1

2 3

4

5

6

7

8

9

10

11

12

Married

Widow

Single Single

Widow

Single

Single

Single

Couple

Family

Family

Family

Family

Age

Grey

Young Young

Grev

Young

Young

Young

Grev

Middle

Middle

Middle

Middle

Home

Rent

Rent Rent

Rent

Ruv

Rent

Own

Own

Buy

Buy

Buy

Buy

Inuume

-

-

-

-

-

-

Medium

Low

Low

High

High

Class

E

AB E

C2D

C2D

C2D

BG1

AB

C2D

BC1

BC1

AB

Employment

-

Student

-

-

Self

Self

-

-

-

-

-

overcomes some of its limitations without sacrificing the benefits. Its principal applications so far are in direct marketing, but its applications are expected to widen.

SOURCES: 'High and dry', Tlie.Economist (2ft September 1991), p. 108; Patrick Moynagh, 'Exploiting [he potential of your database to improve customer service and relationship building', Institute Tor International Research Conference on Advanced Customer Profiling. Segmentation and Analysis, London, 10 February 1994; Peter Doyle and Ian Fcnwiek, 'The pitfalls of AID analysis', Journal ofMarketing; 12, 4 (1975), pp. 408-13; Steve Baron and Diannu Phillips, 'Attitude survey data reduction using GHAlDi an example in shopping centre market research'. Journal f>f Marketing Management, 10, 1-3 (1994), pp. 75-88.

A multinational drug company used to segment its market geographically until it found that its sales budgets were limited by legislation. That meant that it had to use its detailers (ethical drug salespeople) more carefully. It developed its multivariate segments using the prescribing habits of doctors for numerous drugs. It identified nine segments of doctors with clear marketing implications. Among them were:

• Initiators who prescribed a wide range of drugs in large volumes, but were also eager to try new ones. They were opinion leaders and researchers, but did not have time to see detailers. This group is hard for detailers to see, but critical to the success of new products. They were recognized as 'thought leaders' and had special, research-based promotions and programmes designed for them.

• Kindersctirecks have quite high prescription rates and were willing to see detailers, but had few children patients. They arc an accessible and attractive target, but not for children or postnatal products.

• Thrifty howseroiws were often married women with children who did not run their medieal practice full time. They had few patients, prescribed very few drugs and were usually unavailable to detailers. This segment was not attractive.

This allowed the drug company to select target markets for campaigns and help detailers when selling to them.35

high-salaried segments compared with the bank population as a whole'.

AID is useful because it can quickly form segments that differ in important ways. It is also good at handling variables like gender, occupation and home ownership. In First Directs case it is revenue, but it could be usage rate, brand loyalty or profitability. The segments are also unique and distinct. AID does, however, have drawbacks. It looks at only one dimension at a time and it can miss segments where the discriminating variables interact. Also, since it keeps splitting the data, it quickly cuts the sample into groups that are too small for further analysis. Clii-square Interaction Detection (CTIAID) is a development of AID that

Online Survey Champion

Online Survey Champion

There are people all over the world trying to find ways to make money online. From stay at home moms looking to make a few extra dollars to college students and entrepreneurs, the allure of making your own hours and working from home or from the local coffee shop is very appealing.

Get My Free Ebook


Post a comment