How Do They Know Their Customers So Well?
Internet portal Yahoo! records every click made by every visitor, accumulating some 400 billion bytes of data per day — the equivalent of 800,000 books.1 Direct marketing giant Fingerhut has 4 million names of repeat customers and stores up to 1,000 attributes on each one. Its data warehouse can hold 4.5 trillion bytes.
Can even a fraction of this data find its way into a thoughtful analysis of customer needs or more personalized, innovative services?
The answer is “yes,” but it may be a mere fraction. Companies are rushing to invest in technologies that enable them to track patterns in customer transactions. Yet when the transaction collection-and-tracking dust settles, most firms have a larger data warehouse but very few additional insights into their customers. In other words, they may know more about their customers, but they don’t know the customers themselves or how to attract new ones.
Interviews with 24 standouts in customer-knowledge management, including Harley-Davidson, Procter & Gamble and Wachovia Bank, reveal that a firm needs more than transaction data to gain such insight. Many of the executives interviewed say that their companies succeed because they consider the person behind the transaction — recording what customers do during sales and service interactions. By examining this “human” data, they can better understand and predict customers’ behaviors, and they can rely less on technologies to collect, distribute and use transaction-driven knowledge.
Not that the transaction data isn’t important. It’s why a firm can flash “Hello, Mary Jones” on the screen — along with a set of suggested products customized to her preferences —when Mary enters its Web site. But the smart firms realize they can’t just collect data. The data has to translate into something meaningful about existing or potential customers. This requires first understanding which transaction-based approaches will provide the right data.
It may also mean mixing transaction and human data, a strategy that the customer-knowledge management leaders say gives them the best results. However, having a mix of approaches doesn’t mean you have to integrate the data. Even the most successful firms seem unsure about how — or even whether — to integrate data types into a comprehensive customer database.
Clearly, questions and uncertainties abound. We have a long way to go in capturing a customer’s psyche. But to compete for customer satisfaction, firms must work harder to collect, distribute and use the right data.
What the Experts Do
The really good customer-knowledge managers aren’t afraid of the tough problems. We talked to executives in 24 firms known for their superior customer knowledge. (See “The Companies We Studied.”) All were attempting to combine transaction and human data. They were using a variety of tactics (depending largely on why they wanted the knowledge), but our interviews —and our past work with other successful companies — revealed that, for the most part, the firms have seven common practices. (See “What the Leaders Do.”)
Focus on the Most Valued Customers
Customer-knowledge management initiatives take time and effort, so a firm has to know which customers are worth the cost. FedEx, US WEST 2 and several banks categorize customers according to profitability: The customers who spend more get more services. In other cases, the customers requiring attention are obvious. In the 1980s, Procter & Gamble, which had been spending most of its effort on end-consumer research, shifted its focus to Wal-Mart and other large retailers because of their growing concentration and power. Likewise, Microsoft began to emphasize the needs of CIOs when senior executives visiting large companies saw that vastly more revenues would come from corporate, rather than individual, software purchases.
Prioritize Objectives
Initiatives are most successful when a firm first defines its customer-relationship objectives and prioritizes them according to business strategies. (See “Meeting Customer-Knowledge Management Objectives,”) The successful firms know which customers to focus on and what new behaviors the customers should exhibit. The aim of a FedEx customer-knowledge initiative, for example, was to increase the company’s share of the small shipper market. However, picking up packages at small businesses is expensive. To serve the new customers cost-effectively, FedEx decided to encourage small shippers to bring their packages to FedEx drop-off points. The clarity of these objectives made it much easier to understand what knowledge to obtain and share about small shippers. FedEx succeeded in persuading these shippers to drop off their own packages and was able to increase its share of the small shipper markets.
Aim for the Optimal Knowledge Mix
Those we interviewed had long forsaken the notion of one all-encompassing solution for knowledge management. We were impressed with the breadth and creativity of techniques in Procter & Gamble, in particular. Long considered one of the world’s best marketers, P&G uses data-driven approaches to gain consumer knowledge and human-centric approaches to understanding retailers.
One of the more interesting methods is to build highly detailed mental maps that capture consumers’ thinking about various products. The maps are based on extended discussions with typical consumers and on the input of P&G marketing people, who walk the floors with shoppers, noting what they say or don’t say and what they do. With thorough mapping, the company can gain insight from just a few consumers. To evaluate products, P&G uses both mental maps and focus groups (an extension of mental maps that involves more consumers but with less detail). It also relies heavily on statistical data from point-of-sale transactions. Recently, it began to focus on managing explicit knowledge about key retailer chains as customers, and P&G encourages members of customer teams to capture and share important retailer insights with team members who serve the same retailer. In some cases, knowledge sharing is face to face. When teams are geographically dispersed, sharing occurs via knowledge repositories and discussion databases.
Because these different types of customer knowledge typically involve different types of customers, P&G doesn’t attempt to integrate them. In the aggregate, however, P&G has compiled a huge amount of customer knowledge.
Avoid One Repository for All Data
Many companies advocate the integrated customer repository.3 The idea is to put everything the organization knows about its customers in one physically or virtually integrated database. Using an integrated database, a salesperson in one location can learn everything about a customer; while talking on the telephone, a customer service representative can find out about a customer by accessing information with one click.
Yet on the basis of our contact with more than 70 firms and our interviews with leading customer-knowledge management firms, the fully integrated customer-knowledge environment seems to be more of an intriguing idea than a practical reality. Surprisingly, most firms aren’t even striving to develop one —despite all the favorable press. Not only would the diverse forms of information be difficult to combine in one set of database records, the ongoing maintenance and structuring of such a complex repository would be almost overwhelming. And even if a firm succeeded in developing a fully integrated environment, it risks having a departing employee walk away with highly developed knowledge. Consequently, not many firms are ready to put what they know about customers in one easily accessible, computerized application.
Instead, customer data is fragmented across multiple systems and locations. One problem is that most firms have many customer types, each with a unique set of needs. How do you effectively integrate that diversity of knowledge or synthesize it into a consistent customer profile? Apparently no one has satisfactorily answered that question; even large firms are hesitating to combine hard (transaction-based) knowledge and soft knowledge into one customer database. Dell Computer, often praised as an exemplar of customer-relationship management, has not integrated the customer knowledge it receives online with what it gets from telephone call centers, for example.4
Clearly, firms need to pursue multiple types of customer information, and they need to become knowledgeable about multiple customer types. In the short run, they should try to integrate a given data type as much as possible — for example, put all transaction data on customers in one place. They should also ensure that each customer has a unique identifier so that everyone can easily locate all data and knowledge about the same customer. But for now, the best approach seems to be a mix of approaches, each in a separate database.
Think Creatively about Human Knowledge
Aside from the lack of integrated data, the most striking result of our interviews was the desire to use human knowledge. In fact, this is the primary factor that separates the leaders from the laggards. One HP executive noted that the company strives “to delight our customers by providing them with what they value and may not even know they find important.” This isn’t an easy objective. Managing human knowledge presents the thorniest of management issues — explaining, in part, the current predilection for solutions based solely on transaction data. (See “Making the Most of Transaction Data,”) Yet we saw many creative solutions to managing both explicit (documented and accessible) and tacit (understood but undocumented and inaccessible) knowledge. (See “Managing Data From Customer Encounters,”)
Explicit knowledge.
Nordstrom’s Personal Touch program is a good example of creative thinking in managing explicit human knowledge. The company knows that a person can extrapolate from past choices to current styles much more reliably than a computer, so it uses Personal Touch shoppers — fashion consultants seeking long-term customer relationships. The company trains the personal shoppers in color use, current fashions and the matching of products to a customer’s appearance, taste and lifestyle. The personal shoppers record customer likes, dislikes, lifestyle and apparel needs, ascertained through telephone contacts or face-to-face conversations. Then they apply their fashion expertise to sell the customers entire ensembles, not just individual items. The personal shoppers can also access a customer’s purchasing history to gain insight into a customer’s tastes or to suggest items that might complement a prior purchase. They can annotate a customer’s records, noting, for example, to check later when alterations will be completed.
McKay Envelopes maintains detailed information about buyers’ likes and dislikes. Fidelity Investments requires sales-people to record in a database knowledge about customers acquired during sales calls. This is particularly useful when a customer account is reassigned for some reason.
In some cases, important customer knowledge must remain largely tacit.
Both Harley-Davidson and the Jeep division of DaimlerChrysler “understand” that they must know their customers well before they can introduce product improvements that befit the brand. Both companies hire marketing professors (with strong backgrounds in the social sciences) to observe and interpret customer behavior at rallies and to discuss this behavior with managers and marketing professionals. This is an attempt to capture subtle observations and impressions in order to convert them into explicit knowledge.5 This ethnographic approach, while capturing the essence of customer attitudes, is not easy to convert into traditional, qualitative terms.
To keep managers in touch, the companies have extensive programs in which executives fraternize with customers at rallies and owner group meetings. Harley executives regularly ride their bikes cross-country to join their customers. They not only learn from the customers they meet, but they experience firsthand how it feels to ride their product over long distances. Harley’s intense and long-term focus on tacit customer knowledge seems to be successful; the company is in the enviable position of selling all the motorcycles it can make, with a long backlog of orders. It has one of the highest levels of customer loyalty of any business.
In General Electric’s appliance division, customer-service personnel often gain tremendous amounts of tacit knowledge by observing how customers use (and sometimes abuse) the company’s products. To convey that tacit knowledge to the right people, GE service reps systematically record comments from customers during service calls and give the comments to product developers for analysis. Thus, both engineers and researchers get a feel for the customer’s voice and moods. Both also listen to incoming customer calls at regular intervals.
Look at the broader context
Most of the leading firms recognize that customer-knowledge initiatives do not exist in a vacuum. Rather, their success depends on the organization’s roles and responsibilities, the workplace culture and the organizational structure.
Establish new roles.
Many of the companies created new roles or functions to help manage customer knowledge. Nordstrom established its Personal Touch shoppers. Hewlett-Packard appointed a “customer data czar” at the corporate level to establish common customer name-and-address information to be shared by disparate business units. (It’s difficult to compile substantial knowledge about a customer unless you’re sure that everyone is referring to the same person!) Clarica, a large Canadian insurance company, created a Customer Knowledge Center to develop strategies for how to manage the firm’s many sources of customer knowledge. P&G is experimenting with dedicated librarians who synthesize customer data from different sources and periodically alert P&G personnel of potentially useful information.
Move toward the customer-centric.
Many of the leading firms have made the difficult switch to a more customer-centric culture6 — something we believe contributes significantly to the success of their customer-knowledge management approaches.
Hewlett-Packard only recently made the switch — in this case, from an organizational culture built on product types, such as “midsize laser printers,” to a culture centered on the “full customer experience,” including customer support, buying, upgrading, selling, servicing, purchasing supplies and maintenance.
HP development engineers now think more broadly about their products. For example, surveys suggested that customers would be more satisfied if the “Low Toner” message were to indicate how many pages could be printed before the toner supply ran out. This design change, which will appear in HP’s future products, required adding sensors to predict how long the toner would last. Development engineers now better understand the need for such features; they see that, even though a change isn’t needed for technical excellence, it can contribute to market leadership.
Carefully restructure.
Most leading firms have realigned their organizations around customer segments, away from product groups. Microsoft (whose managers were not interviewed in this study) is a notable example of reorganization to become more customer-focused. Nearly 50 percent of executives interviewed in another recent survey (sponsored by Accenture) said that their company would organize around customer type by 2002, as compared with only 18 percent today.7
But in reorganizing, a firm should not focus its approach on only one customer type. If it does, it won’t be able to develop valuable generalized solutions that the entire organization can use. One manager of a bank’s customer-knowledge program said that his group had developed an effective approach to managing knowledge about his customer type, but groups with different customers were reluctant to adopt it. “Everyone thinks their customers are different,” he noted.
Establish a process and tools
The need for a process and tools may seem obvious, but many firms seem to stop working after selecting a management strategy — in an attempt to avoid the planning that is critical to implementing it. The leading firms work hard to deliberately manage customer knowledge, using a defined process and creating tools as needed. Marketers at Kraft General Foods developed such a process and tool for managing customer data and translating it into knowledge. Using the “3-Step Category Builder,” they build an overview of a product category, analyze it and identify opportunities for new customers, and form a plan for handling that category. These steps help them work with retailers to build share in product categories. Automated templates and frameworks make it easy to do the analysis and collect the right kind of data. Sales representatives can quickly and efficiently obtain, analyze and apply customer knowledge. A Kraft executive believes that “we have done so well [because] we focus on what is truly critical.”
Kraft’s carefully defined process for collecting, interpreting and communicating customer category knowledge has paid off. By focusing on the data elements that drive sales, sales representatives get superior results using less than 20 percent of the data that competitors need for similar analyses. Stores that use the knowledge have a marginal sales growth of 3 to 4 percent, relative to stores that don’t employ it.
Where To Start
Obviously, not all firms have the resources to do everything the leaders do, but they can learn from them. (See “Gathering and Sharing Data From Customer Encounters.”) For a firm struggling with customer-knowledge management, a few lessons may serve as a good starting point:
- Use transaction data wisely. Continue investing in collection and tracking technology, but look at business objectives to ensure you know what type of data your company needs.
- Try being more human-centric. Every company has some direct interaction with its customers or channel partners. Take advantage of those encounters. Although most firms don’t capture and codify this kind of knowledge, managers can learn how to do so by looking at the way they manage other kinds of knowledge, such as product knowledge, competitive intelligence or best practices. Start by focusing on particular customer types and objectives for better customer relationships.
- Don’t worry about integrating data, but do look at it in a broader context. How does it satisfy your customer-knowledge management objectives?
Above all, don’t be tempted by the easy path. Because there is not one best practice, successful firms are trying a variety of approaches. In any event, keep in mind that customers are more than transactions. The companies who know this and act accordingly will reap a continual harvest of new and returning customers.
References
1. H. Green, “The Information Gold Mine,” Business Week Online, July 26, 1999.
2. US WEST Inc. merged with Qwest Communications International Inc. on June 30, 2000.
3. For example, see: Rashi Glazer, “Winning in Smart Markets,” Sloan Management Review 40 (summer 1999): 59–69.
4. C. Hildebrand, “One to a Customer,” CIO Enterprise, October 15, 1999, 62.
5. This work is described in: J.W. Schouten and J.H. McAlexander, “Subcultures of Consumption: An Ethnography of the New Bikers,” Journal of Consumer Research 22 (June 1995): 43–66.
6. A. Kohli and B. Jaworski, “Market Orientation: The Construct, Research Propositions and Managerial Implications,” Journal of Marketing 54 (April 1990): 1–18.
7. Economist Intelligence Unit, “Managing Customer Relationships: Lessons from the Leaders,” 1998 (www.eiu.com/).
ADDITIONAL RESOURCES
Books
M.A. Berry and G. Linoff present 20 case studies about applying data mining to customer-relationship problems in the book “Mastering Data Mining: The Art and Science of Customer Relationship Management,” published in 1999 by John Wiley. J. Curry and A. Curry provide the basic techniques for customer segmentation, direct marketing and customer-oriented organization, along with an update on Internet applications of customer-relationship management in “The Customer Marketing Method: How to Implement and Profit from Customer Relationship Management,” published in 2000 by Free Press. “Defying the Limits: Reaching New Heights in Customer Relationship Management” (San Francisco: Montgomery Research, 2000) is a collection of articles by academics, consultants and practitioners on leading practices in customer-relationship management. A.M. Hughes’ book, “Strategic Database Marketing” (Probus Publishing, 1994), is a classic introductory text on database marketing. J.H. Gilmore and B.J. Pine edited a compilation of Harvard Business Review articles on customer relationships in a book published in 2000, titled “Markets of One: Creating Customer-Unique Value through Mass Customization.” A book by A. Payne and his colleagues, titled “Relationship Marketing for Competitive Advantage: Winning and Keeping Customers” (Butterworth-Heinemann, 1998), is a useful collection of articles on customer relationship marketing from academics and consultants. D. Peppers and M. Rogers provide concrete approaches and tactics from the founders of the “one to one” school of marketing in “The One to One Manager: Real-World Lessons in Customer Relationship Management” (Doubleday, 1999). In “Customer Connections: New Strategies for Growth” (Harvard Business School Press, 1997), R.E. Wayland and P.M. Cole give an overview of how to develop customer insights that includes a chapter on customer-knowledge management.
Web sites
www.crmproject.com contains the text of “Defying the Limits” (the book version is mentioned above) and a collection of articles on leading-edge practices in customer-relationship management.
www.customerinsight.org presents research from the University of Texas Business School’s Center for Customer Insights.
www.cio.com/forums/crm/ is a collection of CIO magazine articles and other resources on customer-relationship management.