3. Leveraging Proprietary Data for Analytical Advantage

Thomas H. Davenport

It is widely agreed that proprietary information provides competitive advantage, but it is scarcely useful without analysis and application in business processes. Data that no other organization possesses can provide insights and allow decisions and actions of which no other organization is capable. Data by itself normally confers little or no direct advantage, but analytics based on data can be very powerful competitive tools. At a time where traditional bases of competitive differentiation have largely faded away, leveraging unique and proprietary data can be a powerful source of competitive differentiation.

Proprietary data can provide a powerful view into company operations and the preferences and behaviors of customers and markets. In many cases, such data is valuable to other companies, competitors, consumers, and even governments. Internet leaders such as Google and Yahoo! have used proprietary data to spur new businesses and offer opportunities for discovery and growth, demonstrating that the data has value beyond first-line marketing opportunities.

Proprietary data is often a by-product of pursuing another business goal, such as executing payment transactions in banking, managing inventory in retail, fulfilling shipments, operating a communication network, or improving Internet searches. Few companies have invested the time and resources necessary to leverage such proprietary data for other uses. But those that have done so have been able to launch new products, provide outstanding customer service, and outperform their competitors. For example, Capital One mines customer data for new-product development, Progressive insurance uses proprietary data on customer driving behavior in its Snapshot program to accurately price car insurance, and Delta Dental of California analyzes claims data to identify cost savings. In many cases, the discoveries in the data led to new business opportunities that were otherwise not obvious.

Proprietary data is also being used for advantage in sports. Daryl Morey, general manager of the NBA Houston Rockets, is one of the most analytical managers in professional basketball. He argues that “real advantage comes from unique data,” and he employs a number of analysts who classify the defensive moves of opposing players in every NBA game. The Boston Red Sox follow the same philosophy. They have traveled to NCAA headquarters to categorize and quantify the paper-based records of college baseball players to analyze what attributes lead to success in the professional leagues. The Italian professional soccer team AC Milan gathers proprietary data on its players’ movement patterns under different conditions and uses it to predict and prevent injuries.

Recently, new businesses have developed around the goal of creating and mining new types of data for business gain through using social networks, selling data-derived products, or participating as marketplace creators. Many of these organizations refer to themselves as big-data firms. One company, Factual, is attempting to gather a large mass of proprietary data on a variety of seemingly unrelated topics. One account of the company described its data-gathering strategy:

Geared to both big companies and smaller software developers, it includes available government data, terabytes of corporate data and information on 60 million places in 50 countries, each described by 17 to 40 attributes. Factual knows more than 800,000 restaurants in 30 different ways, including location, ownership and ratings by diners and health boards. It also contains information on half a billion Web pages, a list of America’s high schools and data on the offices, specialties and insurance preferences of 1.8 million United States health care professionals. There are also listings of 14,000 wine grape varietals, of military aircraft accidents from 1950 to 1974, and of body masses of major celebrities.¹

However, the role of such data and its potential for spurring innovation, new sources of revenue, and new business and operational risks is still largely unexplored.

A 2009 Accenture survey of 600 executives in the U.S. and U.K. suggests that proprietary data is rare but extremely valuable. Only 10% of the survey respondents said that their company’s proprietary data “far exceeds that of the competition in terms of usefulness or significance, offering us a distinct competitive advantage.” Similarly, 86% said their company data was “about on par with that of the competition.” Yet when asked how valuable proprietary data can be in differentiating a company and its products from the competition, 97% said it was either “very valuable” or “quite valuable.”

Why such high levels of perceived value and low levels of activity with regard to proprietary data? It might be argued that most organizations and managers lack familiarity with the topic and haven’t really embedded it within discussions on strategy and competition.

Issues with Managing Proprietary Data and Analytics

Despite the fact that most managers acknowledge the value of proprietary data and analytics based on them, there are still more questions than answers about the topic. Here are some of the specific questions that organizations should address before actively pursuing proprietary data strategies:

• What are the best sources of proprietary data for my business?

• How should we convert proprietary data into proprietary insights through analytics? How do the opportunities vary by business line and strategy?

• What types of proprietary data have the most potential for competitive differentiation? How are competitors likely to respond?

• Do proprietary data and analytics have the potential in our industry to disrupt and reshape industry dynamics?

• Should we sell our proprietary data or analytics, or keep them to ourselves?

• When should we consider working with an intermediary data provider to market such data or analytics?

• In addition to selling our data, what other means of achieving value from proprietary data and analytics exist?

• How can we leverage data and analysis from third parties and syndicated sources for competitive advantage?

To address and answer these questions systematically and regularly, companies need to develop institutionalized approaches. Some organizations do so via executive-level data steering committees. Others have created Chief Data Officer positions, particularly in financial services. In any case, data-oriented discussions should address not only the problems that organizations encounter in data management, but also the opportunities arising from proprietary data and analytics.

In addition to the strategic opportunities from proprietary data and analytics, there are also organizational and regulatory implications to be explored. Because such data may contain enormous amounts of sensitive customer information, the role of a privacy protocol (especially in the presence of little regulation) is a real concern. Customer expectations brought forth by technology—such as on-demand services, remote banking, frequent-shopper identification, and transportable electronic medical records—further raise important issues. These issues include how a company should manage its data and the circumstances under which data can and should be shared across companies.

To illustrate some of the opportunities and challenges inherent in proprietary data, I’ll describe two cases. One involves a proprietary data initiative in an organization; the other addresses the potential for proprietary data in an entire industry—and the somewhat puzzling failure to achieve it.

Leveraging Proprietary Data in One Organization: PaxIS from IATA

To briefly illustrate some of the potential competitive advantages and perils in using proprietary data, consider the case of PaxIS, which stands for Passenger Intelligence Services, from the International Air Transport Authority (IATA). PaxIS employed proprietary data—or at least data that IATA believed was proprietary—on flights across 163 countries captured through the authority’s billing and settlement plan (BSP). By many accounts, international airlines found the data useful for such purposes as market share analyses, network planning and optimization, fleet planning, pricing and revenue management, marketing planning, and analysis of sales by travel agency channel. IATA sold access to PaxIS but largely relied on its airline customers to analyze the data. The authority also sold information on airline schedules (known as the Schedule Reference Service [SRS]) as a useful companion to the PaxIS passenger demand information.

However, some providers of computerized airline reservations systems (collectively known as global distribution systems [GDSs]) argued that IATA did not actually own the data, given it was collected and transmitted through reservation systems. One GDS, Amadeus, took legal action against the PaxIS offering, arguing that PaxIS constituted a breach of contract by IATA. Amadeus also charged that because new European Commission regulations prohibited it from identifying specific travel agency sales, IATA should not be allowed to do so either. In 2009, an International Chamber of Commerce arbitration panel found in favor of Amadeus and prohibited IATA from using its data in PaxIS. In 2011, the European Commission ruled that PaxIS had to remove all European data from the system. Throughout this period, Amadeus began to market its own proprietary data offering called Amadeus Market Information (previously known as Marketing Information Data Tapes [MIDT]). This offering also compiled data from travel agency flight bookings and could be used for purposes similar to PaxIS.

The case of PaxIS illustrates both the potential and the peril of leveraging proprietary data. Such data can be valuable to many participants in a value chain and can yield additional revenue and profits. But it may be subject to regulation, ownership disputes, competition, and difficulties of aggregation and management. In addition, to be of use either internally or to customers, proprietary data must be analyzed and used in business processes involving decisions and actions.

Leveraging Proprietary Data in an Industry: Consumer Payments

Every day, billions of consumer payments—credit and debit card transactions, checks, money transfers, and online payments—pass through the financial system. Several types of organizations may have access to payment data, including banks, credit card networks (Visa and MasterCard), financial transaction processors (FiServ and First Data), and financial planning and management software firms and websites (Intuit, Wesabe.com). What these institutions have in common is that they don’t take much advantage of the payments data they possess. As one executive at a firm with payments data put it, “We studied the opportunity to exploit payments data. To the team it looked like bags of money just sitting on a table. But my company just didn’t want to do anything with it.” There are many reasons for this reluctance to seize the opportunity that payments data represents, which I describe next.

There are at least three major ways to utilize payments data for positive business advantage. A couple of additional ways, fraud prevention and credit risk analysis, are intended more to prevent business disadvantage and therefore are not covered in detail in this chapter. However, many financial institutions regularly examine payments data for evidence of fraud and cancel a transaction in real time if they suspect a fraudulent payment. Some banks and credit card providers have correlated certain types of payments with higher levels of credit risk. Each of the three more-positive approaches is described next, along with the possible reasons why owners of payment data may not have exploited the opportunity.

Macroeconomic Intelligence and Capital Markets

Organizations with large amounts of payments data can use it to analyze and act on the state of the economy in particular countries or regions. A bank with substantial scale in credit cards, for example, has data on what customers are spending on what products. In many cases it can compile and analyze data faster than government sources. Using this data, the bank (or agents or customers it sells the analysis to) could invest in firms, industries, or financial instruments that benefit from the spending patterns. This is not a hypothetical example; both CitiGroup and Bank of America have used consumer spending data from credit cards to place such bets. All accounts suggest that they tend to be successful. As one banker put it, “We can predict the GDP a couple of weeks before the Fed announces it, and as a result we’ve made lots of money in the hedge markets.” Even this bullish executive, however, admitted that his bank was only scratching the surface of what could be done with payments data in this regard.

What prevents other banks and payments processors from making macroeconomic bets? Many firms that would have such data don’t have in-house capital markets groups that could make the necessary investments. Of course, they could invest through other firms, but this seems less likely to happen in practice. Making investments on macroeconomic data also may not fit with some firms’ business models. Another constraint may be the lack of economic and analytical skills in organizations to do the analysis and make investment decisions. Some banks have also been cautious in this area because they fear objections by regulatory bodies.

Targeted Marketing

Payment data provides a wealth of opportunities for learning about customers and targeting offers to them. Through it an organization can learn about discretionary and nondiscretionary spending, loyalty, life events, price elasticity behavior, and payment behavior. This makes it an ideal tool for targeted marketing to the most desirable consumers for products and services.

Actual uses of payment data for targeted marketing thus far, however, have been somewhat limited. A few banks have explored the potential of payments data to identify cross-selling opportunities. For example, if a bank detects through analyzing check payments that a customer is making payments on credit cards from other banks, the bank can offer the customer a preferred rate on its own credit card. Citizens Bank has employed targeting for online offers based on payment behaviors; the offers are for its own products and those of marketing partners and affiliates.

In addition to targeted offers, payments data can be used to segment customers for differential pricing. Pricing can be based on the usage volume, profitability, or lifetime value of services used. Some credit card firms, such as Capital One, have used this approach to charge different prices for “transactors” (those who pay off their bills in full each month) versus “revolvers,” who use their credit cards to take loans by not paying bills in full.

Payment data analysis also has value in predicting which customers are most likely to leave. A study of payment data by eCom Advisors for one bank found that the customers most likely to leave the bank did not make electronic bill payments or did so rarely and were relatively young. Targeted marketing to specific consumer profiles (young and low activity) can decrease attrition and maximize profitability.

Banks, the most likely users of payments data for targeted marketing, have been reluctant to apply it for this purpose. Many bankers focus primarily on brand-oriented marketing, rather than on targeted direct marketing. In addition, they may be nervous about negative customer reactions to targeted marketing based on payment data analysis. Some firms in other domains (Google, Groupon) have been very successful with targeted marketing based on analyzing consumer data. However, still other firms (Coca-Cola, Facebook, Amazon) have encountered resistance to targeted marketing initiatives based on customer behavior data analysis. In 2012, Bank of America began offering targeted offers (primarily of nonbanking products and services) based on payments data to debit card customers. The bank employed a third party, Cardlytics, to analyze the data.

Enhanced Customer Services

A final alternative in taking advantage of payment data is to provide information-based customer service offerings for personal financial management. A variety of potential services can be provided. Thus far, most of the providers of such services have been online startups (Mint.com, acquired by Intuit; Wesabe; Geezeo) and PC software (Quicken, Microsoft Money) that offer account aggregation, budgeting and investing tools, and financial education. Several of the sites also offer “Web 2.0” services, in which users can discuss their financial situations with others. A few also offer recommendations on products and services that the user already uses, such as a cellular telephone provider with lower rates than the one the user currently uses. Banks (such as Wells Fargo’s “My Spending Report”) and credit card firms offer a somewhat lower level of services involving spending reports and categorizations.

Third-party firms, of course, don’t have direct access to payments data and must get access to customer accounts by obtaining customer permission and linkages to their financial providers. Payment processors also typically don’t have relationships with consumers. Again, banks are the most likely to benefit from enhanced services to customers based on payment data analysis, but they have been curiously slow in pursuing these options.

Data Ownership and Permissions Issues in Payments

Consumers own their financial transaction data and generally must “opt in” to any plan to use data for marketing or enhanced services. Of course, most do so automatically when they open their accounts. There is good reason for the conservative approaches banks have displayed toward payments data. Consumers usually consider their spending habits to be personal and inviolate and probably would react negatively to unsophisticated marketing approaches that don’t provide them with clear benefits. This doesn’t mean, however, that well-planned efforts to employ payment data analysis won’t succeed. There is much opportunity to exploit this resource, but it should be handled carefully and with great attention to the privacy and security of customer data.

Lessons Learned from Payments Data

Many potential benefits are possible from leveraging and analyzing proprietary data, but most opportunities have not been aggressively pursued. It is always difficult to understand why something hasn’t happened, but the reasons organizations have not aggressively pursued this opportunity range from inertia, to lack of understanding of the possibilities, to regulatory uncertainty. To take advantage of the opportunities provided by proprietary data, companies may need an appropriate organizational structure (or partnerships with other firms); this could be another reason why many firms have hesitated. Data ownership and permission for use are other key factors to address in exploiting proprietary data.

Judging from actions by financial services firms thus far, there may be fewer concerns around preventing negative actions (such as fraud and credit default risk) than creating positive benefits (such as targeted marketing and customized offers).

Endnote

1. Quentin Hardy, “Just the Facts. Yes, All of Them,” New York Times, March 24, 2012. Page 1, Sunday Business section.