Truth is an elusive thing, and nowhere is that more evident than in understanding the value — and values — of a customer. Becoming customer-centric is an increasingly popular goal for many companies, but confusion reigns regarding the best way to achieve customer understanding. Some believe that customer understanding comes from thoroughly understanding customer value and profitability. Others see "customer value analysis" as a means of understanding the set of buyer needs and desires that form the basis for purchase decisions. Still others maintain that customer understanding comes from behavioral segmentation.
The turf battles between adherents of one customer profitability vs. customer values/attitudes vs. behavioral segmentation continue to rage, with each proponent claiming their approach is best. What is the best approach for customer understanding? The short answer is "all of the above," with all three — customer profitability, customer buyer values, and customer segmentation — having a place under the "big tent" of customer-centric understanding. The challenge is to understand the differences between the approaches, when to apply them, and how to derive strategic and tactical value from them.
Why be concerned about differences in customer profitability? For many companies, it is an article of faith that all customers are important, and all should be treated equally. While this makes for good public relations, it ignores the reality of customer value and customer contribution to company profit. All customers do not contribute the same value to a company. A simple decile analysis of customer revenue and profitability can reveal surprising insights into relative customer value. The Pareto rule generally applies, in which a small percentage of customers represents a large percentage of total revenues and profits.
Recognizing the importance of a minority of customers translates into several imperatives. First is managing the cost to serve customers. While a reasonable level of customer service is appropriate for all customers, it makes sense to balance the cost of serving a customer with the current or potential value of that customer. Take, for example, the significant differences in customer value among insurance policyholders: highly profitable customers can support agent-made phone calls and other higher cost means of communication, while less profitable customers support white mail or call center contacts. Combining this principle with customer preferences can drive customer profitability, customer satisfaction, and company value.
The second imperative is product strategy. Too many companies make the mistake of managing product offerings based on the perceived needs of the entire customer base, rather than tailoring offerings to the needs of those customers who actually drive profitability. A classic example of this is the mistake many grocery retailers make in allocating shelf and merchandising space to products that appeal to "cherry pickers" and other low profit customers instead of allocating this space for products bought by their best (most profitable) customers.
Measuring only the value of customers at the current point in time is not sufficient. What about customers who have spent a lot in the past but are not currently spending at that level? What about customers who are not currently valuable but will likely be in the future?
Customer long-term value (CLV) analysis provides an approach to measuring customer value over time. CLV captures the net present value of the future stream of revenues less the costs associated with a customer. The CLV formula also helps to identify the factors that drive value creation, including the costs and benefits of acquisition efforts, up-sell and cross-sell marketing, and retention activities, as well as the impact of customer referrals. Although more easily measured when applied to a customer base in the aggregate, CLV can be a valuable measure of customer profitability. It clarifies the extent to which a company should invest in market, sell and serve efforts, and at the disaggregate level better clarifies how much should be spent on a given customer.
CLV is a useful approach for making value-based resource allocations, especially in investments such as customer databases, customer analytics, customer interactions, and other Customer Relationship Management (CRM) related resource decisions. Accenture has developed an approach to mapping the components of customer long-term value to the supporting CRM capabilities. At the core of this approach is an estimate of the value potential derived from improvements to the key drivers of the CLV formula. These drivers may include customer acquisition rate, defection rate, products or services purchased, acquisition cost elements, selling cost elements, servicing cost elements, and customer duration. Value potential is measured using performance versus industry benchmarks or potential performance against stretch goals, depending on available data.
For example, assume that the current new customer acquisition rate is 20 percent, and that the identified potential is an acquisition rate of 22 percent. This two-point gain in new customer acquisition can be easily quantified into profit impact by multiplying the 2 percent times the size of the existing customer base, then applying the current gross profit per customer. The value potential of improvements in each CLV component is determined and then compared to opportunities identified through assessing existing CRM capabilities, resulting in prioritizing CRM investments which will most directly impact value creation.
Customer Buyer Values
A second approach to customer-centric understanding is to determine why customers make specific buying decisions and incorporate that knowledge into specific marketing and service strategies. Known as customer buyer values analysis, it involves:
o Identifying the key values that drive customer behavior
o Understanding customer preferences and the trade-offs they are willing to make
o Segmenting customers based on their values and trade-offs
o Developing product, channel, pricing, and service strategies and value propositions that best serve those segments.
The resulting value-based segmentation can support strategic imperatives including overall corporate strategy, revenue enhancement, cost reduction, process improvements, and increased customer responsiveness to offers.
Typically conducted by surveying a representative subset of customers, the buyer values analysis identifies the key data elements that drive purchasing and measures what is truly important to customers by asking them to make trade-off discussions. Traditional research typically asks the importance of particular features individually; the predictable result is that customers tend to rate each attribute as important, especially price. Trade-off research using techniques such as conjoint analysis better measures the decisions customers face in actual purchase decisions by asking them to choose between pairs of options. For example, a financial services customer may be asked to indicate their preference between opening an account in person at market rates vs. opening an account by phone at a rate 0.5 percent above market rate, among scores of other trade-offs related to the account opening channel, wait time, interest rate and service/problem resolution channels. By conducting multiple paired comparisons, weightings are determined for each variable that determine their relative importance in the buying decision.
When conducting a trade-off analysis of this sort it is critical to ask the right questions in the right way: the old adage "garbage in, garbage out" applies. First, include all the factors that are likely to influence the customer’s decision. Second, combine the factors in ways that make sense to the customer and reflect their actual decision process. Some of the trade-off techniques consider variables in relative isolation, whereas others always present full product concepts to the respondent. Because different types of trade-off analysis will result in different weight being given to certain attributes (especially price), not choosing the right technique runs the risk of missing a critical driver of customer behavior (for example, not identifying a "price sensitive" segment when one actually exists).
Customers are then grouped into segments based on their responses. For example, a segment of customers may value transaction speed and be willing to pay a premium to get it, while another segment’s behavior may be driven by channel preference, and a third segment might value low price. The most important values for each segment can be identified, leading to an understanding of what matters to all customers as well as what matters to specific segments.
For example, a buyer values analysis for a major bank dispelled several myths. The first was the belief that demographic factors such as age and income could be used to identify discrete segments; in reality, demographics were well-distributed across buyer value segments. The second was the belief that customers valued individual relationships with bank personnel; when, in reality, price and speed were found to be more important than personal banking relationships for most customers. The third was that customer satisfaction drives market success; in reality, understanding buyer values and acting on that understanding maximizes profitability.
Buyer values related to preferences and willingness to pay led to a set of client strategies including a new product offering that rationalized pricing and delivery costs; a new high-speed telephone channel targeted at high-profit segments who preferred banking by telephone; a decision to build corporate strategies around the needs of a select number of segments rather than the entire customer base; and a marketing campaign that built demand by focusing on products and delivery channels that appealed to target customers. Process improvements included simplification of new loan documentation and customer information requirements that reduced loan decision time from two weeks to three days. These and other changes around organizational design and technology led to substantial improvements in both revenue enhancement and cost savings.
A closely related approach known as customer value analysis also focuses on understanding customers’ key buying factors, but also incorporates the choices customers make in choosing between competitors. It incorporates both the relative importance of each buying factor, the rating of how well each competitor delivers on the key factors that drive the purchase decision, and how the buying factors, importance weights and relative competitive performance are changing over time.
This approach is based on the idea that customers make buying decisions based on a complex set of tradeoffs between product and service attributes and cost. It can be used to better understand customer satisfaction by looking not only at customer’s satisfaction with a particular company’s offerings but how well those offerings stack up against competitive offerings. It can also serve as a bellwether indicator of competitive threats and lead to strategic measures to counter those threats.
For example, a long-distance telecommunications firm facing a potential price war with its competition used customer value analysis to address the question of whether to maintain its premium price position or become more price competitive. By better understanding customers’ views on competitive price and long distance service quality, the company found that although their quality was perceived as superior and customers were willing to pay more for quality, their price premium was perceived to be excessive and their quality lead was narrowing. This led to a strategy of investments to improve quality, including infrastructure spending and improvements in billing and installation processes, and an advertising campaign to address the price premium perception and highlight the quality improvements.
Customer Behavior Segmentation
A third approach to customer understanding is the use of behavioral segmentation, especially analysis of customer data resident in customer databases. Database-enabled segmentation addresses a critical challenge of customer buyer values analysis — the inability to efficiently assign segment membership to all customers.
Unless all customers can be evaluated, some believe that application of customer value metrics in a useful way is impossible. Segmentations based on surveys are often not actionable except in a strategic sense unless some method of imputing the survey results to the population is developed. This is non-trivial and usually so error-prone that the inferences made are often not reliable. (Accenture, in partnership with Market Advantage, has been successful in imputing attitudinal information to non-survey respondents in the financial services industry to a certain extent.) Consequently, the advent of data mining of customer databases to derive customer-centric understanding is critical. A combination of data-based segmentation and customer value surveys delivers a more robust means of developing actionable customer-centric understanding.
Customer database behavior analysis can generate valuable insight into customers because it reveals what customers actually do — not what they say they will do or their attitudes, but their actual behavior. This analysis presumes the capture of transactional data products or services purchased, quantities, timing, promotional vs. full price purchases — as well as demographics, lifestyle, and life stage data when available.
Effective behavioral segmentation starts first with a set of business requirements, and an understanding of the critical business issue to be resolved. This shapes the choice of variables used for the segmentation and helps produce a meaningful outcome. For example, an online broker had demonstrated the ability to attract customers but was challenged with driving sufficient activity and profit from its customer base. A behavioral segmentation incorporating trading levels, account value, proxies for trading sophistication, and measures of recency, frequency and monetary value (RFM) identified six behavioral segments, ranging from novices to experienced power users. Each customer in the data warehouse was scored with his or her segment membership and demographic, lifestyle and life stage data were appended. This enabled a rich set of analysis of customer behavior, associated demographics and lifestyle characteristics. A further step incorporated survey information that allowed a "universal segmentation" of the investor universe. This universal segmentation allowed penetration analysis, potential customer value estimates and even models predicting potential trading levels.
The level of learning and insight developed from these models and analyses completely changed the way the brokerage thought about its customers. Products and offerings were targeted to particular segments and sub-segments, and segment managers were assigned. Additional heavy use of data mining, modeling, and analysis was successfully implemented and incorporated into virtually every aspect of marketing. The differential successes in marketing were such that the payback for developing the data warehouse and all the data mining activities was less than 18 months.
Many companies have adopted attitudinal segmentation approaches to customer understanding, and effectively use the segmentation to drive advertising and product strategies. Some are resistant to adopting a behavioral segmentation approach in addition to their existing attitudinal segmentation, claiming to "already have a segmentation." This is a common but shortsighted philosophy. In fact, attitudinal and behavioral segmentations can exist in parallel very effectively. Attitudes drive behaviors, and behaviors in turn drive value. Both attitudes and behavior can provide valuable customer insight, but how they are applied differs. Attitudinal segmentation is, in virtually all instances, survey-based and drives broad strategic decision making such as market positioning. Behavioral segmentation can be either survey-based or database-driven; database-driven segmentation allows all customers to be segmented, and can be used to influence all customer interactions. As a result, behavioral segmentation is an enabler of both customer strategies and tactical treatment.
The key to developing segmentations is keeping them separate. A common mistake is mixing demographics like age and gender into a behavioral segmentation. This muddies the resulting segments, often rendering them useless. Our approach is to define pure behavioral variables from the database, chosen for their business information value. Only these variables are used in the segmentation. Then other information can be analyzed in terms of the segments — called "profiling" the segments. Statistical analysis of demographic data can test hypotheses about behavioral segment drivers. The behavioral segmentation provides a powerful stratification for attitude and needs surveys: for example, what are the value drivers of the people in segment six?
Another key to using segmentations is the ability to score every customer with segment identifiers and profile information. Segment-based treatments require identification of the people in the segment. This requires a data warehouse containing all the information used to derive the segments, and the information used to profile the segments. Segment-based marketing experiments allow the development of powerful models capable of predicting customer response to marketing and other treatments and offers.
Finally, the segmentations and models must be used. Sales and marketing personnel must understand how to use the models, how to set up the campaigns and test cells, and how to retain and analyze the responses. This information must be stored in the data warehouse and used to train better models. Campaign protocols must enforce the use of the models and must require evaluation of all results. This often requires developing a new paradigm within the corporate structure. Unfortunately, people being people, new ways must be proven. Acceptance of the new modeling approach may require careful proofs in order to obtain buy-in. Senior executive sponsorship is mandatory, as is a strong advocate capable of leading the charge. If either of these two factors is missing, the best models in the world will not be used and the money and effort wasted.
The segmentations can be used as the foundation of customer portfolios that are assigned to a portfolio manager. Each portfolio may have its own set of priorities and directives. The service-oriented segment may require developing special services and treatments. The novice segments may require education and additional assistance. The low-value segment manager may need to identify a strategy to migrate the members to a higher-value segment. The point is that the segmentations provide the framework from which all CRM activities are defined.
A long-term but elusive goal of many companies is to leverage survey-based attitudinal segmentation by generalizing attitudinal information measured on a small number of customers to every record in the database. Attempts to do so have generally proved fruitless, but an approach pioneered by Accenture and its research partner Market Advantage has shown success with attitudinal imputation among financial services clients. Market Advantage starts with a group of customers for whom both attitudinal and behavioral data is available. Complex multivariate models are developed linking attitudes to behaviors. These models are then applied to customers for whom we have behavioral data but no attitudinal information. While naturally imperfect, the resulting "imputed attitudes" do move us some way towards universal knowledge of customer attitudes. Several enablers are required for this approach, including a rich behavioral database and the ability to determine which of many attitudes are likely to impute well. It is an approach suitable for select instances only, but, when applicable, can facilitate the customization of messages and offers to the "sweet spot" of the customer, with significant benefits.
Our experience indicates that segmentation is the foundation of customer insight. It is the organizing structure from which all else springs. It enables experimental validation of marketing ideas and strategies, and allows for customer portfolio managers who do segment-specific goals. Segemntations can be the platform from which new products and service offerings are designed and tested. Segmentation is not easy, nor is it inexpensive — but it is the first step to a true customer-centric enterprise.