Perceptions persist that using segmentation to solve marketing issues is expensive, requires large samples and produces results that are difficult to interpret and use. These perceptions are myths. In this column, I'll explode all of these myths and show that segmentation can be inexpensive, easy to understand and apply, and highly useful even when applied to a small group of people.
Large-scale segmentation studies are obviously needed in order to quantify the sizes of the segments in a population or to develop finely tuned statistical models for database marketing. These issues, however, are statistical requirements based on sampling theory. If the need is to generate hypotheses and gain new insights, small sample segmentation is highly effective. By dividing small groups of people, products, markets and other areas of interest, new insights and opportunities emerge.
Segmentation is a philosophy of data unrelated to sample size. It holds that all populations are diverse. By understanding differing segments within a population, companies can make better marketing decisions. The averages typically used to describe a population submerge diversity. Only the characteristics of the majority float to the surface.
An example from the utilities industry serves to detonate the myths surrounding segmentation. California's recent energy crisis propelled the results of deregulation into the headlines. While that state's energy problems seem to center on supply and demand, deregulation of this industry takes in a wider range of complex issues.
Since 1996, gas and electric utilities have made attempts to reinvent themselves by offering services other than energy. With wires into customers' homes, utilities can, for example, transmit information as well as energy. A wide range of possibilities exists. Which path do they take? This question is one that every company in this industry has answered or will have to answer. Segmentation can help utilities make better decisions more quickly.
Exploding Segmentation Myths
The management of an electric utility is trying to reach consensus about what new types of services the company should offer. The choices they have decided to consider are phone service, cable TV, appliance repair, home energy audits, home insulation services, satellite TV installation, Internet services, digital subscriber lines (DSLs)/broadband, home repairs and appliances sales.
Although they have met many times over several weeks to consider these options, the eight members of the executive committee remain divided. Because the committee doesn't understand the fundamental differences that separate the members, their decision-making process has been difficult. If they knew the essential issues that divide them, they could concentrate on their real differences and work toward consensus. Although they don't realize it yet, their real differences form a pattern. The committee members can be grouped into segments that share similar opinions.
Because decisions have to be made and they are getting nowhere, the executives agree to rank order a list of 10 services from the one he or she most favors to the one least favored. The data for this segmentation consists of a rank ordering of the 10 services from eight executives.
Segmenting the Rankings
In analyzing this data, we looked at the relationship between the executives and the list of services the utility could provide this was the problem that segmentation could help solve. By determining these relationships and the patterns they formed, we helped the executives understand the strategic differences between them and move their discussions to a higher plane. Before the segmentation was created, the executives discussed each potential service individually. They weren't aware of the philosophical divide that separated them.
When determining the segments, we used MDPREF, a multidimensional scaling algorithm, to evaluate the preference relationship of the services by the executives. Figure 1 shows the executives as points on a grid and the services as arrows. The more an executive's point is located toward the tip of the arrow, the more he or she favors that service.
Figure 1: Using MDPREF to Evaluate the Preference Relationship of the Services
The segments that exist among the eight executives are easily seen in Figure 1. Executives 1, 2, 3 and 4 favor traditional home services; namely, appliance repair and sales, home insulation and repair, and energy audits. In contrast, the segment made up of executives 5, 6, 7 and 8 favor "new economy" services, such as satellite and cable TV, telephone, DSL/broadband and Internet services.
One "discovers" the meaning of the dimensions by interpreting the differences between the services on one side of the axes versus those services on the other side of the axes. Although Figure 1 displays two dimensions, we focused on the dimension represented by the horizontal axis. This dimension breaks out services related to the "old" versus the "new" economy. When the executives understood their two segments and how the issues divided them, they were able to explore their biases and quickly reach consensus.
This example shows that segmentation can be a low-cost tool effectively used on a small sample. Its application is limited only by one's imagination. Instead of viewing a population with a mundane perspective based on an average, segmentation gives us the option of finding exciting subgroups upon which some rewarding action can be taken.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access