by Jeanne W. Ross, Cynthia M. Beath, and Anne Quaadgras
Artwork: Chad Hagen, Graphic Composition No. 2, 2009, digital
Companies are investing like crazy in data scientists, data warehouses, and data analytics software. But many of them don’t have much to show for their efforts. It’s possible they never will.
What’s the problem? To begin with, big data has been hyped so heavily that companies are expecting it to deliver more value than it actually can. In addition, analytics-generated insights can be easy to replicate: A financial services company we studied built a model based on an analysis of big data that identified the best place to locate an ATM, only to learn that consultants had already built similar models for several other banks. Moreover, turning insights from data analytics into competitive advantage requires changes that businesses may be incapable of making. One retailer, for example, learned that it could increase profits substantially by extending the time items were on the floor before and after discounting. But implementing that change would have required a complete redesign of the supply chain, which the retailer was reluctant to undertake.
The biggest reason that investments in big data fail to pay off, though, is that most companies don’t do a good job with the information they already have. They don’t know how to manage it, analyze it in ways that enhance their understanding, and then make changes in response to new insights. Companies don’t magically develop those competencies just because they’ve invested in high-end analytics tools. They first need to learn how to use the data already embedded in their core operating systems, much the way people must master arithmetic before they tackle algebra. Until a company learns how to use data and analysis to support its operating decisions, it will not be in a position to benefit from big data. (See the sidebar “Who Benefits from Big Data?”)
Over the past three years, we’ve conducted seven case studies and interviewed executives at 51 companies to understand how companies generate business value from data. We have found that those that consistently use data to guide their decision making are few and far between. The exceptions, companies that have what we call a culture of evidence-based decision making, have all seen improvements in their business performance—and they tend to be more profitable than companies that don’t have that kind of culture.
The digital economy is all about capturing, analyzing, and using information to serve customers. Most companies can significantly improve their business performance simply by focusing on how operating data can inform day-to-day decision making. So why don’t more companies make better use of data and analysis? One reason may be that their management practices haven’t caught up with their technology platforms. Companies that installed digital platforms—ERP and CRM systems, real-time data warehouses, and homegrown core information systems—over the past 10 to 15 years have not yet cashed in on the information those platforms make available. In addition, adopting evidence-based decision making is a difficult cultural shift: Work processes must be redefined, data must be scrubbed, and business rules must be established to guide people in their work. The good news is that once companies have made the cultural change, they usually don’t go back, and their operating improvements are not easily replicated by competitors.
Our research suggests that companies with a culture of evidence-based decision making ensure that all decision makers have performance data at their fingertips every day. They also follow four practices: They establish one undisputed source of performance data; they give decision makers at all levels near-real-time feedback; they consciously articulate their business rules and regularly update them in response to facts; and they provide high-quality coaching to employees who make decisions on a regular basis.
Before we explore those practices, let’s look at a company that has had a culture of evidence-based decision making since its founding.
Empowering Employees to Make Good Decisions
In the 1970s Southland Corporation, known for pioneering the concept of the convenience store chain with its 7-Eleven shops, divested its Japanese stores, and Seven-Eleven Japan was born. Toshifumi Suzuki, the first CEO, decided early on that the key to profitability for the company’s tiny stores would be rapid inventory turnover. So he placed responsibility for ordering—the single most important decision in the business—in the hands of the stores’ 200,000 mostly part-time salesclerks. Those employees, Suzuki believed, understood their customers and, with good information, could make the best decisions about what would sell quickly.
To support salesclerks’ decision making, he sent each store daily sales reports and supplemental information such as weather forecasts. The reports detailed what had sold the previous day, what had sold the previous year on the same date, what had sold the last day the weather was similar, and what was selling in other stores. Because Seven-Eleven Japan carries fresh food, Suzuki arranged for deliveries three times a day so that the clerks could base their orders on immediate needs. And he connected the clerks with suppliers to encourage the development of items that would suit local customers’ tastes. The result? Seven-Eleven has been the most profitable retailer in Japan for more than 30 years.
This is not a story about big data, or even about big investments in data. This is a story about a lot of little data. More important, it’s about betting your business success on the ability of good people to use good data to make good decisions. Empowering employees in this way, and arming them with the data they need, helps them make better operating decisions on a daily basis. It can also lead to a constant stream of innovation. At Seven-Eleven Japan, approximately 70% of the products on the shelves each year are new, designed by salesclerks in response to customers’ preferences.
In contrast, consider the U.S. department store executive who proudly proclaimed that the company’s systems alerted corporate managers instantly when a store ran out of yellow sweaters and needed inventory to be shifted from stores that were overstocked. When asked, he acknowledged that his systems could not tell him how many orange sweaters would have sold if the company had carried them. Only his salesclerks would know about orange sweater demand—and he had no formal way of collecting their insights.
The Seven-Eleven Japan approach to generating big value from little data relies on providing transparent information to decision makers and setting clear expectations for how they will use it. That is the essence of evidence-based decision making. You could design a computer model to spit out predictions of what might sell quickly, but the computer would not have data on all the requests that couldn’t be fulfilled or insights from casual conversations with customers. There would be far fewer opportunities to identify successful new-product concepts.
Most examples of evidence-based decision making we’ve seen have been in divisions and functions rather than across companies. That’s probably because it’s less daunting to improve how data are used in one unit than to do so throughout an organization. Now let’s examine the four practices.
Agree on a Single Source of Truth
The exemplary organizations we’ve studied do not necessarily have a single data repository, but they do insist on using performance data from just one authorized source. When Ron Williams became the head of operations at Aetna, in 2001 (he became president in 2002 and CEO in 2006), he found that all the divisional heads could show him a spreadsheet with performance data indicating that their divisions had been profitable the previous year—even though Aetna as a whole had recorded a loss of almost $300 million! One of his first initiatives was to mandate a single information system that defined the data everyone would use to measure performance. Senior managers saw the data as seriously flawed at first—some revenue and expense items, they believed, were inaccurately calculated or allocated—but even so they got into the habit of focusing on the metrics Williams had designated. As IT and business leaders cleaned up the data, management gained a better understanding of costs and profitability. Soon executives were creating new health plans with more-targeted pricing and working their way back to profitability. In 2005 Aetna recorded profits of $1.6 billion. In 2006, reflecting on his company’s success, Williams said, “When you have a pre-agreed set of numbers presented in a uniform way, you can train the company how to think about problems. It gives you the context for making choices.”
Getting everyone to accept the single source of data may require appointing one executive to oversee its management. At Foxtel, Australia’s largest provider of pay-TV services, CFO Peter Tonagh (now COO of News Corp Australia, one of Foxtel’s parent companies) maintained primary control over the definitions of the data in the company’s data warehouse. “There is only one source of truth in this business, and that’s what comes out of my team,” he says. Tonagh also keeps a lid on reports in order to focus everyone’s attention on what matters most. “I don’t want people thinking, How many customers have taken multiroom service?” he notes. “I want them to be thinking, How am I going to sell more multiroom services?” Tonagh’s approach has led to a significant decrease in the number of regular reports generated, down to 180 from a high of 600. That in itself has generated cost savings for Foxtel, but the greater benefit has been helping management focus on strategic objectives.
Universal acceptance of one source of truth is the first step in adopting a culture of evidence-based decision making. As both Aetna and Foxtel learned, it’s okay if the data are initially flawed, because it takes time for people to learn how to use a single source. But over time, quality matters, so companies will want to initiate processes for improving data capture. Invariably, that means reviewing business processes and identifying where mistakes enter systems. People required to use data will take an active interest in governance processes designed to clarify data definitions and in learning how information flows through the organization.
Use Scorecards
Perhaps the best way to teach people how to use data to create business benefits is to provide them with data about their own performance. Regular scorecards clarify individual accountability and provide consistent feedback so that individuals know how they are doing.
At PepsiAmericas, a $5 billion bottling company (purchased in 2010 by PepsiCo), management instituted scorecards that informed each person of his or her performance the previous day. At one warehouse, management posted the scorecards, ranking each loader’s performance on both quantity and quality. Most employees checked their rankings as they started work each day, greeting the results with either fist pumps or groans. The warehouse took on an air of friendly competition, which, coupled with new technology and powerful data, increased the accuracy of the loading process by several percentage points, to 99.8%. This approach also eliminated the need for more checkers in the warehouse.
It’s important for scorecards to be based on the right metric. In June 2010, when Tim Whall and his management team took the reins of Protection One, North America’s sixth-largest security provider, the company was enduring its fifth consecutive year of declining revenue. To turn the situation around, they set about switching managers’ attention from P&L to recurring monthly revenue (RMR), the key metric for assessing a subscription business like theirs. Within months, CIO Don Young began distributing a scorecard every day at about 4:30 AM that reported each branch and regional manager’s results in terms of changes in the prior day’s RMR. The scorecard guides Whall’s decisions about how to spend his time: which managers to call, what to ask, and what help to offer. Now his managers use the scorecard every morning to do the same thing. It takes time to change the entrenched habits of longtime employees, but the new management team has already turned around customer and employee satisfaction scores while increasing revenue by more than 10% in an industry where 3% or 4% increases are the norm.
The most important characteristic of the scorecard is that it focuses on results that individuals can control; these are not summaries of the company’s financial performance or stock price. A targeted scorecard allows the group to identify problems before they show up on the bottom line, and it helps individuals understand how their activities contribute to business success. To be sure, the metrics are more nuanced for employees at higher levels of an organization, where success on one metric (such as customer satisfaction) may come at the expense of another (negotiated price). But individuals with experience using scorecards can learn to adapt to greater ambiguity.
Explicitly Manage Your Business Rules
Little data can have a big effect on performance when managers use the data (about customers, products, transactions, and so on) to continually assess and improve the business rules that govern their operations. Business rules are the mechanism for specifying what actions should be taken in a given circumstance. They may be broad (“Do whatever it takes to make the customer happy”) or quite granular (“Accept returns from customers only if they bring a receipt and the receipt shows that they purchased the item in the past 30 days”).
Ideally, business rules align the actions of operational decision makers with the strategic objectives of the company. But that happens only when relevant individuals understand the rules and management regularly adjusts them in response to new information.
Companies with a culture of evidence-based decision making see to it that business rules are continually assessed and improved by articulating them clearly and ensuring consistency across the company. Consider Citrix Systems, a $2.1 billion technology firm that has 250,000 customers in 100 countries. Most of Citrix’s customers are served directly by one of the company’s 10,000 business partners. Citrix has traditionally offered its best partners discounts on Citrix products to encourage and reward their loyalty. But company executives found wide variation in managers’ discounting practices and increasingly observed negative impacts on revenue. So Citrix established a new companywide set of business rules that award rebates on the basis of how many Citrix product certifications (which attest to the ability to service a product) the partner firm’s employees have collectively earned. Management anticipated that these rules would optimize revenue and, by encouraging partners to earn product certifications, improve partners’ capabilities.
Having instituted new business rules, Citrix can analyze their impact. If results aren’t as anticipated, the company can change its rules again. That kind of analysis doesn’t involve the massive processing associated with big data, nor does it engage data scientists in sophisticated statistical modeling. Instead, it involves ordinary managers’ close monitoring of changes in key indicators. That is how a company uses its little data to improve performance.
Business rules become complex as they become more granular: An airline’s elite customers can check a bag free of charge; other customers must pay. Some tickets are refundable; others are not. Companies address the complexity of their business rules by embedding many of them in software. For example, an airline passenger’s elite status is stored electronically so that the system will calculate the accurate baggage fee. Retailers can store customers’ purchase data so that computers can check whether a given return qualifies for a refund.
Embedding business rules in software—automating them—frees people from routine decisions, allowing them to focus on activities that demand individual discretion. Citrix automated its partner certification rules so that the partners are not required to track eligibility for rebates. The system does the tracking and grants the rebates. It even has a built-in grace period for partners that temporarily fall below thresholds for rebates. Automating business rules also permits increasing granularity, because systems can deal with more details than people can. It tends to be easier to test the effects of changes in automated business rules than in rules that are not automated. (For more on the upside and the downside of automation, see the sidebar “Business Rules Are Running Your Company, and You Don’t Even Know It.”)
Use Coaching to Improve Performance
It might seem that a combination of well-defined expectations, performance data, and clearly articulated business rules would be sufficient to help people make evidence-based decisions on a daily basis. Not so! The secret sauce is continual coaching aimed at improving the performance of every individual. In fact, as far as we can tell from the companies we’ve studied, there is no point in addressing the first three requirements if you don’t commit to coaching. It’s not enough to tell people what the new rules or goals are. You have to help them shift from basing their decisions on instinct to basing them on data. With customer-facing employees, this often involves helping them realize the importance of their own behavior—teaching them that they can, for example, do more to improve customer satisfaction by watching to make sure the customer uses the product correctly and listening to what the customer has to say than by demonstrating how to use the product.
At Seven-Eleven Japan, counselors visit each of the company’s 16,000 stores twice a week, helping salesclerks learn to use data effectively. The counselors compare each individual’s hypotheses about what would sell during the prior week and what actually sold. They then discuss how that individual might improve his or her performance in the coming week. Counselor is a full-time position to which high-performing salesclerks can be promoted.
At Protection One, rather than creating a new role, senior executives decided that coaching should become the primary responsibility of all managers. Some managers caught on quickly; others took much longer. From the beginning, the CEO has coached senior executives tirelessly, explaining and re-explaining why RMR is important, what each manager’s RMR-component responsibilities are, how to read and understand scorecards, and, crucially, what an executive can do today to improve performance by month’s end. Whall also models data-driven behavior. If he hears a complaint about something, he says, “Let’s look at the data.” The company’s leaders are focused on developing the coaching skills of first-level supervisors—such as branch managers, sales managers, and call center managers—who directly affect many people’s lives but typically have little experience in motivating and teaching others. Whall has mandated monthly conversations between the managers and each of their reports. The objective of these conversations is to identify how each employee can address gaps between goals and outcomes and how the manager can help.
A Gradual Shift
In a culture of evidence-based decision making, people who perform routine work suddenly find themselves more responsible for outcomes than for the number of hours they put in. Many people need to acquire coaching skills, which will lead to new and different relationships. In most organizations it will not be possible to overlay this new culture on existing structures, roles, and processes. The change will be a disruptive one.
The temptation may be to treat this cultural shift like any other major business change initiative, starting at the top by defining and communicating goals, establishing metrics, assigning accountability, and training people. But we’ve found that it is best to begin more modestly. Although Aetna was able to start near the top of the company, many business leaders would be wise to aim lower. Pick important repetitive work that includes some discretion and some application of rules—service work is a good example. Imagine how that work would be performed if people had clear business rules and metrics, along with all the data they needed to make good decisions. Then assign coaches to those employees and coach the coaches. These early efforts may reveal misguided business rules, low-quality data, and dysfunctional metrics.
Over time, the culture can spread to many, maybe even most, roles. Much of the hype around big data focuses on getting more information and more people to analyze it. But the opportunity presented by the information economy is best tapped by getting allpeople to use data more effectively. That may seem like an expensive and risky endeavor. But it’s actually a cheap and powerful way of taking advantage of all the big—and little—data you are accumulating.
No comments:
Post a Comment