CS14 Statistical techniques for identifying consumer segments: Guide
Overview
This guide will help you understand the different statistical techniques used in the consumer segmentation process.
The key messages are:
- Cluster analysis and CHAID or Classification Trees are the most commonly used techniques for consumer segmentation. Other lesser-used techniques are Discriminant Analysis, Factor Analysis, and Conjoint Analysis.
- Cluster analysis is the most commonly used statistical technique for psychographic profiling while conjoint is widely used for needs-based segmentation.
- From time to time, the research agency should update the client on new statistical techniques and their utility.
In this guide, you learn about the different statistical techniques and software available that help effectively identify the target segments for a brand.
Sections:
Comparing techniques
A comparison of the different techniques based on their applications is provided below:
Discriminant Analysis & Factor Analysis
Both techniques look for underlying dimensions in responses to questions about product attributes. However, discriminant analysis builds these underlying dimensions based on differences between the attributes rather than similarities between them. Discriminant analysis is also different from factor analysis in that it is not an interdependence technique; a distinction between independent variables and dependent variables must be made in discriminant analysis.
Discriminant Analysis & Cluster Analysis
In discriminant analysis the groups (clusters) are determined beforehand and the objective is to determine the combination of independent variables which best discriminates among the groups. Thus, it provides answers to what makes the groups different from each other. In cluster analysis the groups (clusters) are not predetermined and in fact the objective is to determine the best way in which cases may be clustered into groups. But cluster analysis does not provide any explanation of why/how the groups are distinct.
Conjoint Analysis and Cluster Analysis
Conjoint analysis is often an input into cluster analysis. Conjoint analysis identifies what product features/ attributes drive preferences for different individuals. Cluster analysis then groups individuals with similar needs/ requirements to produce distinct segments.
CHAID and Discriminant/Logistical Regression Analysis
CHAID is an exploratory tool that identifies which explanatory variables have the most influence on a dependant variable. Discriminant/ Logistic regression is a confirmatory tool used after CHAID to quantify and test significance of the relationship between the explanatory variable identified (by CHAID) and the dependant variable.
Statistical software
| Software | Description | Highlights |
| SPSS (Statistical Package for Social Sciences) | Amongst the most widely used programs for statistical analysis across sectors.
The program has a base software with data management and data documentation features. Advanced analysis is available through separate modules. |
Amongst all packages it seems to be the easiest to use for the most widely used statistical techniques.
Can be used either with a Windows point-and-click approach or through syntax (i.e., writing out of SPSS commands). |
| SAS (Statistical Analysis System) | The SAS System comprises products for managing large databases and statistical analyses of most classical statistical problems, including multivariate analysis, linear models, and clustering as well as data visualization and plotting features. | All SAS statistical analyses may be interfaced with the graphical products to produce relevant graphical descriptions of the data.
The SAS System is available on PC and UNIX-based platforms, as well as on mainframe computers. Many applications can be accomplished using simple point-and-click operations. |
| MINITAB | Provides tools to analyze data across a variety of disciplines and users, including scientists, business and industry, and education through an array of general statistics. | Data can be imported directly from a variety of file formats, including Lotus, Excel, Symphony, Quattro Pro, dBase and text (ASCII) files.
It is available for most computer platforms, including Windows, DOS, Macintosh, OpenVMS, and Unix. One can transparently transition from the Macintosh version of MINITAB to the Windows version. It is easy to learn and use with pull-down menus and dialog boxes to assist. |
| S-Plus | A high-level programming language designed for easy implementation of statistical functions with capabilities for multivariate analysis, cluster analysis.
It also offers extensive graphics and hardcopy capability. |
Flexible with regard to the implementation of user-defined functions and the customization of one’s environment.
It has dedicated modules targeted at specific application areas. S-PLUS runs on both PC and UNIX-based platforms. |