Cluster Analysis
Cluster Analysis Software
Cluster analysis segmentation is a method of grouping people into similar attitudinal or behavioral types. Using a selection of variables, for example, Simmons Psychographics, it is possible to use Cluster analysis software to segment a particular audience into a number of distinct groups, each of which has its own set of attitudes, behavior patterns and media preferences. Cluster analysis groups respondents together so they are as similar as possible within groups, yet as different as possible between groups.
Although most clients implement Audience segmentation segment by Lifestyle or Psychographics, it is possible to cluster using any variables you like (e.g. TV programs, sporting activities), as long as you are using scaled data. In the case of Simmons Psychographics, respondents are given a score of 1 to 5 depending upon whether they are ‘Far Above Average’ (1), ‘Above Average’ (2), ‘Average’ (3), ‘Below Average’ (4) or ‘Far Below Average’ (5) with each psychographic. Individuals with similar patterns of responses or ‘scores’ are then clustered together, as shown below:
|
Resp. 1
|
Resp. 2
|
Resp. 3
|
Resp. 4
|
|
| Enjoy eating junk food |
1
|
5
|
1
|
4
|
| Health foods are for fanatics |
1
|
5
|
1
|
5
|
| I take regular exercise |
5
|
1
|
5
|
2
|
| I eat less fat these days |
4
|
2
|
5
|
2
|
When Should Cluster Analysis be Used?
Publishers, broadcasters and advertising agencies use cluster analysis segmentation to target groups of consumers who cannot be defined in simple demographic terms. For example, an advertiser might wish to reach people with ‘green’ attitudes, or ‘early adopters’ of technology. In each of these cases, our target audience segmentation would be difficult to define demographically.
Of course, an advertiser could attempt to define ‘early adopters of technology’ using straight demographics (for example, by using a definition of ‘ABC1 men aged 18-34′), but this method has a number of flaws.
• Some people included in the above definition will have no interest in new technology.
• The above definition excludes ‘early adopters’ in other demographic groups e.g. women, or people aged 35 and over.
Statistical Cluster analysis enables you to target a distinct group of individuals without wastage.
