Competing on Analytics PDF - Harvard Business Review February 2006

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/7327312 Competing on Analytics Article in Harvard Business Review · February 2006 Source: PubMed CITATIONS READS 1,034 73,103 1 author: Thomas H. Davenport Babson College 277 PUBLICATIONS 62,766 CITATIONS SEE PROFILE All content following this page was uploaded by Thomas H. Davenport on 01 August 2014. The user has requested enhancement of the downloaded file. www.hbrreprints.org Some companies have built their very businesses on their ability to collect, analyze, and Competing on act on data. Every company can learn from what these Analytics firms do. by Thomas H. Davenport Included with this full-text Harvard Business Review article: 1 Article Summary The Idea in Brief—the core idea The Idea in Practice—putting the idea to work 2 Competing on Analytics 11 Further Reading A list of related materials, with annotations to guide further exploration of the article’s ideas and applications Reprint R0601H This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics The Idea in Brief The Idea in Practice It’s virtually impossible to differentiate your- To become an analytics competitor: self from competitors based on products same way—applying metrics to compensa- alone. Your rivals sell offerings similar to Champion Analytics from the Top tion and rewards. yours. And thanks to cheap offshore labor, Acknowledge and endorse the changes in you’re hard-pressed to beat overseas com- culture, processes, and skills that analytics Hire the Right People petitors on product cost. competition will mean for much of your work- Pursue and hire analysts who possess top- force. And prepare yourself to lead an analyt- How to pull ahead of the pack? Become an notch quantitative-analysis skills, can express ics-focused organization: You will have to un- analytics competitor: Use sophisticated complex ideas in simple terms, and can inter- derstand the theory behind various data-collection technology and analysis to act productively with decision makers. This quantitative methods so you can recognize wring every last drop of value from all your combination may be difficult to find, so start their limitations. If you lack background in sta- business processes. With analytics, you dis- recruiting well before you need to fill analyst tistical methods, consult experts who under- cern not only what your customers want positions. stand your business and know how analytics but also how much they’re willing to pay can be applied to it. and what keeps them loyal. You look be- Use the Right Technology yond compensation costs to calculate your Prepare to spend significant resources on tech- Create a Single Analytics Initiative workforce’s exact contribution to your bot- nology such as customer relationship manage- Place all data-collection and analysis activities tom line. And you don’t just track existing ment (CRM) or enterprise resource planning under a common leadership, with common inventories; you also predict and prevent (ERP) systems. Present data in standard formats, technology and tools. You’ll facilitate data future inventory problems. integrate it, store it in a data warehouse, and sharing and avoid the impediments of incon- make it easily accessible to everyone. And ex- Analytics competitors seize the lead in their sistent reporting formats, data definitions, and pect to spend years gathering enough data to fields. Capital One’s analytics initiative, for standards. conduct meaningful analyses. example, has spurred at least 20% growth Example: in earnings per share every year since the Example: Procter & Gamble created a centrally man- company went public. It took Dell Computer seven years to create aged “überanalytics” group of 100 analysts COPYRIGHT © 2005 HARVARD BUSINESS SCHOOL PUBLISHING CORPORATION. ALL RIGHTS RESERVED. a database that includes 1.5 million records Make analytics part of your overarching drawn from many different functions. It ap- of all its print, radio, broadcast TV, and cable competitive strategy, and push it down to plies this critical mass of expertise to press- ads. Dell couples the database with data on decision makers at every level. You’ll arm ing cross-functional issues. For instance, sales for each region in which the ads ap- your employees with the best evidence sales and marketing analysts supply data peared (before and after their appearance). and quantitative tools for making the best on growth opportunities in existing mar- The information enables Dell to fine-tune decisions—big and small, every day. kets to supply-chain analysts, who can then its promotions for every medium—in every design more responsive supply networks. region. Focus Your Analytics Effort Channel your resources into analytics initia- tives that most directly serve your overarching competitive strategy. Harrah’s, for instance, aims much of its analytical activity at improv- ing customer loyalty, customer service, and re- lated areas such as pricing and promotions. Establish an Analytics Culture Instill a companywide respect for measuring, testing, and evaluating quantitative evidence. Urge employees to base decisions on hard facts. Gauge and reward performance the page 1 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Some companies have built their very businesses on their ability to collect, analyze, and act on data. Every company can learn from what these firms do. Competing on Analytics by Thomas H. Davenport We all know the power of the killer app. Over Organizations are competing on analytics the years, groundbreaking systems from com- not just because they can—business today is panies such as American Airlines (electronic awash in data and data crunchers—but also be- COPYRIGHT © 2005 HARVARD BUSINESS SCHOOL PUBLISHING CORPORATION. ALL RIGHTS RESERVED. reservations), Otis Elevator (predictive main- cause they should. At a time when ﬁrms in tenance), and American Hospital Supply (on- many industries offer similar products and use line ordering) have dramatically boosted their comparable technologies, business processes creators’ revenues and reputations. These her- are among the last remaining points of differ- alded—and coveted—applications amassed entiation. And analytics competitors wring and applied data in ways that upended cus- every last drop of value from those processes. tomer expectations and optimized operations So, like other companies, they know what to unprecedented degrees. They transformed products their customers want, but they also technology from a supporting tool into a stra- know what prices those customers will pay, tegic weapon. how many items each will buy in a lifetime, Companies questing for killer apps generally and what triggers will make people buy more. focus all their ﬁrepower on the one area that Like other companies, they know compensa- promises to create the greatest competitive ad- tion costs and turnover rates, but they can also vantage. But a new breed of company is up- calculate how much personnel contribute to or ping the stakes. Organizations such as Ama- detract from the bottom line and how salary zon, Harrah’s, Capital One, and the Boston Red levels relate to individuals’ performance. Like Sox have dominated their ﬁelds by deploying other companies, they know when inventories industrial-strength analytics across a wide vari- are running low, but they can also predict ety of activities. In essence, they are transform- problems with demand and supply chains, to ing their organizations into armies of killer achieve low rates of inventory and high rates apps and crunching their way to victory. of perfect orders. harvard business review january 2006 page 2 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics And analytics competitors do all those compete on quantitative turf. As one would ex- things in a coordinated way, as part of an over- pect, the transformation requires a signiﬁcant arching strategy championed by top leadership investment in technology, the accumulation of and pushed down to decision makers at every massive stores of data, and the formulation of level. Employees hired for their expertise with companywide strategies for managing the numbers or trained to recognize their impor- data. But at least as important, it requires exec- tance are armed with the best evidence and utives’ vocal, unswerving commitment and the best quantitative tools. As a result, they willingness to change the way employees make the best decisions: big and small, every think, work, and are treated. As Gary Love- day, over and over and over. man, CEO of analytics competitor Harrah’s, Although numerous organizations are em- frequently puts it, “Do we think this is true? Or bracing analytics, only a handful have achieved do we know?” this level of proﬁciency. But analytics competi- tors are the leaders in their varied ﬁelds—con- Anatomy of an Analytics sumer products, ﬁnance, retail, and travel and Competitor entertainment among them. Analytics has been One analytics competitor that’s at the top of its instrumental to Capital One, which has ex- game is Marriott International. Over the past ceeded 20% growth in earnings per share every 20 years, the corporation has honed to a science year since it became a public company. It has al- its system for establishing the optimal price for lowed Amazon to dominate online retailing and guest rooms (the key analytics process in hotels, turn a proﬁt despite enormous investments in known as revenue management). Today, its am- growth and infrastructure. In sports, the real se- bitions are far grander. Through its Total Hotel cret weapon isn’t steroids, but stats, as dramatic Optimization program, Marriott has expanded victories by the Boston Red Sox, the New En- its quantitative expertise to areas such as con- gland Patriots, and the Oakland A’s attest. ference facilities and catering, and made re- At such organizations, virtuosity with data is lated tools available over the Internet to prop- often part of the brand. Progressive makes ad- erty revenue managers and hotel owners. It has vertising hay from its detailed parsing of indi- developed systems to optimize offerings to fre- vidual insurance rates. Amazon customers can quent customers and assess the likelihood of watch the company learning about them as its those customers’ defecting to competitors. It service grows more targeted with frequent pur- has given local revenue managers the power to chases. Thanks to Michael Lewis’s best-selling override the system’s recommendations when book Moneyball, which demonstrated the certain local factors can’t be predicted (like the power of statistics in professional baseball, the large number of Hurricane Katrina evacuees ar- Oakland A’s are almost as famous for their riving in Houston). The company has even cre- geeky number crunching as they are for their ated a revenue opportunity model, which com- athletic prowess. putes actual revenues as a percentage of the To identify characteristics shared by analyt- optimal rates that could have been charged. ics competitors, I and two of my colleagues at That ﬁgure has grown from 83% to 91% as Mar- Babson College’s Working Knowledge Re- riott’s revenue-management analytics has search Center studied 32 organizations that taken root throughout the enterprise. The have made a commitment to quantitative, fact- word is out among property owners and fran- based analysis. Eleven of those organizations chisees: If you want to squeeze the most reve- we classiﬁed as full-bore analytics competitors, nue from your inventory, Marriott’s approach Thomas H. Davenport (tdavenport@ meaning top management had announced is the ticket. babson.edu) is the President’s Distin- that analytics was key to their strategies; they Clearly, organizations such as Marriott don’t guished Professor of Information Tech- had multiple initiatives under way involving behave like traditional companies. Customers nology and Management at Babson complex data and statistical analysis, and they notice the difference in every interaction; em- College in Babson Park, Massachusetts, managed analytical activity at the enterprise ployees and vendors live the difference every the director of research at Babson Exec- (not departmental) level. day. Our study found three key attributes utive Education, and a fellow at Accen- This article lays out the characteristics and among analytics competitors: ture. He is the author of Thinking for a practices of these statistical masters and de- Widespread use of modeling and optimiza- Living (Harvard Business School Press, scribes some of the very substantial changes tion. Any company can generate simple de- 2005). other companies must undergo in order to scriptive statistics about aspects of its busi- harvard business review january 2006 page 3 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics ness—average revenue per employee, for ex- tions—even those, like marketing, that have ample, or average order size. But analytics historically depended on art rather than sci- competitors look well beyond basic statistics. ence—can be improved with sophisticated These companies use predictive modeling to quantitative techniques. These organizations identify the most proﬁtable customers—plus don’t gain advantage from one killer app, but those with the greatest proﬁt potential and rather from multiple applications supporting the ones most likely to cancel their accounts. many parts of the business—and, in a few They pool data generated in-house and data cases, being rolled out for use by customers acquired from outside sources (which they an- and suppliers. alyze more deeply than do their less statisti- UPS embodies the evolution from targeted cally savvy competitors) for a comprehensive analytics user to comprehensive analytics com- understanding of their customers. They opti- petitor. Although the company is among the mize their supply chains and can thus deter- world’s most rigorous practitioners of opera- mine the impact of an unexpected constraint, tions research and industrial engineering, its ca- simulate alternatives, and route shipments pabilities were, until fairly recently, narrowly around problems. They establish prices in real focused. Today, UPS is wielding its statistical time to get the highest yield possible from skill to track the movement of packages and to each of their customer transactions. They cre- anticipate and inﬂuence the actions of peo- ate complex models of how their operational ple—assessing the likelihood of customer attri- costs relate to their ﬁnancial performance. tion and identifying sources of problems. The Leaders in analytics also use sophisticated UPS Customer Intelligence Group, for exam- experiments to measure the overall impact or ple, is able to accurately predict customer de- “lift” of intervention strategies and then apply fections by examining usage patterns and com- Employees hired for their the results to continuously improve subse- plaints. When the data point to a potential quent analyses. Capital One, for example, con- defector, a salesperson contacts that customer expertise with numbers ducts more than 30,000 experiments a year, to review and resolve the problem, dramatically with different interest rates, incentives, direct- reducing the loss of accounts. UPS still lacks the or trained to recognize mail packaging, and other variables. Its goal is breadth of initiatives of a full-bore analytics their importance are to maximize the likelihood both that potential competitor, but it is heading in that direction. customers will sign up for credit cards and that Analytics competitors treat all such activities armed with the best they will pay back Capital One. from all provenances as a single, coherent ini- evidence and the best Progressive employs similar experiments tiative, often massed under one rubric, such as using widely available insurance industry data. “information-based strategy” at Capital One or quantitative tools. As a The company deﬁnes narrow groups, or cells, “information-based customer management” at of customers: for example, motorcycle riders Barclays Bank. These programs operate not result, they make the best ages 30 and above, with college educations, just under a common label but also under decisions. credit scores over a certain level, and no acci- common leadership and with common tech- dents. For each cell, the company performs a nology and tools. In traditional companies, regression analysis to identify factors that most “business intelligence” (the term IT people use closely correlate with the losses that group en- for analytics and reporting processes and soft- genders. It then sets prices for the cells, which ware) is generally managed by departments; should enable the company to earn a proﬁt number-crunching functions select their own across a portfolio of customer groups, and uses tools, control their own data warehouses, and simulation software to test the ﬁnancial impli- train their own people. But that way, chaos cations of those hypotheses. With this ap- lies. For one thing, the proliferation of user- proach, Progressive can proﬁtably insure cus- developed spreadsheets and databases inevita- tomers in traditionally high-risk categories. bly leads to multiple versions of key indicators Other insurers reject high-risk customers out within an organization. Furthermore, research of hand, without bothering to delve more has shown that between 20% and 40% of deeply into the data (although even traditional spreadsheets contain errors; the more spread- competitors, such as Allstate, are starting to sheets ﬂoating around a company, therefore, embrace analytics as a strategy). the more fecund the breeding ground for mis- An enterprise approach. Analytics compet- takes. Analytics competitors, by contrast, ﬁeld itors understand that most business func- centralized groups to ensure that critical data harvard business review january 2006 page 4 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics and other resources are well managed and that projects. Meanwhile, masterful number different parts of the organization can share crunching has become part of the story P&G data easily, without the impediments of incon- tells to investors, the press, and the public. sistent formats, deﬁnitions, and standards. Senior executive advocates. A companywide Some analytics competitors apply the same embrace of analytics impels changes in cul- enterprise approach to people as to technol- ture, processes, behavior, and skills for many ogy. Procter & Gamble, for example, recently employees. And so, like any major transition, created a kind of überanalytics group consist- it requires leadership from executives at the ing of more than 100 analysts from such func- very top who have a passion for the quantita- tions as operations, supply chain, sales, con- tive approach. Ideally, the principal advocate sumer research, and marketing. Although is the CEO. Indeed, we found several chief ex- most of the analysts are embedded in business ecutives who have driven the shift to analytics operating units, the group is centrally man- at their companies over the past few years, in- aged. As a result of this consolidation, P&G cluding Loveman of Harrah’s, Jeff Bezos of can apply a critical mass of expertise to its Amazon, and Rich Fairbank of Capital One. most pressing issues. So, for example, sales and Before he retired from the Sara Lee Bakery marketing analysts supply data on opportuni- Group, former CEO Barry Beracha kept a sign ties for growth in existing markets to analysts on his desk that summed up his personal and who design corporate supply networks. The organizational philosophy: “In God we trust. supply chain analysts, in turn, apply their ex- All others bring data.” We did come across pertise in certain decision-analysis techniques some companies in which a single functional to such new areas as competitive intelligence. or business unit leader was trying to push ana- The group at P&G also raises the visibility of lytics throughout the organization, and a few analytical and data-based decision making were making some progress. But we found within the company. Previously, P&G’s crack that these lower-level people lacked the clout, analysts had improved business processes and the perspective, and the cross-functional scope saved the ﬁrm money; but because they were to change the culture in any meaningful way. squirreled away in dispersed domains, many CEOs leading the analytics charge require executives didn’t know what services they of- both an appreciation of and a familiarity with fered or how effective they could be. Now the subject. A background in statistics isn’t nec- those executives are more likely to tap the essary, but those leaders must understand the company’s deep pool of expertise for their theory behind various quantitative methods so Going to Bat for Stats The analysis-versus-instinct debate, a favor- James, the famous baseball statistician who rale. In his recent book, Three Nights in August, ite of political commentators during the last popularized that term. Analytic HR strategies Buzz Bissinger describes that balance: “La two U.S. presidential elections, is raging in are taking hold in European soccer as well. Russa appreciated the information generated professional sports, thanks to several popular One leading team, Italy’s A.C. Milan, uses pre- by computers. He studied the rows and the books and high-proﬁle victories. For now, dictive models from its Milan Lab research columns. But he also knew they could take you analysis seems to hold the lead. center to prevent injuries by analyzing physio- only so far in baseball, maybe even confuse Most notably, statistics are a major part of logical, orthopedic, and psychological data you with a fog of overanalysis. As far as he the selection and deployment of players. Mon- from a variety of sources. A fast-rising English knew, there was no way to quantify desire. eyball, by Michael Lewis, focuses on the use of soccer team, the Bolton Wanderers, is known And those numbers told him exactly what he analytics in player selection for the Oakland for its manager’s use of extensive data to eval- needed to know when added to twenty-four A’s—a team that wins on a shoestring. The uate players’ performance. years of managing experience.” New England Patriots, a team that devotes an Still, sports managers—like business lead- That ﬁnal sentence is the key. Whether enormous amount of attention to statistics, ers—are rarely fact-or-feeling purists. St. Louis scrutinizing someone’s performance record won three of the last four Super Bowls, and Cardinals manager Tony La Russa, for exam- or observing the expression ﬂitting across an their payroll is currently ranked 24th in the ple, brilliantly combines analytics with intu- employee’s face, leaders consult their own ex- league. The Boston Red Sox have embraced ition to decide when to substitute a charged- perience to understand the “evidence” in all “sabermetrics” (the application of analysis to up player in the batting lineup or whether to its forms. baseball), even going so far as to hire Bill hire a spark-plug personality to improve mo- harvard business review january 2006 page 5 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics that they recognize those methods’ limita- when to run with the numbers and when to tions—which factors are being weighed and run with their guts. which ones aren’t. When the CEOs need help grasping quantitative techniques, they turn to Their Sources of Strength experts who understand the business and how Analytics competitors are more than simple analytics can be applied to it. We interviewed number-crunching factories. Certainly, they several leaders who had retained such advisers, apply technology—with a mixture of brute and these executives stressed the need to ﬁnd force and ﬁnesse—to multiple business prob- someone who can explain things in plain lan- lems. But they also direct their energies to- guage and be trusted not to spin the numbers. ward ﬁnding the right focus, building the right A few CEOs we spoke with had surrounded culture, and hiring the right people to make themselves with very analytical people—pro- optimal use of the data they constantly churn. fessors, consultants, MIT graduates, and the In the end, people and strategy, as much as in- like. But that was a personal preference rather formation technology, give such organizations than a necessary practice. strength. Of course, not all decisions should be The right focus. Although analytics compet- grounded in analytics—at least not wholly so. itors encourage universal fact-based decisions, Personnel matters, in particular, are often well they must choose where to direct resource- and appropriately informed by instinct and an- intensive efforts. Generally, they pick several ecdote. More organizations are subjecting re- functions or initiatives that together serve an cruiting and hiring decisions to statistical anal- overarching strategy. Harrah’s, for example, ysis (see the sidebar “Going to Bat for Stats”). has aimed much of its analytical activity at in- But research shows that human beings can creasing customer loyalty, customer service, make quick, surprisingly accurate assessments and related areas like pricing and promotions. of personality and character based on simple UPS has broadened its focus from logistics to observations. For analytics-minded leaders, customers, in the interest of providing supe- then, the challenge boils down to knowing rior service. While such multipronged strate- THINGS YOU CAN COUNT ON Analytics competitors make expert use of statistics and modeling to improve a wide variety of functions. Copyright © 2005 Harvard Business School Publishing Corporation. All rights reserved. Here are some common applications: FUNCTION DESCRIPTION EXEMPLARS Supply chain Simulate and optimize supply chain flows; reduce Dell, Wal-Mart, Amazon inventory and stock-outs. Customer selection, Identify customers with the greatest profit potential; Harrah’s, Capital One, loyalty, and service increase likelihood that they will want the product or Barclays service offering; retain their loyalty. Pricing Identify the price that will maximize yield, or profit. Progressive, Marriott Human capital Select the best employees for particular tasks or jobs, New England Patriots, at particular compensation levels. Oakland A’s, Boston Red Sox Product and service Detect quality problems early and minimize them. Honda, Intel quality Financial Better understand the drivers of financial performance MCI, Verizon performance and the effects of nonfinancial factors. Research and Improve quality, efficacy, and, where applicable, safety Novartis, Amazon, Yahoo development of products and services. harvard business review january 2006 page 6 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics gies deﬁne analytics competitors, executives quency and carrying inventory. we interviewed warned companies against be- The right culture. Culture is a soft concept; coming too diffuse in their initiatives or losing analytics is a hard discipline. Nonetheless, an- clear sight of the business purpose behind alytics competitors must instill a company- each. wide respect for measuring, testing, and evalu- Another consideration when allocating re- ating quantitative evidence. Employees are sources is how amenable certain functions are urged to base decisions on hard facts. And they to deep analysis. There are at least seven com- know that their performance is gauged the mon targets for analytical activity, and speciﬁc same way. Human resource organizations industries may present their own (see “Things within analytics competitors are rigorous You Can Count On”). Statistical models and al- about applying metrics to compensation and gorithms that dangle the possibility of perfor- rewards. Harrah’s, for example, has made a mance breakthroughs make some prospects es- dramatic change from a rewards culture based pecially tempting. Marketing, for example, has on paternalism and tenure to one based on always been tough to quantify because it is such meticulously collected performance rooted in psychology. But now consumer prod- measurements as ﬁnancial and customer ser- ucts companies can hone their market re- vice results. Senior executives also set a consis- search using multiattribute utility theory—a tent example with their own behavior, exhibit- tool for understanding and predicting con- ing a hunger for and conﬁdence in fact and sumer behaviors and decisions. Similarly, the analysis. One exemplar of such leadership was advertising industry is adopting economet- Beracha of the Sara Lee Bakery Group, known rics—statistical techniques for measuring the to his employees as a “data dog” because he lift provided by different ads and promotions hounded them for data to support any asser- In traditional companies, over time. tion or hypothesis. The most proﬁcient analytics practitioners Not surprisingly, in an analytics culture, departments manage don’t just measure their own navels—they also there’s sometimes tension between innovative help customers and vendors measure theirs. or entrepreneurial impulses and the require- analytics —number- Wal-Mart, for example, insists that suppliers ment for evidence. Some companies place less crunching functions use its Retail Link system to monitor product emphasis on blue-sky development, in which movement by store, to plan promotions and designers or engineers chase after a gleam in select their own tools and layouts within stores, and to reduce stock-outs. someone’s eye. In these organizations, R&D, train their own people. E.&J. Gallo provides distributors with data and like other functions, is rigorously metric- analysis on retailers’ costs and pricing so they driven. At Yahoo, Progressive, and Capital One, But that way, chaos lies. can calculate the per-bottle proﬁtability for process and product changes are tested on a each of Gallo’s 95 wines. The distributors, in small scale and implemented as they are vali- turn, use that information to help retailers op- dated. That approach, well established within timize their mixes while persuading them to various academic and business disciplines (in- add shelf space for Gallo products. Procter & cluding engineering, quality management, and Gamble offers data and analysis to its retail cus- psychology), can be applied to most corporate tomers, as part of a program called Joint Value processes—even to not-so-obvious candidates, Creation, and to its suppliers to help improve like human resources and customer service. responsiveness and reduce costs. Hospital sup- HR, for example, might create proﬁles of man- plier Owens & Minor furnishes similar services, agers’ personality traits and leadership styles enabling customers and suppliers to access and and then test those managers in different situ- analyze their buying and selling data, track or- ations. It could then compare data on individu- dering patterns in search of consolidation op- als’ performance with data about personalities portunities, and move off-contract purchases to determine what traits are most important to to group contracts that include products dis- managing a project that is behind schedule, tributed by Owens & Minor and its competi- say, or helping a new group to assimilate. tors. For example, Owens & Minor might show There are, however, instances when a deci- a hospital chain’s executives how much money sion to change something or try something they could save by consolidating purchases new must be made too quickly for extensive across multiple locations or help them see the analysis, or when it’s not possible to gather data trade-offs between increasing delivery fre- beforehand. For example, even though Ama- harvard business review january 2006 page 7 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics zon’s Jeff Bezos greatly prefers to rigorously express complex ideas in simple terms and quantify users’ reactions before rolling out new have the relationship skills to interact well features, he couldn’t test the company’s search- with decision makers. One consumer products inside-the-book offering without applying it to company with a 30-person analytics group a critical mass of books (120,000, to begin looks for what it calls “PhDs with personal- with). It was also expensive to develop, and that ity”—people with expertise in math, statistics, increased the risk. In this case, Bezos trusted his and data analysis who can also speak the lan- instincts and took a ﬂier. And the feature did guage of business and help market their work prove popular when introduced. internally and sometimes externally. The head The right people. Analytical ﬁrms hire ana- of a customer analytics group at Wachovia lytical people—and like all companies that Bank describes the rapport with others his compete on talent, they pursue the best. group seeks: “We are trying to build our people When Amazon needed a new head for its glo- as part of the business team,” he explains. “We bal supply chain, for example, it recruited want them sitting at the business table, partici- Gang Yu, a professor of management science pating in a discussion of what the key issues and software entrepreneur who is one of the are, determining what information needs the world’s leading authorities on optimization businesspeople have, and recommending ac- analytics. Amazon’s business model requires tions to the business partners. We want this the company to manage a constant ﬂow of [analytics group] to be not just a general util- new products, suppliers, customers, and pro- ity, but rather an active and critical part of the motions, as well as deliver orders by promised business unit’s success.” dates. Since his arrival, Yu and his team have Of course, a combination of analytical, busi- been designing and building sophisticated ness, and relationship skills may be difﬁcult to The most proficient supply chain systems to optimize those pro- ﬁnd. When the software company SAS (a spon- cesses. And while he tosses around phrases like sor of this research, along with Intel) knows it analytics practitioners “nonstationary stochastic processes,” he’s also will need an expert in state-of-the-art business good at explaining the new approaches to Am- applications such as predictive modeling or re- don’t just measure their azon’s executives in clear business terms. cursive partitioning (a form of decision tree own navels—they also Established analytics competitors such as analysis applied to very complex data sets), it Capital One employ squadrons of analysts to begins recruiting up to 18 months before it ex- help customers and conduct quantitative experiments and, with the pects to ﬁll the position. vendors measure theirs. results in hand, design credit card and other ﬁ- In fact, analytical talent may be to the early nancial offers. These efforts call for a special- 2000s what programming talent was to the ized skill set, as you can see from this job de- late 1990s. Unfortunately, the U.S. and Euro- scription (typical for a Capital One analyst): pean labor markets aren’t exactly teeming High conceptual problem-solving and with analytically sophisticated job candidates. quantitative analytical aptitudes…Engineer- Some organizations cope by contracting work ing, financial, consulting, and/or other analyti- to countries such as India, home to many sta- cal quantitative educational/work background. tistical experts. That strategy may succeed Ability to quickly learn how to use software ap- when offshore analysts work on stand-alone plications. Experience with Excel models. Some problems. But if an iterative discussion with graduate work preferred but not required (e.g., business decision makers is required, the dis- MBA). Some experience with project manage- tance can become a major barrier. ment methodology, process improvement The right technology. Competing on ana- tools (Lean, Six Sigma), or statistics preferred. lytics means competing on technology. And Other ﬁrms hire similar kinds of people, but while the most serious competitors investigate analytics competitors have them in much the latest statistical algorithms and decision greater numbers. Capital One is currently science approaches, they also constantly mon- seeking three times as many analysts as opera- itor and push the IT frontier. The analytics tions people—hardly the common practice for group at one consumer products company a bank. “We are really a company of analysts,” went so far as to build its own supercomputer one executive there noted. “It’s the primary because it felt that commercially available job in this place.” models were inadequate for its demands. Such Good analysts must also have the ability to heroic feats usually aren’t necessary, but seri- harvard business review january 2006 page 8 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics ous analytics does require the following: all the computer maker’s print, radio, network A data strategy. Companies have invested TV, and cable ads, coupled with data on Dell many millions of dollars in systems that snatch sales for each region in which the ads ap- data from every conceivable source. Enter- peared (before and after their appearance). prise resource planning, customer relation- That information allows Dell to ﬁne-tune its ship management, point-of-sale, and other sys- promotions for every medium in every region. tems ensure that no transaction or other Business intelligence software. The term signiﬁcant exchange occurs without leaving a “business intelligence,” which ﬁrst popped up mark. But to compete on that information, in the late 1980s, encompasses a wide array of companies must present it in standard for- processes and software used to collect, ana- mats, integrate it, store it in a data warehouse, lyze, and disseminate data, all in the interests and make it easily accessible to anyone and ev- of better decision making. Business intelli- eryone. And they will need a lot of it. For ex- gence tools allow employees to extract, trans- ample, a company may spend several years ac- form, and load (or ETL, as people in the indus- cumulating data on different marketing try would say) data for analysis and then make approaches before it has gathered enough to those analyses available in reports, alerts, and reliably analyze the effectiveness of an adver- scorecards. The popularity of analytics compe- tising campaign. Dell employed DDB Matrix, tition is partly a response to the emergence of a unit of the advertising agency DDB World- integrated packages of these tools. wide, to create (over a period of seven years) a Computing hardware. The volumes of data database that includes 1.5 million records on required for analytics applications may strain the capacity of low-end computers and serv- ers. Many analytics competitors are convert- ing their hardware to 64-bit processors that You Know You Compete on Analytics churn large amounts of data quickly. When... The Long Road Ahead Most companies in most industries have excel- 1. You apply sophisticated information systems and rigorous analysis not only to lent reasons to pursue strategies shaped by an- your core capability but also to a range of functions as varied as marketing and alytics. Virtually all the organizations we iden- human resources. tiﬁed as aggressive analytics competitors are 2. Your senior executive team not only recognizes the importance of analytics capa- clear leaders in their ﬁelds, and they attribute bilities but also makes their development and maintenance a primary focus. much of their success to the masterful exploi- tation of data. Rising global competition in- 3. You treat fact-based decision making not only as a best practice but also as a part tensiﬁes the need for this sort of proﬁciency. of the culture that’s constantly emphasized and communicated by senior executives. Western companies unable to beat their In- 4. You hire not only people with analytical skills but a lot of people with the very best dian or Chinese competitors on product cost, analytical skills—and consider them a key to your success. for example, can seek the upper hand through optimized business processes. 5. You not only employ analytics in almost every function and department but also Companies just now embracing such strate- consider it so strategically important that you manage it at the enterprise level. gies, however, will ﬁnd that they take several 6. You not only are expert at number crunching but also invent proprietary metrics years to come to fruition. The organizations in for use in key business processes. our study described a long, sometimes arduous journey. The UK Consumer Cards and Loans 7. You not only use copious data and in-house analysis but also share them with cus- business within Barclays Bank, for example, tomers and suppliers. spent ﬁve years executing its plan to apply ana- 8. You not only avidly consume data but also seize every opportunity to generate in- lytics to the marketing of credit cards and formation, creating a “test and learn” culture based on numerous small experiments. other ﬁnancial products. The company had to make process changes in virtually every aspect 9. You not only have committed to competing on analytics but also have been build- of its consumer business: underwriting risk, ing your capabilities for several years. setting credit limits, servicing accounts, con- 10. You not only emphasize the importance of analytics internally but also make trolling fraud, cross selling, and so on. On the quantitative capabilities part of your company’s story, to be shared in the annual re- technical side, it had to integrate data on 10 port and in discussions with ﬁnancial analysts. million Barclaycard customers, improve the harvard business review january 2006 page 9 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics quality of the data, and build systems to step campaigns and loyalty programs.) Existing em- up data collection and analysis. In addition, the ployees, meanwhile, will require extensive train- company embarked on a long series of small ing. They need to know what data are available tests to begin learning how to attract and re- and all the ways the information can be ana- tain the best customers at the lowest price. lyzed; and they must learn to recognize such pe- And it had to hire new people with top-drawer culiarities and shortcomings as missing data, du- quantitative skills. plication, and quality problems. An analytics- Much of the time—and corresponding ex- minded executive at Procter & Gamble sug- pense—that any company takes to become an gested to me that ﬁrms should begin to keep analytics competitor will be devoted to techno- managers in their jobs for longer periods be- logical tasks: reﬁning the systems that produce cause of the time required to master quantita- transaction data, making data available in tive approaches to their businesses. warehouses, selecting and implementing ana- The German pathologist Rudolph Virchow lytic software, and assembling the hardware famously called the task of science “to stake and communications environment. And be- out the limits of the knowable.” Analytics com- cause those who don’t record history are petitors pursue a similar goal, although the doomed not to learn from it, companies that universe they seek to know is a more circum- have collected little information—or the scribed one of customer behavior, product wrong kind—will need to amass a sufﬁcient movement, employee performance, and ﬁnan- body of data to support reliable forecasting. cial reactions. Every day, advances in technol- “We’ve been collecting data for six or seven ogy and techniques give companies a better years, but it’s only become usable in the last and better handle on the critical minutiae of two or three, because we needed time and ex- their operations. perience to validate conclusions based on the The Oakland A’s aren’t the only ones playing data,” remarked a manager of customer data moneyball. Companies of every stripe want to analytics at UPS. be part of the game. And, of course, new analytics competitors will have to stock their personnel larders with Reprint R0601H fresh people. (When Gary Loveman became To order, see the next page COO, and then CEO, of Harrah’s, he brought in or call 800-988-0886 or 617-783-7500 a group of statistical experts who could design or go to www.hbrreprints.org and implement quantitatively based marketing harvard business review january 2006 page 10 This article is made available to you with compliments of SAS. Further posting, copying or distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886. Competing on Analytics Further Reading ARTICLES Diamonds in the Data Mine ture a share of the additional value your peo- by Gary Loveman ple create for customers. Harvard Business Review Countering the Biggest Risk of All May 2003 by Adrian J. Slywotzky and John Drzik Product no. 3647 Harvard Business Review Gaming giant Harrah’s CEO Loveman de- April 2005 scribes how his company uses analytics to win Product no. 977X its clientele’s devotion and supercharge reve- This article presents ideas for using analytics nues. Harrah’s acquires extensive customer in- to understand and mitigate a particularly formation through a transactional database grave strategic risk—sudden shifts in cus- that records each customer’s activity at vari- tomer tastes that redefine your industry. Miti- ous points of sale, then slices and dices the gate this risk by gathering and analyzing pro- data finely to develop strategies for encourag- prietary information to detect potential shifts. ing customers to visit Harrah’s casinos regu- And conduct fast, cheap experiments to iden- larly. It identifies core customers by calculat- tify attractive offerings for different customer ing their lifetime value, and rewards them for microsegments. For example, Coach won- spending more. Thanks to analytics, Harrah’s dered whether its customers would remain scored 16 straight quarters of same-store rev- loyal if it offered trendier styles. It conducted enue growth. in-store product tests and market experi- The Surprising Economics of a “People ments to gauge the impact of new pricing, Business” features, and offers by competitive brands. It by Felix Barber and Riner Strack used the information to quickly alter product Harvard Business Review designs, drop unappealing items, and create June 2005 new lines featuring different fabrics and col- Product no. R0506D ors. The upshot? Coach retained its traditional fans and attracted new customers. The authors explain how to use analytics to manage your company’s human resources more effectively. In “people businesses”— companies with high employee costs, low capital investment, and limited spending on activities intended to generate future reve- nue—you need to use the right metrics to assess performance. Avoid relying on capital- oriented metrics (such as return on assets or To Order return on equity); they mask weak perfor- mance or indicate market volatility where it For Harvard Business Review reprints and may not exist. Instead, use financially rigorous, subscriptions, call 800-988-0886 or people-oriented metrics—such as a reformu- 617-783-7500. Go to www.hbrreprints.org lation of a conventional calculation of eco- nomic profit—so you’re gauging people’s pro- For customized and quantity orders of ductivity. Reward excellent performance Harvard Business Review article reprints, through variable compensation schemes, and call 617-783-7626, or e-mai price products and services in ways that cap- [email protected] page 11 This article is made available to you with compliments of SAS. Further posting, copying or View publication stats distributing is copyright infringement. To order more copies go to www.hbr.org or call 800-988-0886.

Competing on Analytics PDF - Harvard Business Review February 2006

Document Details

Tags

Related

Summary

Full Transcript

Upgrade to continue