Marketers Must Understand Customer Value to Make Segmentation Pay

It sounds un-egalitarian, but some customers simply are more valuable than others. In fact, at most companies about a third of the customers aren't profitable&#x2014;and two-thirds of those never will be, according to <i>CFO Magazine,</i> "Playing Favorites," January 2009. In many companies, the bottom 20 percent of customers can drain profits by 80 percent, while the top 20 percent can generate 150 percent of a company's profit, ("Angel Customers and Demon Customers," Larry Selden & Geoffrey Colvin, 2003). In the current economic climate, understanding which customers drive value and which destroy it is more important than ever.<br clear="none" /> <br clear="none" /> Marketers increasingly are recognizing this reality. However, many lack sufficient insight on which customers are actually most valuable. Marketers often do not have access to the data needed to determine profitability at the customer level, or sometimes not even at the segment level. A recent CMO Council survey indicates that fewer than half of marketers worldwide have "good insights into retention rates, customer profitability, and lifetime value." And more than three-quarters believe they are not realizing the full potential of their customers, as noted in "CMO Council Unveils Preliminary Findings of New Study," Market Wire, October 2008.<br clear="none" /> <br clear="none" /> As a result, when marketing is charged with segmentation, it usually is done with an insufficient understanding of customer value. All too often segmentations are derived primarily by attitudinal, behavioral, demographic, and/or psychographic dimensions&#x2014;with customer value either missing or playing an insignificant role. If implemented, these types of segmentations are not only non-actionable, they can perpetuate unprofitable relationships or, even worse, result in acquisition of (more) unprofitable customers<br clear="none" /> <br clear="none" /> <b>Balancing Customer Needs and Value</b><br clear="none" /> <br clear="none" /> Don't misunderstand&#x2014;we recognize that to be actionable, customer segments must always share a <i>common set of needs</i> that can be met with a common value proposition. We also note from experience that segments developed based <i>solely</i> on value metrics usually fall short on these criteria. However, developing segments with insufficient insight into their value is a risky proposition&#x2014;you may end up targeting customers you really don't want. This article will share ideas on how to incorporate customer value metrics into segmentation to make it both more actionable to line management and more useful to the organization as a whole.<br clear="none" /> <br clear="none" /> <b>Recommended Approach</b><br clear="none" /> <br clear="none" /> Based on our work with market-leading companies across a variety of industries, HawkPartners suggests the following approach for developing and implementing segmentation more firmly rooted in customer value:<br clear="none" /> <br clear="none" /> <b>STEP 1: Align on how the segmentation will be used.</b><br clear="none" /> It may seem obvious, but it is critical to gain alignment among relevant stakeholders on the business objectives and potential marketing applications of the segmentation <i>before</i> conducting any market research. Failure to align on these issues often leads to unmet expectations. A segmentation that promises to be "all things to all people" almost always ends up meeting the needs of no one. One way of avoiding this trap is to hold one or more cross-functional working sessions&#x2014;including stakeholders in marketing, sales, market research, and other relevant functions&#x2014;in which these issues are addressed early in the process. Key issues to consider include:<br clear="none" /> <br clear="none" /> &#x2022; Will the segmentation be used to drive acquisition of new customers or retention of existing ones?<br clear="none" /> &#x2022; How important is it to be able to target individual customers (vs. targeting at the segment level)?<br clear="none" /> &#x2022; What type of marketing campaigns will the segmentation support?<br clear="none" /> &#x2022; What types of measures (value or otherwise) currently are used to segment customers?<br clear="none" /> &#x2022; What does a valuable (and not valuable) segment look like?<br clear="none" /> <br clear="none" /> Understanding how the segmentation will be used should influence the design of the segmentation, as well as which dimensions are used to segment the market, thus dramatically increasing the odds it will actionable.<br clear="none" /> <br clear="none" /> <b>STEP 2: Understand which customers are most valuable.</b><br clear="none" /> While it may seem counterintuitive to understand customers' value <i>before</i> assessing their needs, there are several reasons why this is optimal. First, an understanding of which customers account for the lion's share of value can be eye-opening. Second, it is critical to know how behaviors and other customer characteristics differ among the least and most valuable segments&#x2014;these insights will be useful in constructing actionable needs-based segments. Third, it is important to analyze a representative sample of customers across the value continuum when assessing customer needs (or perhaps focus analysis on only the most valuable).<br clear="none" /> <br clear="none" /> This issue leads back to a central question&#x2014;which value metrics should be used? Answering this question involves both art and science, and requires making trade offs based on data availability, time, and budget constraints. For example, in some organizations, profitability is tracked at the customer level; in others, it is not known. As a first step in developing segmentation, marketers should investigate what data is available, at what level, and for which customers. This likely will involve reaching across functions to counterparts in finance and accounting, or even evaluating secondary sources. Based on this assessment, marketers then can evaluate which value metrics can be used to develop and profile the segments.<br clear="none" /> <br clear="none" /> The following provides an overview of the most commonly used customer value measures for segmentation, from least to most complex:<br clear="none" /> <br clear="none" /> <i>1. Revenue:</i> Prior year sales or revenue typically is the most readily available value metric in almost any industry. In some cases, revenue can be highly correlated with profitability, in which case it represents a good proxy. However, there are many instances where high-revenue generating customers actually drain value due to an even higher cost to serve. Think of a retail customer who buys frequently, but only items on sale, or a financial services customer with a high account balance who is constantly in the branch or on the phone with customer service. We, therefore, challenge marketers and market researchers to identify better value metrics whenever possible.<br clear="none" /> <br clear="none" /> <i>2. Market opportunity/revenue potential:</i> A more holistic way to think about value includes both actual and <i>potential</i> customer revenue. For example, a frequent business traveler might fly a particular airline infrequently, but would be open to using that carrier much more often if a stronger value proposition were provided. Focusing only on the current revenue this traveler generates for the airline would underestimate his potential future value as a customer. Potential revenue can be estimated based on overall spend in the category and/or by potential growth in revenue over time. Both measures can be modeled from primary market research or syndicated industry sources such as IRI (retail) and IMS (pharmaceuticals). Revenue potential is most appropriately used as a value metric when share growth is the primary business objective. However, its primary drawback is that it still does not account for profitability.<br clear="none" /> <br clear="none" /> <i>3. Current profitability:</i> Customer profitability is a better value metric than revenue alone because it accounts for the cost to serve the customer. It is an especially useful metric in industries where cost to serve varies significantly by customer. There are many different ways to calculate customer profitability, but all involve allocating costs such as sales and marketing acquisition costs, service/support, and R&D, ideally at the customer level. While this is feasible for companies that have a relatively small number of large customers (e.g., Boeing), or those with very sophisticated customer management systems (e.g., Fidelity Investments), it may not be feasible for all companies.<br clear="none" /> <br clear="none" /> <i>4. Customer Lifetime Value (CLV):</i> This is a more comprehensive, but also more complex method of assessing customer value. CLV treats customers as corporate assets&#x2014;recognizing that the cost of acquiring a customer is an investment that may be recouped over time. CLV is calculated by estimating the future profit stream (revenues - service and product costs) generated by the customer, discounting it to its present value at the firm's cost of capital, and then subtracting the cost to acquire the customer. There are many available models to calculate CLV, as well as many software programs to compute the values. While CLV may be the most accurate value metric, it is not always feasible to estimate, especially at the customer level.<br clear="none" /> <br clear="none" /> <i>5. Customer influence:</i> Customer influence metrics can take many forms. Their primary benefit is that they encompass the holistic value a customer provides by including not only the value of goods and services they purchase, but also how much they influence other customers to buy. These metrics are most useful in industries where a small number of customers exert a disproportionate share of influence on others' buying decisions. For example, pharmaceutical companies place high value on Key Opinion Leaders (KOLs)&#x2014;physicians who are viewed by their peers as experts in a specific therapeutic area&#x2014;because they influence other doctors to prescribe certain medications. Influence can be measured in many ways&#x2014;including level of customer knowledge about the product/category, degree of enthusiasm in promoting or demoting a product, and size of influence network. Influence can be tricky to reliably measure, however. Marketers, even those with extensive relationship-marketing programs, often are limited by databases that don't take into account the influence potential of consumers.<br clear="none" /> <br clear="none" /> Note that the measures above are not mutually exclusive, and marketers need not limit evaluation to just one value metric. In fact, multivariate segmentation enables incorporation of multiple value metrics in the segmentation itself. What's most important is the organization buys into them. In general, "simple" metrics that have broad acceptance are sometimes better than more complex measures based on assumptions that are not understood.<br clear="none" /> <br clear="none" /> Once the most appropriate value metric(s) have been selected, the relevant customer set should be "scored" using these measures. This may require utilizing different data sources depending on the business objectives. Internal customer data most likely will be used if the objective is improving retention. If the objective is enhanced acquisition, secondary market data and/or survey data most likely will be needed to supplement internal customer data.<br clear="none" /> <br clear="none" /> Once all customers are assigned a value, it is useful to array them into quintiles or deciles based on the proportion of value they contribute. While the groups created here are not needs-based segments, the insights derived from analyzing them will be instrumental in developing the segmentation.<br clear="none" /> <br clear="none" /> <b>STEP 3: Segment based on customer needs, attitudes, behaviors, and value.</b><br clear="none" /> <br clear="none" /> You may be thinking, why can't we stop here? Why not just segment the market on customer value alone? While there are some instances where a purely value-based segmentation is appropriate&#x2014;for example, when making decisions about sales resource allocation&#x2014;in most cases this approach is not sufficient to meet <i>marketing</i> objectives. Since the goal of most marketing segmentations is to develop customized messages and/or value propositions, a viable marketing segment must have differentiated needs, attitudes, or behaviors. Unfortunately, most segments developed using value metrics alone do not fulfill this criterion.<br clear="none" /> <br clear="none" /> Thus, an understanding of "softer" measures&#x2014;including customers' buying needs, behaviors, and attitudes about the product and category&#x2014;will be critical to develop customized marketing communications. This type of information is best collected through quantitative market research with customers or prospects. In many cases, an online survey with a representative sample of customers is the most cost-effective way to get this data.<br clear="none" /> <br clear="none" /> When designing the survey, insights are needed around:<br clear="none" /> &#x2022; Customer (or prospective customer) attitudes, emotions, and perceptions about the brand and the category<br clear="none" /> &#x2022; Purchase behaviors&#x2014;brand loyalty, competitor brand usage, future purchase intent, etc.<br clear="none" /> &#x2022; Unmet needs&#x2014;how they can be better served<br clear="none" /> &#x2022; Service requirements<br clear="none" /> &#x2022; Targeting information&#x2014;demographics/firmographics, channel and media usage, geography, etc.<br clear="none" /> <br clear="none" /> Once this data has been collected, the next step is to understand how needs, attitudes, and behaviors map to key value measure(s). This can be done through several methods. One approach is to compare how needs, attitudes, and behaviors differ among the most and least valuable customers. This analysis provides clues as to the type of customers desirable for acquisition, or the types of behavior to reward. For example, a pharmaceutical company recently segmented the market for a branded prescription medication. Customers (in this case, psychiatrists) were grouped into quintiles based on the value of their prescriptions in the category over the last two years. The most valuable prescribers had some common characteristics: They were more likely to initiate therapy themselves, were earlier adopters of new medications, preferred branded medications across all categories, and were more aggressive in terms of dosing. These attributes were used to segment the market and support development of targeted messaging for the client's sales force.<br clear="none" /> <br clear="none" /> Another effective approach utilizes correlation analysis. Again, the goal is to uncover the needs, attitudes, and behaviors that are correlated with customer value. When conducting multivariate segmentation, this exercise can be especially useful to narrow down the list of potential variables to drive the segmentation. For example, a leading mobile communications company recently segmented the market for prepaid wireless services. The project began by assessing which needs and behaviors correlate with customer profitability. It turned out that the most profitable customers were also the most family oriented (many had large extended families overseas), made a high percentage of international calls, and were most likely to value an unlimited calling plan for friends and family. Multivariate segments then were developed based on these dimensions and were targeted with tailored products and messages.<br clear="none" /> <br clear="none" /> Lastly, one or more of the value metrics themselves should be included as "active" or "driver" variables in the segmentation. This will further differentiate segments by their value. Additionally, it potentially can result in segments that have different <i>sources</i> of value (i.e., one segment may be valuable because it has low churn and low service costs, whereas a different segment might be valuable because it has enormous revenue potential, despite relatively low influence).<br clear="none" /> <br clear="none" /> <b>Application and Tracking of the Segmentation</b><br clear="none" /> <br clear="none" /> None of the approaches above should be considered "revolutionary." However, by segmenting based on the needs, attitudes, and behaviors that are <i>most correlated with value,</i> marketers and market researchers can avoid the trap of non-actionable segments that are differentiated on attitudes but provide little insight into which customers to target. This approach ensures the resulting segments will be differentiated based on their needs <i>and</i> the potential value they offer, which can be used to support several business/marketing objectives.<br clear="none" /> <br clear="none" /> If the goal is customer <i>acquisition,</i> the segmentation can enable marketers to target new customers with the most potential to enhance profitability with a compelling value proposition. If customer <i>retention</i> is the objective, marketers can develop tactics to reward the most valuable customers and encourage other customers to adopt their behaviors. Segmentation also can be useful in determining which customers <i>not</i> to target. In some extreme cases, it can lead to policies that result in "firing" unprofitable customers, or, at a minimum, changing the terms of the relationship to those more amenable to the provider.<br clear="none" /> <br clear="none" /> Lastly, incorporating value metrics facilitates better measurement and tracking of segment performance. Key marketing metrics such as awareness, consideration, response rates, and even ROI should be tracked at the segment level. Even shareholder value is tracked by segment in the most customer-centric organizations. The ability to demonstrate improvement in these areas, especially among the most valued customers, provides a strong business case for senior management.<br clear="none" /> <br clear="none" /> Ultimately, a successful value-based segmentation should result in improved ability to target the right customers, pinpoint their needs, and offer them a compelling value proposition. In today's business environment, this kind of competitive advantage can be the difference between lackluster and superior business performance.<br clear="none" /> <br clear="none" /> <i>Scott Wilkerson is a partner at HawkPartners in the Washington, D.C. office. He may be reached at <a href=""></a> or by phone at 301.951.7046.</i>