|
In general, we believe that multi-attribute segmentation models more accurately reflect
the general goals that companies have when embarking on a program to segment their customer
base. Most purchase decisions are not made on the basis of one single characteristic. More
commonly, customers bring a host of different concerns into the picture when considering a
purchase. The company that identifies the unique profiles that customers have across all of
these characteristics when considering a purchase will do a better job of designing products
that are tailored to their needs, and communicating their unique competitive advantage to
potential customers.
Multi-attribute segmentation models are used in a wide variety of areas. While
usually developed on the basis of survey data, it is also possible to develop these
models on the basis of internal company databases. Some examples of multi-attribute
segmentation models include:
- Product Use models, where the volume of sales across multiple product lines is used to
develop the segmentation scheme. A key benefit of this approach is the identification of
opportunities for cross-selling.
- Customer Satisfaction models, where the segments are based on the key
drivers of overall satisfaction with the company and its products. These models
permit identification of totally disenchanted customers who are upset with all or
most of the key drivers versus those who have unique, targeted concerns.
- Product Design segmentation models can be developed from conjoint analysis
studies to identify groups for whom different product characteristics are most important. Rather
than developing a single product, product lines that vary in specific features can be
developed using this approach.
- Attitude and Behavior models seek to maximize sales penetration by linking
attitudinal data with the customers' behaviors. Marketing messages are tailored based
on the attitudinal information of the best customer segments, and these messages are
targeted at potential customers who share the same attitude but have not yet
purchased the product(s).
These four are merely example; a host of other areas are also applicable. Multi-attribute segmentation models
are usually developed as one of the final stages of an extensive statistical analysis. These models
are usually created using either agglomerative or splitting methods of cluster analysis, of which there
are many different specific methods, each with their own unique strengths and ability to capture
different nuances in the data.
Multi-attribute segmentation models are complex but give you the best overall picture of your
customers. The complexity of the model mirrors the complexity of the decision making process
that customers use when making purchase decisions, and it is this fact that gives these
models their power.
Proceed to Previous Page
|
|
|
