Tutorial: Using R for Customer Segmentation

Abstract

In today's noisy world, successful marketers communicate with prospects and customers using relevant messages synchronized to the individual's changing needs.
Segmentation along a number of "orthogonal" dimensions allows construction of messages, selection of offers, and the targeting of recipients which approach one-to-one customization.

• life-stage in relation to organization and offering:
  one life-stage sequence could be: welcome, nurture, build, retain, and reactivate
• preferences among various product specifics offered:
  model lines, style, colors, related products are some examples
• demographic and psycho-demographic properties:
  socio-economic grouping, family stage, gender, age, etc.

This tutorial will work through applying R clustering packages to the customer segmentation problem using actual (but sanitized) data from our practice.

Sample date, R code, and slides will be provided to students.

We will show examples using the packages/methods:

• cluster
• kmeans
• flexclust
• mclust
• flexmix

Goal

Upon completion, students will be able to apply these methods to their own customer segmentation problems or other clustering projects.

Skill level

No marketing background is assumed. Basic R skills. This tutorial is a sequel to my useR! 2006 tutorial (which will is available for review here) and will assume R skills as covered in Part II - EDA and Basic Statistics.