Market Segmentation in 3 Steps
Market segmentation research can be very complex – keep a simple perspective.
Conducting a thorough market segmentation project involves many elements as well as much mental, spiritual and procedural effort. A segmentation study can be done for many different business purposes:
- identify new market opportunities
- develop new products
- create better products
- clarify positioning
- communicate more clearly and cogently
- a stimulus to new ideas and business concepts
Don’t get confused by these different goals. Treat them like you would treat an app on an iPhone. An iPhone app applies the iPhone’s computing capabilities to a particular purpose. Goals focus segmentation process power on a particular organization touch point to obtain defined business outcomes. Market segmentation research can affect an organization at many different touch points, each point leads to its own kind of meaningful results. A specific business situation provides context and content. Segmentation research provides tools for processing content to improve the context.
Confused? Regardless of the particular “app” you find yourself in, you can become oriented by the fact that segmentation always involves 3 steps.
The first step in market segmentation involves selecting characteristics that might be important in defining marketplace differences. What are these characteristics and where do they come from? They come from observation and imagination… characteristics can be just about anything observable or even NOT! Observable characteristics are directly visible differences in products or services, price levels, demographics, consumption behavior, as well as communication and distribution channels. Some characteristic cannot directly seen or heard including attitudes, beliefs, personality, or other so-called latent traits. (Even though there are times I can see such things in people, especially in expressives, precise definition and measurement is required. “who-ah” is insufficient for segmentation.)
It can be really hard and take a long time to come up with what one believes is a good set of characteristics. The main idea is that one needs to come up with a set of characteristics that have a reasonable likelihood to identify important marketplace differences. If one comes up with too many independent or detailed characteristics it will be hard to find groups of respondents who give similar answers across the characteristics. On the other hand, if the characteristics are all highly correlated and too coarse then there won’t be much variety from which to glean meaningful segments.
Step two in market segmentation involves sorting the marketplace on your set of selected characteristics. To do this we typically take some kind of measurement and then sort the measurements using a statistical method to create groups of items. A successful outcome is when items are placed into collections of similar items which are also separated from other collections of different items. There are lots and lots of different techniques. For measurement there are yard-sticks, surveys, transaction data bases, click streams, phone interviews and so forth. Statistical methods for sorting and grouping include:
- cluster analysis
- factor analysis
- interaction detection (e.g. CHAID, CART, etc.)
- latent class analysis
- perceptual mapping
- … and many others
Discovering meaning, the third step in a market segmentation study, might be the most misunderstood aspect of market segmentation research. The first two steps produce a somewhat guaranteed output: items grouped in some way. That way might or might not be meaningful, useful, or a complete answer. Said differently market segmentation analyses do not result in mathematical tautologies (http://en.wikipedia.org/wiki/Tautology_(logic) ) Rather, a variety of outcomes can be obtained depending on the particular set of characteristics, measurements, and statistical algorithms used.
This means there are a number of possible segment solutions and interpretations hidden within any particular segmentation study’s data. One can “look at” the data multiple ways and obtain a different answer each time. Therefore, the final task in segmentation becomes one of discovering the most meaningful solution.
In essence, with market segmentation one creates a collection of maps that reveal different “buyer routes”. Then one chooses the most relevant and meaningful route. There is no one, correct answer in a market segmentation study. Still, there is no need for confusion. At the end, there is but one destination, and that is to increase sales and improve business performance. If the project transforms your thinking from concerns about sales routes to insights on buyer routes then you have a successful market segmentation project.
For more information about market segmentation check out Market Segmentation The Basics in English.