Editor's note: Jeanne G. Harris is a senior executive research fellow and director of research at the Accenture Institute for High Performance Business, where she leads research in the areas of information, technology, and strategy. During her thirty years at Accenture, Ms. Harris has led Accenture’s business intelligence, analytics, performance management, knowledge management, and data warehousing consulting practices. She has worked extensively with clients seeking to improve their managerial information, decision-making, analytical, and knowledge management capabilities. Ms. Harris is co-author of Competing on Analytics: The New Science of Winning.

In the 1980s, two financial services consultants, Richard Fairbank and Nigel Morris, identified a major problem in the credit card industry, as well as a potential solution. The problem was that the industry lacked a focus on the individual customer, and the solution came in the form of technology-driven analytics.

Fairbank and Morris believed that insights from data analysis would enable a company to discover, target, and serve the most profitable credit customers while leaving other firms with less profitable customers. They pitched this idea, their "information-based market strategy," to more than fifteen national retail banks before Virginia-based Signet Bank hired them to work in its bank card division. Signet was hardly a leading competitor in credit cards at the time.

Over the next two years, the duo ran thousands of analytical tests on Signet's customer database - much to the chagrin of the company's previous, and largely intuitive, experts. They discovered that the most profitable customers were people who borrowed large amounts quickly and then paid off the balances slowly. At the time, the credit card industry treated such customers just as they treated people who made small purchases and paid their balances off in full every month. Recognizing an opportunity, the team created the industry's first balance-transfer card. As the first card that targeted debtors as valued, not just valuable, customers, it quickly took off within the industry. Ultimately, Fairbank and Morris's success with analytics led Signet to spin off its bank card division as a company called Capital One.

Today, Capital One runs about three hundred experiments per business day, on average, to improve its ability to target individual customers. These tests provide a relatively low-cost way for the company to judge how successful products and programs would be before it engages in full-scale marketing. In its savings business, for example, Capital One found that its experiments in terms of CD interest rates, rollover incentives, minimum balances, and so forth had very predictable effects on retention rates and new money coming into the bank. Through such analyses, the savings business increased retention by 87 percent and lowered the cost of acquiring a new account by 83 percent.

Through this analytical approach to marketing, Capital One is able to identify and serve new market segments before its peers can. The key to this ability is the company's closed loop of testing, learning, and acting on new opportunities. The firm's knowledge of what works and what doesn't forms the basis of a strategic asset that enables it to avoid approaches and customers that won't pay off. Few companies are truly set up to apply the principles of this test-and-learn approach, but Capital One's entire distinctive capability is built on it.

Capital One's analytical prowess has transformed the organization into a Fortune 200 company with an enviable record of growth and profitability. The value of its stock has increased by 1,000 percent over the past ten years, outpacing the S&P 500 index by a factor of 10. By comparison with its largest competitors, Capital One's stock has increased two to four times faster over the same period. Analytics are at the heart of the company's ability to consistently outperform its peers and sustain its competitive advantage.

Now consider a long-established company that has also become an analytical competitor: Marriott International, the global hotel and resort firm. Marriott's focus on fact-based decision making and analytics is deeply embedded in the corporate culture and lore. As one senior executive put it, "Everything is based on metrics here." This orientation was instilled as early as the 1950s, when founder J. Willard Marriott used to observe the occupancy of cars pulling into his motel's parking lot in order to charge the rate for a double room, if appropriate.

Over the last twenty years, Marriott has built on J. W. Marriott's early labor-intensive foray into revenue management — the process by which hotels establish the optimal price for their rooms (the industry’s "inventory"). The economics are simple: if a hotel can predict the highest prices that will still lead to full occupancy, it will make more money than it would if too-high prices led to unoccupied rooms or too-low prices filled the building but essentially gave money back to customers unnecessarily. Marriott introduced revenue management to the lodging industry, and over the past two decades has continued to refine its capability with the help of analytics — even as most competitors are constantly a step behind in their ability to optimize revenues.

Recent enhancements make the system work faster so that pricing could be easily and frequently adjusted for hotel rooms, and they have allowed Marriott to extend revenue management into its restaurants, catering services, and meeting spaces — an approach Marriott calls "total hotel optimization." In late 2003, the company began using a new revenue management system and began to use a new metric—revenue opportunity — that relates actual revenues to optimal revenues. In 2005, Marriott had a revenue opportunity figure of 91 percent — up from 83 percent in 2003. While the company prefers its franchisees to use the system, it has given its regional "revenue leaders" the power to override the system's recommendations to deal with unanticipated local events, such as the arrival in Houston of a large number of Hurricane Katrina evacuees.

A successful revenue management system has helped Marriott achieve consistently strong financial performance. Marriott employs an enterprise-wide revenue management system called One Yield. The system automates the business processes associated with optimizing revenue for more than 1,700 of the company's 2,600 properties.

Marriott hotels that have installed One Yield have seen an increase of up to 2 percent in revenue from leisure travelers in 2004, providing an annual profit increase for individual hotels totaling $86 million. In 2003, operating income rose 17 percent, while Marriott added 185 new hotels and over 31,000 rooms, approximately one-third of which were conversions from competing hotel brands. Marriott attributes these results in part to One Yield.

In addition to revenue management, Marriott has embedded analytics into several other customer-facing processes. The company has identified its most profitable customers through its Marriott Rewards loyalty program and targets marketing offers and campaigns to them. Marriott also maintains a sophisticated Web analytics capability for its online channel, through which it did $4 billion in business last year. The Web analytics group is constantly doing tests to understand the impact of changes in its Web site. Analytics has become such a focus within sales and marketing that all analytical people were recently combined into one organizational unit. Partly as a result of Marriott's analytical prowess, the company has been named the most admired firm in its industry for seven straight years in Fortune magazine's ranking.

Another analytical competitor whose innovations have kept it ahead of its rivals is Progressive Insurance. Progressive's top managers relentlessly hunt for undiscovered insurance markets and business models that have been ignored by companies that perform only conventional data analysis.

Progressive was the first insurance company to offer auto insurance online in real time and the first to allow online rate comparisons — the company is so confident in its price setting that it assumes that companies offering lower rates are taking on unprofitable customers. It has even pioneered a program that would offer discounts to safer drivers who voluntarily used the company's TripSensor technology to measure such factors as how often they make sudden stops and the percentage of time they drive more than 75 miles per hour. By digging deeper into customer information and doing it faster and earlier than the competition, the company uncovers new opportunities and exploits them before the rest of the industry takes notice. These and other tactics have paid off handsomely, as the company's market capitalization has doubled over the past four years to $23 billion.

What do the stories of Capital One, Marriott International, and Progressive Insurance have in common? They demonstrate not only the concept of competing on analytics but also the connection between the extensive use of analytics and business performance.

The success of companies like Capital One, Marriott, and Progressive demonstrates that the use of analytics can lead to better business performance and, indeed, competitive advantage. While academic work remains to be done to quantify the benefits of this still-new way of doing business, preliminary research indicates that individual analytical projects pay major dividends, and survey data confirms that analytical approaches are correlated with high performance.

Reprinted by permission of Harvard Business School Press. Excerpt from Competing on Analytics: The New Science of Winning by Thomas H. Davenport and Jeanne G. Harris. Copyright 2007 Harvard Business School Publishing; All rights reserved.