Success or failure? Every segmentation effort, whether for market research or on a customer database, ultimately will be judged by management along this dimension. To maximize the likelihood of success, one's segmentation should follow five rules:
Rule 1: Define the problem with clear objectives. No one would recommend building a new house without blueprints. However, too many analysts begin work on major projects without clearly understanding the strategic implications of their work. Make sure management understands and is willing to accept how the work is to be accomplished and how it will impact the business.
The cost of a project is always a concern. If the implications of the project are understood, one can estimate its impact in terms of dollars and cents. A simple rule to estimate the value of a project is to multiply an expected return on investment by the likelihood that the project will succeed. For example, if the project is likely to return $1 million with a 50 percent probability of success, one should not spend more than $500,000 on the project.
Time is just as important as cost. If management clearly understands the implications of the segmentation, it is more likely that sufficient time will be allocated for completing the work satisfactorily. If the work is important enough, sufficient time must be allocated to doing it efficiently and correctly.
Rule 2: Determine how the outcomes will be used. Segmentation results are often complex. As such, analysts must be prepared to communicate and champion the findings, thereby ensuring that managers understand how they will use the segmentation results. Knowing the limitations to any segmentation is especially important. Often, a large-scale project produces an expectation that it will solve all problems when the results are known.
Rule 3: Select the relevant population to be segmented. A critical part of the planning of a segmentation project is to clearly define the population and ensure that all members have an equal opportunity to be part of the investigation.
Too often, marketers segment their customers without giving a thought to investigating potential non-customers. Part of the planning process must involve considering acquisition, as well as up-sell and cross-sell. The strategy I advocate is to evaluate all of the possible prospects to which the enterprise could sell goods or services.
Of course, if interest is limited to a particular sub-population, one should take care not to include people or households that are not relevant.
Rule 4: Assess relevant data. Data relevance is critical. Successful segmentations require sufficient, relevant, valid and reliable data.
If a company wants to expand its share, it cannot rely solely on data from its own customers. An acquisition strategy requires knowledge of the competition's customers, as well as its own. Furthermore, one cannot assume that the data in one's customer database is complete. Banking customers, for example, usually have numerous financial relationships. Without augmenting a database with data from a third-party source, any segmentation would be greatly flawed.
Another problem in completing a segmentation is an over-reliance on demographics. Because demographics are easy to collect and are among the most populated fields in any database, they are the most commonly used identifiers of markets. Unfortunately, there are few cause-and-effect relationships between demographics and product usage, with baby food and women's lingerie as obvious exceptions. One should make certain any demographic relationships are fundamentally sound and not just statistical flukes.
Scales must be valid and reliable. A valid scale measures what it purports to measure. A reliable scale delivers the same results every time it is used. Analysts should use scales that have a history with the organization. In addition to being familiar to management, it is likely the strengths and weaknesses of those scales will be known.
Lastly, one should make sure there is sufficient data. As one subdivides a market into segments, the sample sizes become smaller. Whether the data resides in a customer database or is collected in a survey, one should determine what relationships are to be assessed and make sure there are sufficient data points.
Rule 5: Choose the appropriate analytic process. Everyone uses segmentation. Logical groups such as heavy versus light users, demographic categories and recent/frequency/monetary classifications are segments. Empirical groups discovered through cluster analysis, or some other process, are also segments. These are all reasonable ways to divide markets. The critical issue is that the analytic process must be most effective in maximizing the goals of the project.
Too often, analysts are eager to try the latest rage in segmentation. One can use neural networks, fuzzy classification or some other sophisticated technique, if appropriate. However, the objectives of the project must dictate the tools to be used and not the other way around. Make sure the technique fits the problem and that management will understand and accept the results.
Segmentation of one's market through research or database analysis is critical to an enterprise's success. By providing clear direction to management about best customers, their characteristics, and likes and dislikes of products, efficiencies can be determined. Having such a system in place ensures the likelihood of reaching the best customer with the most relevant message and closing sales.
To achieve these goals, however, requires careful planning and execution of one's segmentation strategy. Following these five rules will provide the structure needed to maximize the probability of success.
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