Business Corner

Alternative Market Segmentation Models

By Thomas P. Frauman | January 17, 2011

Using the voice of the customer to shape impactful business strategy.

One of the most fundamental tenets of marketing is market segmentation. The story goes that all sources of competitive advantage drive from the firm’s ability to deliver superior customer value on benefit attributes that are both relevant and important to clients. To accomplish that end, the firm should understand the universe of current and potential customers—both who they are and what they want. Without this understanding, how might coatings companies deliver differentiated product and service offerings unique from competitors and the universe of other potential substitutes?
Given the nature of a diverse and complex world, segmenting our business into “bite‐sized” chunks represents a generally accepted approach to focus organizational resources and offerings that more directly speak to the needs of a specific grouping of clients. Most commonly, coatings firms employ an industry sector approach to group customers; however, the following list contains three typical means of partitioning a firm’s universe:
Industry Sector: Standard and Poor’s Global Industrial Classification (GICS), Standard Industrial Classification (SIC), or North America Industry Classification (NAICS);

Technology Platform: Epoxy, acrylic, UV curable, etc; and

Geography: Asia Pacific, North America, Europe Middle East and Africa (EMEA), etc.

Marketing executives with professional titles and roles attached to an industry sector are more common in corporate headquarters than Starbucks cups or day old Wall Street Journals. At the corporate level, following an industry sector approach affords the benefit of presenting a logical organization structure to the financial community and supporting benchmarking against well-studied economic sectors. Organizations are aligned to specialize in markets with names like transportation, building and construction, energy and alike; but, does this sector‐based structure stage impact business strategy and execution?

Throughout my career, I have witnessed numerous well-intentioned efforts to infuse the voice of the customer into strategic planning and resource allocation. With extensive data gathering employing various types of survey instruments, results are scored and compared only to reach the startling conclusion that customers in segments as diverse as waste water treatment and metal furniture all ascribe highest value to quality, price and delivery. While I don’t argue these are important benefit attributes, I do suggest that this view of the world falls significantly short of providing actionable market segment strategies that can guide the firm in fortifying and positioning competitive advantage to expand share.
Within a sector, a deeper look at customer survey data typically reveals a longer list of attributes, in aggregate not as important as the big three, but still important to some and not as important to others, puzzling, right? As numerical values are assigned to survey responses these outliers commonly have lower mean scores but stand out because of higher standard deviation values.
The acid test for the validity of a market segmentation structure is individuals within a segment should have homogeneous benefit affinities; in other words, they should want and value the same things. Furthermore, this grouping of like-minded firms should have discernibly different preferences than firms in the other segments. Let’s explore this concept further in the context of the traditional industry sector approach. As an example, if we take a look at the aerospace market is it remotely plausible that manufacturers producing high‐end corporate jets would have the same needs and wants as a manufacturer producing rockets to launch telecommunications satellites? Might it be likely that product attributes that are aesthetic in nature would be more relevant for the corporate jet and of little to no importance for the rocket? What about unit price sensitivity of the corporate jet producer, versus a mass-market producer of single engine propeller planes?
In an effort to counter this heterogeneity of preferences within a segment, often market segment managers take the approach of defining and redefining their world in more granular detail, creating ever more specific sub‐segments. Although this approach does solve some of the problems of an industry sector structure it is inherently inefficient, undermining economies of scale. Excruciatingly long business review meetings of dozens of important sub‐segments fail to inspire senior corporate leaders to invest in growing specific niches because of perceived low business impact.
Need‐based segmentation represents an alternative if we agree to reject the traditional paradigm. A variety of approaches may be employed to redraw market segment boundaries to create a new structure based on customer preferences. A more formal but fairly common methodology utilizes statistical tools from the marketing science discipline including conjoint and cluster analysis. In a less complex business context, individual clients can be grouped by team consensus based on similarities in preference data.
Regardless of the approach employed, the re-segmentation effort must drive from sound current customer preference data—garbage in leads to garbage out. The importance individual clients ascribe to specific product and service attributes forms the foundation for further analysis. Before embarking on a project of this import, marketers must determine what benefit attributes to include in their query. A preliminary listing of attributes can be assembled through multiple thoughtfully placed and well-executed focus group sessions. Respondents from the survey population can then score the importance of these attributes by completing either a simple survey (Likert scale), or through ranking hypothetical benefit combinations in an orthogonal array experimental design (conjoint analysis). The later approach may provide a more robust assessment of the trade‐offs customers make when presented with multiple benefit combinations.
Regardless of the scoring methodology, the preference data once gathered is compared in order to establish a manageable number of groupings (ideally five or less).1 Using a multivariate statistical analysis technique known as cluster analysis, respondents are clustered by calculating the minimum squared Euclidean distance between all clustering variable.2 Likewise, a simpler approach is possible in less complex businesses. The illustration above details a simple example.
For sake of illustration, call the sectors in the top table anything you like. Nonetheless, keep in mind these are typical industry sectors such as automotive, petrochemical, building and construction, etc. In this illustration very simplistic preference data was scored based on importance—high, medium, low—for three preference attributes identified as important in our focus groups.
Looking only at the top table, imagine you were the marketing director for one of those sectors. What would you do to craft and execute a game changing strategy? Go ahead and take your time.

Now turn your attention to the bottom table. If we abandon our previous paradigm, creating new segments based only on clustering the preference data we come up with an alternative segment that is actionable and supports specific strategies that speak to the needs of the clients within the segment. My non‐traditional segment names may sound funny, as was my intent, but the point is segmented in this way you get a much clearer picture of how to address these customers in a way that creates competitive advantage and supports market share expansion.

The illustration was highly simplified, intended only to paint the most general picture of the concepts I have discussed. In a technology driven context like the coatings industry, a rigorous assessment should include both product and service attributes.

I want to acknowledge Professor David Reibstein, of the Wharton Business School for opening my eyes to these ideas in an Executive Education program at the University of Pennsylvania.

1. May be more or less depending on the actual data.
2. Quick Cluster a SPSS statistical clustering program simplifies this work.

Thomas P. Frauman is a member of the Coatings World editorial advisory board and independent coatings industry consultant. Frauman has more than 20 years experience in senior leadership roles developing and executing strategy. He is a respected leader in the coatings industry and an associate of the Chemark Consulting Group. Frauman can be reached at