Harvard Business Review

Executive Summaries

Cover Feature

When Growth Stalls

Matthew S. Olson, Derek van Bever, and Seth Verry Reprint: R0803C

An abrupt and lasting drop in revenue growth is a crisis that can strike even the most exemplary organization. The authors' comprehensive analysis of growth in Fortune 100-size companies over the past half century revealed, in fact, that 87% of them had stalled out at least once. The record shows that if management cannot turn a company around within a few years, the odds are that it will never again see healthy top-line growth.

Fortunately, Olson, van Bever, and Verry, of the Corporate Executive Board, have uncovered and categorized the most common causes of growth stalls. The majority of these standstills are preventable because, according to the authors, they arise from management choices about strategy or organizational design; external factors, such as regulatory actions or economic downturns, account for only 13%. Four categories predominate:

Premium-position captivity. When a firm's world-class offering has won the most demanding customers in the market, it often fails to respond effectively to new, low-cost competitive challenges or shifts in customer valuation of product features.

Innovation management breakdown. Because most large corporations generate sequential product innovations, any systemic inefficiency or dysfunction in the innovation chain can cause extremely serious problems that last for years.

Premature core abandonment. Managers may conclude too quickly that a core market is saturated. Or they may incorrectly interpret operational impediments in the core business as evidence that it's time to move into new competitive terrain.

Talent bench shortfall. Insufficient capabilities--particularly at the executive level and typically in areas of acute and specialized need--will stop growth dead in its tracks.

The authors also identified a common culprit in detailed case studies of 50 stalled companies--failure to adapt the assumptions that drive company strategy to changes in the external environment. Two tools can help managers avoid growth stalls: a self-test to diagnose impending stalls and a choice of practices to explicitly identify strategic assumptions and test them for ongoing relevance. Forethought

Megaregions: The Importance of Place

Richard Florida Reprint: F0803A

The world isn't just flat. Alongside the dispersing centrifugal forces of globalization there are equally powerful centripetal forces that trigger economic concentration in a few dozen megaregions--areas like the Boston-New York-Washington corridor and the Shanghai-Nanjing-Hangzhou triangle--which account for the bulk of the globe's economic activity and innovation.

In E-Commerce, More Is More

Andreas B. Eisingerich and Tobias Kretschmer Reprint: F0803B

Most managers believe that filling their websites with a broad array of information diverts attention from their company's core offerings. A new global study, however, has revealed just the opposite: that such information increases customer engagement. The research also shows that exploiting consumers' desire for engagement is the strongest predictor of superior shareholder value for e-commerce companies.

Mining Unconscious Wisdom

Ian Ayres Reprint: F0803C

The true power in the collective wisdom of crowds isn't in people's expressed opinions--it's buried deep inside your company's database. New tools are allowing organizations to mine their data for the "unconscious wisdom" that the crowd itself may never have thought to share.

Rudeness and Its Noxious Effects Reprint: F0803D

The mere thought of being on the receiving end of verbal abuse hurts people's ability to perform complex tasks requiring creativity, flexibility, and memory recall, new research reveals.

The Hidden Risk in Cutting Retail Payroll

Zeynep Ton Reprint: F0803E

When retailers' sales slip, the biggest opportunity to boost profits comes from improving execution. To do that, research shows, managers may actually need to increase staff.

Avoid Hazardous Design Flaws

Hari Bapuji and Paul W. Beamish Reprint: F0803F

The vast majority of product safety recalls are due not to problems with Chinese manufacturing processes but to highly preventable design mistakes that Western companies keep making over and over again.

A Conversation with Jennifer Daley Reprint: F0803G

Jennifer Daley agreed to take on a dramatic overhaul of Tenet Healthcare's clinical quality under two conditions: She would work only on behalf of the patients, and she would consider every day to be her last.

Fledgling Firms Offer Hope on Health Costs

Julia Adler-Milstein and Ashish Jha Reprint: F0803H

Regional health information organizations--RHIOs--are springing up in the United States to meet a vital need: to connect the nation's disparate patient-health information systems. If RHIOs can find a viable business model, they stand to improve the quality and decrease the cost of U.S. health care dramatically.

The Best Advice I Ever Got

Kris Gopalakrishnan Reprint: F0803J

The head of Infosys Technologies talks about the power of a well-placed word of encouragement and what it taught him about the CEO's toughest challenge--motivating people.

Reviews

Featuring Here Comes Everybody: The Power of Organizing Without Organizations, by Clay Shirky. HBR Case Study

Authenticity: Is It Real or Is It Marketing?

David Weinberger Reprint: R0803A

Marty Echt, the new head of marketing at Hunsk Engines, is determined to bring the motorcycle maker back to its roots. He says it's not enough to project authenticity to customers--employees must personally subscribe to the brand's values. Should the company's CEO support Marty's "real deal" vision? Five experts comment on this fictional case study.

Bruce Weindruch, the founder and CEO of the History Factory, says that an authenticity-based campaign can be effective--but only if it's truly drawn from history. Marketers like Marty often remember their organization's past in a golden haze. Weindruch recommends exploring old engineering drawings, ads, and product photos in order to understand what customers and employees really valued back in the day.

Gillian Arnold, a consultant to luxury fashion and fine jewelry brands, thinks Marty's approach is right: People in key marketing posts must be passionate about their products and know them inside and out. She argues that the CEO needs to commit more fully to the new campaign and address the significant gap between the staff and the brand.

James H. Gilmore and B. Joseph Pine II, the cofounders of Strategic Horizons, point out that Hunsk needs to manage customers' perceptions rather than trying to be a "real company" or forming a management team whose personal interests match the brand. People purchase a product if it conforms to their self-image; that alone determines the brand's authenticity.

Glenn Brackett of Sweetgrass Rods, a maker of bamboo fly-fishing rods, says Marty seems to be one of the few people who understand Hunsk motorcycles. If employees bring blood, sweat, heart, and soul to a product, it will manifest that spirit, and customers will line up for it. Features

Timeless Leadership

A Conversation with David McCullough Reprint: R0803B

The historian David McCullough, a two-time Pulitzer Prize winner and well-known public television host, has spent his career thinking about the qualities that make a leader great. His books, including Truman, John Adams, and 1776, illustrate his conviction that even in America's darkest moments the old-fashioned virtues of optimism, hard work, and strength of character endure.

In this edited conversation with HBR senior editor Bronwyn Fryer, McCullough analyzes the strengths of American leaders past and present. Of Harry Truman he says, "He wasn't afraid to have people around him who were more accomplished than he, and that's one reason why he had the best cabinet of any president since George Washington....He knew who he was." George Washington--"a natural born leader and a man of absolute integrity"--was unusually skilled at spotting talent. Washington Roebling, who built the Brooklyn Bridge, led by example: He never asked his people to do anything he wouldn't do himself, no matter how dangerous. Franklin Roosevelt had the power of persuasion in abundance.

If McCullough were teaching a business school leadership course, he says, he would emphasize the importance of listening--of asking good questions but also noticing what people don't say; he would warn against "the insidious disease of greed"; he would encourage an ambition to excel; and he would urge young MBAs to have a sense that their work matters and to make their good conduct a standard for others.

Transforming Strategy One Customer at a Time

Richard J. Harrington and Anthony K. Tjan Reprint: R0803D

A decade ago, the Thomson Corporation, like most B2B companies, had a much better understanding of the people who purchased its newspapers, journals, and textbooks for their organizations than of the people who actually used them in their daily jobs. Facing an internet shakeup of its market, Thomson realized it needed to bridge that critical knowledge gap. The company began systematically scrutinizing its end users--in much the same way that Procter & Gamble tackles consumer research--as part of a new front-end customer strategy that would become the cornerstone of the firm's transformation.

In this article, Harrington, Thomson's CEO, and Tjan, a consultant who advised him, describe how the company adopted a user-centric mind-set--initially in the Thomson Financial division and then throughout the organization. First came a redefinition of the division's market, which was mapped not by type of purchaser but by eight end-user segments. That gave Thomson a clear view of the division's real, addressable market and of corresponding opportunities. After conducting surveys and "day in the life" observations of users, Thomson charted their entire work flow, beginning with what they were doing three minutes before and after using a product, and saw where the organization could add value. Then, through cluster and conjoint analysis, the company determined how pain points and product preferences varied among the users. With that information, Thomson was able to identify three clusters of customers in one segment and develop three categories of offerings.

Since beginning to implement this approach, Thomson has changed radically. Its revenue now comes mostly from digital, not print, products, and it generates twice the operating profit and four times the free cash flow it did 10 years ago. In a market that changes by the day, Thomson's revenue is unusually predictable and profitable.

Talent Management for the Twenty-First Century

Peter Cappelli Reprint: R0803E

Most firms have no formal programs for anticipating and fulfilling talent needs, relying on an increasingly expensive pool of outside candidates that has been shrinking since it was created from the white-collar layoffs of the 1980s. But the advice these companies are getting to solve the problem--institute large-scale internal development programs--is equally ineffective.

Internal development was the norm back in the 1950s, and every management-development practice that seems novel today was routine in those years--from executive coaching to 360-degree feedback to job rotation to high-potential programs. However, the stable business environment and captive talent pipelines in which such practices were born no longer exist. It's time for a fundamentally new approach to talent management. Fortunately, companies already have such a model, one that has been well honed over decades to anticipate and meet demand in uncertain environments: supply chain management.

Cappelli, a professor at the Wharton School, focuses on four practices in particular. First, companies should balance make-versus-buy decisions by using internal development programs to produce most--but not all--of the needed talent, filling in with outside hiring. Second, firms can reduce the risks in forecasting the demand for talent by sending smaller batches of candidates through more modularized training systems in much the same way manufacturers now employ components in just-in-time production lines. Third, companies can improve their returns on investment in development efforts by adopting novel cost-sharing programs. Fourth, they should seek to protect their investments by generating internal opportunities to encourage newly trained managers to stick with the firm. Taken together, these principles form the foundation for a new paradigm in talent management: a talent-on-demand system.

How Local Companies Keep Multinationals at Bay

Arindam K. Bhattacharya and David C. Michael Reprint: R0803F

A substantial number of local companies in emerging markets have managed to hold their own--or better--in the face of competition from global Goliaths. Bhattacharya and Michael of the Boston Consulting Group show how these domestic Davids have achieved that impressive feat. The secret is to adopt most, if not all, elements of a six-part strategy.

One, the homegrown winners customize products and services to meet local needs and initially go after economies of scope. Two, they develop business models to overcome market-specific obstacles and gain competitive advantage in the process. Three, they create or buy the latest technologies and use them effectively. Four, they find ways to benefit from low-cost labor and train workers in-house to overcome shortages of skilled employees. Five, they scale quickly by going national before regional rivals can challenge them. Six, they invest in top management talent in order to sustain rapid growth. No element on its own is groundbreaking, but in the aggregate the strategy is a potent one, as the authors illustrate with the story of Ctrip, China's largest online travel agent.

Successful as homegrown champions have been--and this article identifies 50 of them--a few multinationals, such as Yum Brands, Nokia, and Hyundai, have managed to beat the locals at their own game by using the six-part strategy. Global companies would do well to study these models of achievement and, armed with acquired wisdom, rethink their own strategies before local rivals shut them out of lucrative emerging markets.

A More Rational Approach to New-Product Development

Eric Bonabeau, Neil Bodick, and Robert W. Armstrong Reprint: R0803G

Companies often treat new-product development as a monolithic process, but it can be more rationally divided into two parts: an early stage that focuses on evaluating prospects and eliminating bad bets, and a late stage that maximizes the remaining candidates' market potential. Recognizing the value of this approach, Eli Lilly designed and piloted Chorus, an autonomous unit dedicated solely to the early stage. This article demonstrates how segmenting development in this way can speed it up and make it more cost-effective.

Two classes of decision-making errors can impede NPD, the authors say. First, managers often ignore evidence challenging their assumptions that projects will succeed. As a result, many projects go forward despite multiple red flags; some even reach the market, only to fail dramatically after their introduction. Second, companies sometimes terminate projects prematurely because people fail to conduct the right experiments to reveal products' potential.

Most companies promote both kinds of errors by focusing disproportionately on late-stage development; they lack the early, truth-seeking functions that would head such errors off. In segmented NPD, however, the early-stage organization maintains loyalty to the experiment rather than the product, whereas the late-stage organization pursues commercial success.

Chorus has significantly improved NPD efficiency and productivity at Lilly. Although the unit absorbs just one-tenth of Lilly's investment in early-stage development, it delivers a substantially greater fraction of the molecules slated for late Phase II trials--at almost twice the speed and less than a third of the cost of the standard process, sometimes shaving as much as two years off the usual development time.

Is Yours a Learning Organization?

David A. Garvin, Amy C. Edmondson, and Francesca Gino Reprint: R0803H

An organization with a strong learning culture faces the unpredictable deftly. However, a concrete method for understanding precisely how an institution learns and for identifying specific steps to help it learn better has remained elusive. A new survey instrument from professors Garvin and Edmondson of Harvard Business School and assistant professor Gino of Carnegie Mellon University allows you to ground your efforts in becoming a learning organization.

The tool's conceptual foundation is what the authors call the three building blocks of a learning organization. The first, a supportive learning environment, comprises psychological safety, appreciation of differences, openness to new ideas, and time for reflection. The second, concrete learning processes and practices, includes experimentation, information collection and analysis, and education and training. These two complementary elements are fortified by the final building block: leadership that reinforces learning.

The survey instrument enables a granular examination of all these particulars, scores each of them, and provides a framework for detailed, comparative analysis. You can make comparisons within and among your institution's functional areas, between your organization and others, and against benchmarks that the authors have derived from their surveys of hundreds of executives in many industries.

After discussing how to use their tool, the authors share the insights they acquired as they developed it. Above all, they emphasize the importance of dialogue and diagnosis as you nurture your company and its processes with the aim of becoming a learning organization. The authors' goal--and the purpose of their tool--is to help you paint an honest picture of your firm's learning culture and of the leaders who set its tone.

Radically Simple IT

David M. Upton and Bradley R. Staats Reprint: R0803J

Many managers think that developing and rolling out a major IT system is like putting up a warehouse: You build it and you're done. But that does not work for IT anymore. Taking that approach results in rigid, costly systems that are outdated from the day they are turned on. What's needed for today's businesses is IT that serves not only as a platform for existing operations but also as a launchpad for new functions and businesses.

In this article, the authors present a path-based approach that addresses the primary challenges of IT: the difficulty and expense of mapping out all requirements before a project starts because people often cannot specify everything that they need beforehand; the other unanticipated needs that almost always arise once a system is in operation; and the tricky task of persuading people to use and "own" it.

Japan's Shinsei Bank emerged during the authors' research as a standout among the companies applying the path-based method. The firm designed, built, and rolled out its system by forging together, not just aligning, business and IT strategies; employing the simplest possible technology; making the system truly modular; letting it sell itself to users; and ensuring that users influence future improvements. Some of the principles are variations on old themes, while others turn the conventional wisdom on its head.

When Growth Stalls

Successful companies lose momentum for four main reasons. All are within management's control if spotted in time.

by Matthew S. Olson, Derek van Bever, and Seth Verry

cD- VIDEO: Watch the author interview and go to an online diagnostic tool.

Senior management at Levi Strauss & Company could be forgiven for not seeing it coming. The year was 1996. The company had just achieved a personal best, with sales cresting $7 billion for the first time in its history. This performance extended a run of growth in which overall revenue had more than doubled within a decade. Since taking the company private in 1985, management had relaunched the flagship 501 brand, introduced the Dockers line of khaki pants, and increased international sales from 23% to 38% of revenue and more than 50% of profits. Growth in 1995 was the strongest it had been in recent years.

And then came the stall. From that high-water mark of 1996, company sales went into free fall. Year-end revenue results for 2000 were $4.6 billion--a 35% decline from four years prior. Market value declined even more precipitously: Analysts estimate that it went from $14 billion to $8 billion in those four years. The company's share of its core U.S. jeans market dropped by half over the 1990s, falling from 31% in 1990 to 14% by decade's end. Today, with a new management team in place, Levi Strauss has undergone a companywide transformation. It may be regaining its footing, but it has yet to return to growth.

While more dramatic than many, this is the story of a revenue growth stall--a crisis that can hit even the most exemplary organizations. It shares many elements with other stalls, at companies as varied as 3M, Apple, Banc One, Caterpillar, Daimler-Benz, Toys "R" Us, and Volvo. What these companies would surely recognize in the story is the stall's suddenness. Like Levi Strauss, most organizations actually accelerate into a stall, experiencing unprecedented progress along key measures just before growth rates tumble. When the momentum is lost, it's as if the props have been knocked out from under their corporate strategy. (See the exhibit "No Soft Landings.") Typically, few on the senior team see the stall coming; core performance metrics often fail to register trouble on the horizon.

Sidebar Icon No Soft Landings (Located at the end of this article)

As part of our ongoing research into growth, the Corporate Executive Board recently completed a comprehensive analysis of the growth experiences of some 500 leading corporations in the past half century, focusing particularly on "stall points"--our term for the start of secular reversals in company growth fortunes, as opposed to quarterly stumbles or temporary corrections. The companies in our study included more than 400 that have appeared on the Fortune 100 since that index was created, some 50 years ago, along with about 90 non-U.S. companies of a similar size. The study revealed patterns in the incidence, costs, and root causes of growth stalls. (Our research approach is described briefly in the sidebar "The Search for Stall Points.")

Sidebar Icon The Search for Stall Points (Located at the end of this article)

On the quantitative record alone, we can attest that Levi Strauss is in good company: 87% of the companies in this group have suffered one or more stall points. We can also appreciate the consequences of such events. On average, companies lose 74% of their market capitalization, as measured against the S&P 500 index, in the decade surrounding a growth stall. More often than not, the CEO and senior team are replaced in its aftermath. And unless management is able to diagnose the causes of a stall and get the company back on track quickly--turning it around in a matter of several years--the odds are against its ever returning to healthy top-line growth.

Deeper analysis sheds light on the most common causes of growth stalls, which turn out to be preventable for the most part. There is a common assumption that when the fortunes of great companies plunge, it must be owing to big, external forces--economic meltdowns, acts of God, or government rulings--for which management cannot be held accountable. In fact most stalls occur for reasons that are both knowable and addressable at the time. The exhibit "The Root Causes of Revenue Stalls" reveals the factors that lay behind the stalls of 50 companies we went on to study in depth; clearly, a company can falter in many ways. One might almost think that sustaining growth in a very large company depends on doing absolutely everything right. But the root causes of stalls are not so varied or complex that we can't see patterns.

Sidebar Icon The Root Causes of Revenue Stalls (Located at the end of this article)

What the exhibit demonstrates is that the vast majority of stall factors result from a choice about strategy or organizational design. They are, in other words, controllable by management. Further, even within this broad realm, nearly half of all root causes fall into one of four categories: premium-position captivity, innovation management breakdown, premature core abandonment, and talent bench shortfall.

In this article we'll offer advice for avoiding these hazards, drawing from practices currently in use at large, high-growth companies to foresee possible stalls and head them off. More generally we will explore why management is so often blindsided by these events. As we will show, a large number of global companies may at this moment be perilously close to their own stall points. Knowing how to avoid growth stalls begins with understanding their causes. Let's look at each of the four categories.

When a Premium Position Backfires

By far the largest category of factors responsible for serious revenue stalls is what we have labeled premium-position captivity: the inability of a firm to respond effectively to new, low-cost competitive challenges or to a significant shift in customer valuation of product features.

We use the term "captivity" because it suggests how management teams can be hemmed in by a long history of success. A company that solidly occupies a premium market position remains insulated longer than its competitors against evolution in the external environment. It has less reason to doubt its business model, which has historically provided a competitive advantage, and once it perceives the crisis, it changes too little too late. When the towering strengths of a firm are transformed into towering weaknesses, it's a cruel reversal.

Readers will recognize the intellectual kinship between our notion of premium-position captivity and the patterns of technology disruption described by Clayton M. Christensen in his landmark book The Innovator's Dilemma (Harvard Business School Press, 1997). As we scan the broad data set of the Fortune 100 over the past half century, we are struck by Christensen's acumen. In documenting premium-position captivity in leading enterprises, we saw a cycle of disdain, denial, and rationalization that kept many management teams from responding meaningfully to market changes.

Price and quality leaders such as Eastman Kodak and Caterpillar, for example, have found themselves unable (or unwilling) to formulate a timely, effective response to the threat posed by foreign entrants. The owners of iconic brands, such as American Express, Heinz, and Procter & Gamble, may assume that the decades-long investments they have made in their brands will protect their premium prices against lower-cost entrants. Both Compaq and Philip Morris (now part of Altria) failed to respond to signs of trouble in the early 1990s because they relied on performance metrics designed around generous margins.

We saw premium-position captivity at work in the Levi Strauss stall when the company failed to spot a strategic inflection in customer demand. In cases like this one, organizations and their multiple sophisticated market-sensing activities simply don't recognize the importance of an emerging behavior or customer preference in their core markets. They continue to place their bets on product or service attributes that are in decline, while disruptive entrants emphasizing different, underrecognized features gain ground.

In the early 1990s Levi Strauss enjoyed surging revenues even as its relationships with the Gap and other distributors faltered and as designers and retailers introduced jeans products at the high and low ends of the market. The rise of house brands and superpremium designer jeans looked manageable--or ignorable--as long as healthy revenue growth continued. By the time the growth stall had become evident, the company found itself with an expensive retailing strategy and a product line that was out of step with both ends of the denim jeans market.

The market data relating to this growth stall were not hidden from Levi Strauss executives; the challenge was to separate the signal from the noise. The company's years of success warped its interpretation of what it was seeing. Its story illustrates how difficult it is to respond to a threat in the absence of a burning platform: If your sales are continuing to rise, how do you focus concern? In 1999 Gordon Shank, then the company's chief marketing officer, admitted ruefully, "We didn't read the signs that all was not well. Or we were in denial."

Although the onset of premium-position captivity is gradual, there are often clues that trouble is afoot, both in the external market and in executive attitudes and behaviors. (See the sidebar "When Does a Premium Position Become a Trap?") Easiest to spot in marketing data are pockets of rapid market share loss, particularly in narrow customer segments, and increasing resistance among key customers to solutions wrappers and other bundling of services. It can also be revealing to focus on metrics different from those you ordinarily emphasize. If you normally track profit per customer, for example, you are content when it rises. But would you notice if customer acquisition costs increased even more rapidly? When it comes to management attitudes, your ears may pick up the strongest clues: Listen closely to the tone in the executive suite when conversation turns to upstart competitors or to successful rivals that are viewed as less capable. Is it acceptable, or routine, to dismiss them as unworthy? Do your processes for gathering intelligence about your competitors ignore some of these market participants because of their size or perceived lack of quality? Indulging in such behavior is common, but it's a luxury that no market leader can afford.

Sidebar Icon When Does a Premium Position Become a Trap? (Located at the end of this article)

When Innovation Management Breaks Down

The second most frequent cause of growth stalls is what we call innovation management breakdown: some chronic problem in managing the internal business processes for updating existing products and services and creating new ones. We saw manifestations of this at every major stage along the activity chain of product innovation, from basic research and development to product commercialization.

Where revenue growth stalls could be attributed to innovation breakdown, the problems emphatically did not center on individual product launch failures; a New Coke may occasionally belly flop, but the result is typically a temporary growth stumble rather than a fateful turning point in a company's growth history. By contrast, the secular growth stalls we identified were attributable to systemic inefficiencies or dysfunctions. Given that most large corporations rely on business models that have evolved to generate sequential product innovations, when things go wrong here--at the heart of these organizations' most important business process--extremely serious, multiyear problems result.

For firms shifting the bulk of their R&D activities out to their business units, our case studies provide a strong cautionary tale. The logic behind such shifts is clear: The closer R&D is to markets and individual unit strategies, the higher its return on investment should be. But problems seem to arise when decentralization is combined with an explicit (or implicit) metric that demands a high share of revenue growth from new-product introductions. The result can be an overallocation of resources to ever smaller incremental product opportunities, at the expense of sustained R&D investment in larger, future product platforms.

A stark example of this occurred at 3M in the 1970s, when the company experienced a revenue stall after decades of robust top-line growth. Since its founding, in 1902, 3M had followed a clear formula for success, developing innovative products with industrial applications that supported a premium position and then leapfrogging to the next opportunity as the market matured. This strategy, which has been characterized as "the corporate millipede" ("Make a little, sell a little, make a little more"), had by the early 1970s produced a portfolio of more than 60,000 products (the majority of them with sales under $100 million), while more than 25% of total corporate sales came from products less than five years old.

The growth potential inherent in this niche-jumping strategy began to dwindle in the late 1970s, as the firm approached $5 billion in revenue. With the recession of the early 1980s looming, 3M management decided to hold R&D expenditures below historical averages of just over 6% of annual sales and to push most of the R&D budget down to the company's 42 divisions (usually organized around individual product lines).

Total growth slowed as divisions focused on ever narrower niche-segment opportunities. From 1979 to 1982 the company saw its annual growth rate fall from 17% to just over 1%, with sales per employee creeping downward simultaneously. Because the bulk of R&D was controlled by product-centric business units, major new-product development activity was replaced by incremental product line extensions. The former CEO Allen F. Jacobson observed of that era, "Historically, our drive for profit and our preference for developing premium-priced products aimed at market niches meant that we were not comfortable competing only on price. As a result, we never fully developed our manufacturing competencies. And when competitors followed us, we would refuse to confront them--it was always easier to innovate our way into a new niche."

As we looked at the variety of ways in which problems in the innovation management process can eventually produce major revenue stalls, we were struck by the fragility of this chain of activities, and by how vulnerable the whole process is to management decisions made to achieve perfectly valid corporate goals. There are some powerful clues, however, when a company is at serious risk. Most significant is probably not the overall level of R&D spending but how those dollars are being spent. Is the senior team able to look into funding decisions at the business unit level to monitor the balance between incremental and next-generation investments? Are R&D and other innovation resources at the corporate level budgeted separately from incremental innovation? Is some portion of innovation funding allocated to creating lower-cost versions of existing products and services? Given the long lead times characteristic of the innovation process, flaws are slow to surface--and time-consuming to remedy.

When a Core Business Is Abandoned

The third major cause of revenue stalls is premature core abandonment: the failure to fully exploit growth opportunities in the existing core business. Its telltale markers are acquisitions or growth initiatives in areas relatively distant from existing customers, products, and channels.

This category has received significant attention in the recent business literature. Perhaps as a result, stalls attributed to premature core abandonment cluster in the period before 1990. We are tempted to credit the management consulting industry for having hammered home the need for attention to core businesses. In particular, Chris Zook, of Bain & Company, has stayed on this issue with ferocity.

That is not to say that Fortune 100-size firms have mastered the art of generating continuous growth in their core businesses. Quite the contrary: The recent wave of private equity takeovers suggests that many public companies still struggle in their efforts to grow established businesses. Almost without exception, these take-overs are based on strategies for growing the core--strategies that public-company executive teams are either unable or unwilling to pursue.

The two most common mistakes we saw in this category were believing that one's core markets are saturated and viewing operational impediments in the core business model as a signal to move on to new, presumably easier competitive terrain. Either situation invariably ended badly, with some competitor moving in to displace the incumbent.

In the late 1960s Robert Sarnoff, the CEO of RCA and son of David Sarnoff, the legendary force behind the company, came to the mistaken belief that "the age of the big breakthroughs in consumer electronics--the age in which

• had built RCA--had passed." James Hillier, the head of the company's labs, asserted, "The physicists have discovered about all they are going to for consumer application in the near future."

One can hardly blame Sarnoff when even the physicists were advocating moving on--and move on he did. He pursued initiatives in three new, presumably higher-growth directions. First, mainframe computers seemed a logical choice, given that technology-driven big bets had powered RCA's growth since the 1920s. Second, he decided that marketing was the future and deployed huge resources to acquire companies in the consumer products sector. Third, the company redirected internal resources from consumer electronics research into marketing and brand management projects. Meanwhile, Steve Jobs and Bill Gates were on the road to starting companies that would launch a revolution in RCA's former core markets.

Just as interesting as getting it wrong on core business growth prospects is the tendency of executive teams to simply give up on apparently intractable problems in their core businesses. The most intriguing example of this occurred at Kmart. A highly successful challenger to Sears as a general-merchandise big-box retailer, Kmart relentlessly stole its formerly indomitable competitor's market share through the 1960s and 1970s.

In 1976 Kmart reached a peak in new store openings, adding 271 facilities to its countrywide network. That would prove to be its limit. Over the next decade the company reined in expansion in its core business, convinced that the U.S. market was saturated. Its chairman, Robert Dewar, created a special strategy group whose purpose was to study new growth avenues and, in the parlance of the time, far-out ideas. He also established a performance goal for the company: 25% of sales should come from new ventures by 1990.

What's most disturbing about Kmart's choices is not that management was tempted to diversify in search of growth--however misguided this appears in hindsight, given Wal-Mart's concurrent gathering of strength. Rather, it is that the executive team failed to monitor and match the distribution and inventory management capabilities that its rival was pioneering in Bentonville, Arkansas. In the early 1980s, while Wal-Mart was installing its first point-of-service system with a satellite link for automatic reorders, Kmart was acquiring Furr's Cafeterias of Texas, the Bishop's Buffet chain, and pizza-video parlors as outlets for its retained earnings. Throughout the next decade Wal-Mart continued to invest in its cross-docking distribution system, while Kmart pursued a range of disparate businesses, including PayLess Drug Stores, the Sports Authority, and OfficeMax. By the end of the 1980s Kmart was at least 10 years behind Wal-Mart in its logistical capabilities, handing Wal-Mart a "gimme" advantage of more than 1% of sales in inbound logistics costs. As Kmart lagged ever further behind, its imagined need for outside-the-core growth platforms became real.

Of all the red flags signaling stall risk, one of the most obvious is management's use of the term "mature" to refer to any of its product lines, business units, or divisions. (The disinvestment in the core implied by the "cash cow" cell of the growth-share matrix does modern managers no favor.) Established businesses should be managed against significant revenue and earnings goals, and business leaders should actively explore the potential of new business models to rejuvenate even the most "mature" businesses.

When Talent Comes Up Short

Our fourth major category is talent bench shortfall: a lack of leaders and staff with the skills and capabilities required for strategy execution.

Talent bench shortfall merits careful definition, because it has become a fact of daily life in many industries and functions. Indeed, at this writing, shortages of critical talent are the primary concern of human resources departments globally, not just in high-growth markets but in a range of specialty skill categories, and they are expected to get worse. What stops growth dead in its tracks, however, is not merely a shortage of talent but the absence of required capabilities--such as solutions-selling skills or consumer-marketing expertise--in key areas of a company, most visibly at the executive level.

Internal skill gaps are often self-inflicted wounds, the unintended consequence of promote-from-within policies that have been too strictly applied. Such policies, often most fervent in organizations with strong cultures, can accelerate growth in the heady early days of executing a successful business model. But when the external environment presents novel challenges, or competition intensifies, these policies may be a severe drag on progress.

One important element in this category is a narrow experience base at the senior executive level that prevents a timely response to emerging strategic issues. The most common marker of this lack of experience is managers' tendency to follow a well-worn internal path from a dominant business, market, or function to the executive suite. Hitachi, which went into a growth stall in 1994, illustrates this problem. At the time, Hitachi accounted for 2% of Japan's GNP and 6% of its corporate R&D spending. The downward slide in the company's revenue was devastating. Executive management has consistently come up from the energy and industrial side of the company, but Hitachi's growth prospects lie elsewhere. This narrowness extends to functional pedigree: The firm has historically had an engineering culture, with none of its top executives holding an MBA or other business degree. As Hitachi looks toward its centennial in 2010, however, change may be in the offing: Kazuo Furukawa, who was named president and chief operating officer in 2006, came up through the telecom and information systems sectors. He is the company's first president with no exposure to its heavy electrical machinery business.

Few companies formally monitor the balance in the executive team between company lifers and newer hires who offer fresh perspectives and approaches. Furthermore, large companies have a fairly poor track record on incorporating new voices into senior management. Most studies agree that 35% to 40% of senior hires wash out within their first 18 months--a statistic that is improving glacially as we adopt new practices in talent management. And management development programs all too often focus on replicating the skill sets of the current leadership, rather than on developing the novel skills and perspectives that tomorrow's leaders will need to overcome evolving challenges.

We have identified a simple way to ensure balance in the senior executive ranks--what we call mix management. Our analysis of company growth rates and senior leaders' backgrounds suggests that the sweet spot for external talent is somewhere between 10% and 30% of senior management. That is a good target for the CEO and the board to use with the firm's executive committee and for human resources to use with the top 5% of the workforce.

When What You Know Is No Longer So

As noted, the four categories we have outlined account for nearly half of all the root causes we cataloged. A host of other, less common causes that came up in our analysis crossed a broad terrain, including failed acquisitions, key customer dependency, strategic diffusion, adjacency failures, and voluntary growth slowdowns. A powerful observation can be distilled from this array: One culprit in all our case studies was management's failure to bring the underlying assumptions that drive company strategy into line with changes in the external environment--whether because of a lack of awareness that the gap existed or was widening, or because of faulty prioritization.

The lack of awareness is particularly vexing, because it is so insidious. Strategic assumptions begin life as observations about customers, competitors, or technologies that arise from direct experience. They are then enshrined in the strategic plan and translated into operational guidance. Eventually they harden into orthodoxy. This explains why, when we examine individual case studies, we so often find that those assumptions the team has held the longest or the most deeply are the likeliest to be its undoing. Some beliefs have come to appear so obvious that it is no longer politic to debate them.

Part of the reason that few top teams question assumptions is that doing so goes against the nature of the senior executive mandate: The CEO and his or her executive team are paid to develop a vision and execute it--with resolve. Another part is human nature: Introspection and self-doubt don't often appear in the personality profiles of top executives at large enterprises. A third part is process: CEOs have very few opportunities to safely express their midnight anxieties. And the one opportunity for stock taking that is built into the annual calendar of most firms--the review of the strategic plan for the coming year--all too often fails to stimulate deep, searching conversation. Indeed, the "assumptions and risks" section of virtually all strategic plan templates is generally treated as a pro forma exercise rather than an occasion to go deep.

Articulating and Testing Strategic Assumptions

To assist executives in spotting signs of vulnerability to growth stalls in their own organizations, we offer two kinds of tools. The first is a diagnostic self-test we developed at the conclusion of our research. Hoping to determine how companies might foresee a stall, our team spent considerable time looking at various financial metrics, from margin erosion to patterns in R&D spending. This effort was fruitless: Financial metrics--at least those available to the public--are as likely to lag behind as lead an organization's change in strategic vitality.

What we did find helpful was asking, What could the company's senior managers have seen in their markets, in their competitors' behavior, in their own internal practices, that might have alerted them to an impending stall? We looked at our detailed case histories for warning signs before the stall point that perhaps hadn't received the scrutiny they deserved, and uncovered 50 red flags, all rooted in the real experience of the companies we studied. Our 20/20 hindsight may enable you to spot signs faster in your own organization. (See "Red Flags for Growth Stalls.")

Sidebar Icon Red Flags for Growth Stalls (Located at the end of this article)

Also included in our tool kit are four practices drawn from those we've seen management teams use. The first two are effective in making strategic assumptions explicit, and the latter two are designed to test those assumptions for ongoing relevance and accuracy. A hallmark of these practices is that they are embedded in the work flow of the firm--the job of some individual or team--or otherwise built into core operating systems.

Commission a core-belief identification squad.

This practice is simple to execute and involves calling on a diverse, cross-functional working group to go hunting for the firm's most deeply held assumptions about itself and the industry in which it operates. (Gary Hamel and his colleagues at Strategos have led the way on this practice.) The best-functioning squads include a significant share of younger, newer employees, who are less likely to be invested in current orthodoxies. Their efforts are most fruitful when the team is prepared to raise thorny issues and challenge entrenched beliefs, using methods ranging from reality checks--What industry are we in? Who are our customers?--to more provocative explorations: What 10 things would you never hear customers say about our business? Which firms have succeeded by breaking the established "rules" of the industry? What conventions did they overturn?

One leading consumer-goods company told us that it had used this practice to kick off an inquiry into long-term growth pathways and to challenge conventions that had taken hold through the years. We like the practice for two reasons. First, it seems to strike the right balance between traditional, closed-door strategy discussions and all-company "jams," which tend to lose credibility and edge in direct proportion to the number of participants involved. Second, it manages to simultaneously address areas of universal agreement and issues that are in play.

Conduct a premortem strategic analysis.

Many leaders have found it useful to charge teams with developing competing visions of the future success--or failure--of the company as it would be reported in a business periodical five years hence. (See Gary Klein, "Performing a Project Premortem," Forethought, HBR September 2007.) The process typically takes place over one or two days at regularly scheduled offsite management gatherings, and teams senior executives with high-potential staffers from around the world. By seeing which issues the scenarios have in common, leadership teams can identify the subset of core beliefs that should be most closely examined and monitored.

Appoint a shadow cabinet.

Pioneered by a Fortune 250 manufacturing company, the shadow cabinet is a standing group of high-potential employees who tend to be in midcareer and are often in line for promotion to the director level. They usually meet the day before an executive committee meeting, and their agenda matches as closely as possible the agenda for the following day, with presenters delivering dry runs of their material to the group and then providing whatever follow-up is needed to support the group's deliberations and decision making. The members of the shadow cabinet are invited to executive committee meetings on a rotating basis.

The benefits of this practice are manifold. Because it provides such powerful seasoning for the employees who participate, it becomes a mainstay of the leadership development curriculum. And because senior executives are usually most attached to the assumptions underlying current strategy (it is their strategy, after all), they find the fresh perspectives offered by this creditable, well-informed constituency extremely valuable. That said, most executives to whom we've presented this idea respond that it would never work in their organizations. "The executive agenda is too confidential," they say, or "Our executive team is too impatient," or "It looks like too much work." We agree that this practice is not for everyone; in fact, we have visited boardrooms where speaking candidly about shortcomings in company strategy would be a truly career-limiting move. Organizations where this is the case should pass on the idea. Not only will it fail to achieve the desired effect but it may cause more harm than good to the morale of staff members involved in the initiative.

Invite a venture capitalist to your strategy review.

An effective way to bring an external perspective to bear on strategy assumptions is to ask a qualified venture capitalist to sit in on business unit strategy and investment reviews and probe for potential weaknesses. The benefits for business unit managers come primarily from specific challenges but more generally from the practical, payback-focused lens that the VC brings to the review. What's more, the impact of the venture capitalist approach can live on well after the exercise. (Recording all the questions and methods the VC uses to gather information will preserve the essentials of the approach for later reuse.)

The obvious difficulty in implementing this practice is identifying an external party who is knowledgeable enough to add value to the conversation but "safe" enough to be allowed in the room. (In the current climate, representatives from the private equity community might easily meet the first requirement but miserably fail the second.) The organization that brought this idea to our attention was coventuring with a VC and so had begun to build some operating trust.

Unlike corporate investors, VCs are accustomed to serving on the boards of portfolio companies; acting in a similar capacity for a corporate partner isn't much of a stretch. For the corporate partner, however, the experience can be nothing short of eye-opening. The VC's perspective provides an in-the-moment test of assumptions about markets, customers, and competitors and brings an urgency to corporate processes that often feel routine. Deliberation around investment proposals takes on a very different tone. For a venture capitalist, each decision to fund is optional; the usual approach is to release additional funding only when meaningful milestones have been achieved. Freedom to operate for a quarter--not a year--is the norm.

Renewing Competence in Strategy

The practices we recommend in this article compete for space on an already overcrowded executive agenda. What gives force to our advocacy is that growth stalls can have dire consequences: They bring down even the most admired companies; they exact a sizable financial and human toll; and their impact may be permanent. After a stall sets in, the odds against recovery rise dramatically with the passage of time. (See the exhibit "The Long-Term Effects of Stalls.")

Sidebar Icon The Long-Term Effects of Stalls (Located at the end of this article)

Compounding this urgency, all signs point to an increasing risk of stalls in the near future. Of particular concern today is the shrinking half-life of established business models. The importance of spotting change early enough to react in time is rising exponentially. The practices we outline here create that early-warning capability. As critical, they make the strategy conversation ongoing, rather than once a quarter or once a year, and charge line managers at all levels of the firm with leading that conversation. Clay Christensen argued in these pages a decade ago that competent strategic thinking was atrophying in the executive suite because it occurred so infrequently relative to other regular activities. (See "Making Strategy: Learning by Doing," HBR November-December 1997.) As students of strategy-making in large corporations since then, we have found that the problem has only worsened.

Whatever other concerns are on the strategy agenda, guarding against growth stalls should be at the top. The tools we offer will enable the executive team to continually test the accuracy of its worldview and to flag any flawed assumptions that might trigger a stall if they go uncorrected. We know of no more powerful investment for managing controllable risk. No Soft Landings

An analysis of the growth histories of Fortune 100 and Global 100 companies that experienced stalls between 1955 and 2006 reveals this composite pattern. After a burst of energy, growth does not descend gradually; it drops like a stone.

The Search for Stall Points

To understand the prevalence of serious growth crises in large companies, as well as their costs and causes, we analyzed the experiences of more than 400 companies that have been listed on the Fortune 100 since its inception, in 1955, and of about 90 comparable non-U.S. companies. Some 500 companies over 50 years gave us 25,000 years' worth of historical data and information to mine for insights. A pattern that emerged from these histories yielded the useful construct of the stall point--that moment when a company's growth rate slips into what proves to be a prolonged decline.

We began by analyzing the revenue growth records of every company in our study to identify which companies had experienced stall points and when. Specifically, we calculated the compound annual growth rate (CAGR) of each company's revenue for 10 years before and 10 years after every year in the past half-century for which data were available. To qualify as having stalled in a given year, a company must have enjoyed compound annual growth of at least 2% in real dollars for the 10-year period prior to the potential stall point; the difference in CAGR for the 10 years preceding and the 10 years following must have been at least four percentage points; and the CAGR of the subsequent 10 years must have fallen below 6% in real dollars. One stall point identified in this manner is shown below.

We then turned our attention to why companies stall. Out of the 500 companies, we selected for in-depth case research 50 that were representative of the whole in terms of industry mix and age. We assembled comprehensive dossiers on all of them, drawing on the public record of financial reports and published materials, on case studies, and on personal interviews. This enabled us to identify the top three factors contributing to each company's growth stall. After all these analyses we were able to identify the root causes of stalls and the major categories they fell into. We arrived at our framework purely inductively, from the bottom up. (See "The Root Causes of Revenue Stalls.")

Readers may be wondering why we chose revenue rather than profit, value, or some other measure on which to focus our analysis. That is a fair question, and we considered our choice at length. It rests on two premises. The first is that revenue growth, more than any other metric, is the primary driver of long-term company performance. This is not to say that revenue growth without profits is desirable, but high growth through margin management alone is unsustainable. The second premise is more mundane: It's hard to manipulate the top line over time, and market value and profit measures are much more variable. Revenue growth guided us to the most meaningful turning points in corporate growth history.

We would be pleased to discuss any aspect of this methodology or detail of our findings with analysts wishing to learn more or to replicate our approach. We maintain an updated list of FAQs about this initiative on our website, at www.stallpoints.executiveboard.com.

One Company's Stall Point

Tracking the growth of the BF Goodrich Corporation over a 20-year period, we can clearly see its stall point. Annual growth rates are shown for a decade before and a decade after what proved to be the stall year. The turning point in Goodrich's fortunes came in 1979, after which the company's growth fell into secular decline.

The Root Causes of Revenue Stalls

A careful analysis of 50 representative companies that experienced growth stalls revealed nearly as many root causes for them: 42 external, strategic, and organizational factors, which can be grouped into categories as shown here. We identified the top three factors contributing to each company's stall and considered those results as a whole in determining how large a role (indicated by percentage) each category played. The clustering that is at the heart of our findings is clear: Four categories account for more than half the occurrences of root causes we cataloged--premium-position captivity, innovation management breakdown, premature core abandonment, and talent bench shortfall.

When Does a Premium Position Become a Trap?

At the top of every industry are companies that have built premium positions for themselves, dominating the market among the most demanding customer segments and providing products or services that lead the field in performance, thus commanding higher prices. The organizational strengths in product development, brand management, and marketing that created these top positions are sources of great pride to the firms that cultivated them.

But attack from new competitors with significantly lower cost structures, or changes in customer preferences that start slowly and then reach tipping points, can actually transform these dependable sources of competitive advantage into weaknesses. Product innovation loses its ability to protect pricing premiums, and presumed brand and marketing strengths no longer dependably protect market share. All the firm's business processes and activities, developed and honed for the top end of the market, become impediments to refreshing strategy.

It is possible to spot the onset of premium-position captivity. The six yes-or-no questions below probe awareness of threatening market dynamics, an executive team's blind spots regarding competitive threats, and intelligence capabilities for recognizing an impending encroachment on premium turf.

Clues in Market Dynamics

• Are we losing market share to nonpremium rivals in subsegments of our markets?
• Are key customers increasingly resistant to paying price premiums for product enhancements?

Clues in Executive Team Attitudes

• Does the senior executive team resist the proposition that nonpremium players operate in the same business or product category that we do?
• Do we commonly dismiss the possibility that nonpremium rivals and low-end entrants will penetrate the upper ends of our markets?

Clues in Market and Competitor Research

• Do we fail to track shifts in secondary and tertiary customer-group behavior with the same rigor we use for our higher-end segments?
• Do we exclude nonpremium players and low-end entrants from our tracking of competitive threats?

A "yes" to two or more of these questions suggests the need to refocus research into markets and competitors. The goal should be to map premium features and low-end competitor performance. A "yes" to four or more suggests an immediate need for contingency planning: How might the firm modify its current business model (including its margin requirements and cost basis) to respond to a low-cost entrant within 18 months? Red Flags for Growth Stalls

Are you about to hit a stall point? A diagnostic survey of 50 red flags can help signal the danger in time. Below is a sampling of red flags relating to premium-position captivity; other parts of the survey highlight other hazards. To the extent that your senior team and high-potential managers see these as areas for concern, you may be headed for a free fall.

• Our core assumptions about the marketplace and about the capabilities that are critical to support our strategy are not written down.
• We haven't revisited our market definition boundaries, and therefore our list of current and emerging competitors, in several years.
• We haven't refreshed our working definition of our core market, and therefore our understanding of our market share, in several years.
• We test only infrequently for shifts in key customer groups' valuation of our product/service attributes.
• We are less effective than our competitors at translating customer insights into new product and service categories.
• Core customers are increasingly unwilling to pay a premium for our brand reputation or superior performance.

To watch the authors discuss their complete list of red flags and how to use them to diagnose impending growth stalls, go to stallpoints.multimedia.hbr.org. There you can link to the full diagnostic survey, at www.stallpoints.executiveboard.com. The Long-Term Effects of Stalls

Fortune 100 and Global 100 Companies, 1955-2006

The overwhelming majority (87%) of companies in our study had experienced a stall. Fewer than half of those (46%) were able to return to moderate or high growth within the decade. When slow growth was allowed to persist for more than 10 years, the delay was most often fatal: Only 7% of the companies in that category ever returned to moderate or high growth.

Transforming Strategy One Customer at a Time

How B2B giant Thomson Corporation reinvented itself by embracing a P&G mind-set.

by Richard J. Harrington and Anthony K. Tjan

How does a business-to-business company find out exactly what end users do with its products? That was the question we wrestled with at the Thomson Corporation, because the people who buy from us are not the same people who actually use our products in their daily work. For Thomson, the answer has been to combine multiple methods of deep customer inquiry, from market surveys to observing users directly in their workplace. Those efforts have been part of a front-end customer strategy that has become the cornerstone of the company's transformation. This strategy has included asking lawyers, accountants, financial analysts, investment managers, scientific researchers, and other professionals who use our products and services what they do on a minute-by-minute basis. Then we've systematically sought to deliver solutions that meet their needs during each of those hours. By doing so, we've learned how to help end users with their work in ways that might otherwise never have occurred to us.

Sidebar Icon Getting to Know Users (Located at the end of this article)

Such scrutiny of the end user wouldn't be unusual if we were a consumer products company. P&G is known for following consumers around stores and observing them in their kitchens. But like most other B2B companies, Thomson historically had a much better understanding of its buyers than of its end users. We knew a fair amount about, say, financial services information managers, who were responsible for making purchasing decisions for an entire department, but little about the individual brokers or investment bankers who used our data, research, and other resources daily to make investment decisions for their clients.

The transformation of Thomson began a little over a decade ago. At the time, Thomson was a nearly 70-year-old holding company with $8.7 billion in revenue. We published more than 200 newspapers, along with textbooks, law books, and professional journals, and operated the largest leisure travel business in the United Kingdom. Thomson was a prosperous leader in its markets, but we were concerned about the long-term viability of our business portfolio. First, our markets were not equal in terms of growth potential. Leisure travel, for example, was becoming increasingly competitive and turning into a commodity. To realize Thomson's full potential, we needed to become less diversified and more focused on the business model with the best prospects for the future.

Second, as we looked around the corner we could see the beginnings of a radical change in market dynamics. In particular, it appeared that the rise of the internet would change the newspaper and publishing markets forever. The worth of our considerable paper assets was in jeopardy.

Today Thomson is squarely focused as a global information services company, selling to businesses and professionals in the financial, legal, tax and accounting, scientific, and health care sectors. When its proposed acquisition of Reuters obtains regulatory approval and closes, Thomson will become the largest information company in the world. While the company has had approximately the same amount of revenue as it did 10 years ago (before accounting for the sale of its learning division, which closed in the summer of 2007), the makeup and productivity of that revenue are very different. A dollar of revenue now yields approximately twice the operating profit it did back then. Thomson today has an operating profit level of approximately 20%, and free cash flow is approximately four times what it was in 1997. The company's revenue is also more stable. Eighty-three percent of it is subscription based, and renewal rates often exceed 90%. In a market that is changing by the day, this revenue is unusually repeatable, predictable, and profitable.

Over the past decade, the company has seen its market capitalization triple. The sources of Thomson's revenue have shifted dramatically as well--from print to digital. The company's electronic information products, software, and services now account for 80% of its revenue, completely the reverse of its model a decade ago.

The transformation began with the divestiture of businesses that didn't fit our strategic focus on information publishing services--and with the acquisition of professional information publishing assets that did, along with investments in the technology needed to build and deliver products and services online. The real breakthroughs, however, came a few years into the transformation process, in 2001, when we realized we needed to focus more closely on customers than ever before. These advances were driven by the changing needs of our end users and, by extension, our buyers. In this article, we'll describe Thomson's customer strategy, which combined traditional and nontraditional research methods to produce a more intimate understanding of the front-end user. Shifting gears in this way was not as premeditated or as neat as the framework we'll describe suggests. We had to learn along the way. But in the end, seeing the world through the eyes of the ultimate user led Thomson--initially at Thomson Financial and later throughout the organization--to change market definitions, evolve product development strategy, significantly modify pricing models, and even redefine who was considered the real customer.

Step 1: Map Out Your Real Market

Our first step in devising a front-end customer strategy was getting a clear picture of the real, addressable market for a given business--not the entire universe of potential customers but those whose needs we could realistically serve, given the capabilities and products we had on hand.

When we began the analysis at Thomson Financial, in 2001, we used third-party reports to build estimates of our market size, as most firms do. The traditional, or at least readily available, data split an approximately $15 billion financial-information market into three broad categories: firms on the buy side (those investing in company stock), firms on the sell side (those selling company stock), and corporate clients (those issuing stock). Those market segments were so general that they were not all that useful. They didn't help us understand where we were strongest or where we had the most opportunity to grow.

To get a clearer view, Thomson Financial recast the market by breaking it down into new segments we believed more closely reflected the ultimate users of our financial products--groups such as institutional equity advisers, fixed-income advisers, and investment bankers. We began by hypothesizing which people used our products and services most frequently. While that sounds a little basic, customer segmentations were frequently categorized more by sales channel or geography and much less so by end user. Further, where end-user information existed, it often was not granular enough to be meaningful. We had hypothesized that there were as few as four segments and as many as 12, but we felt that there were consistently well-defined differences between eight segments. (See the exhibit "A Better Way to Map the Market.")

Sidebar Icon A Better Way to Map the Market (Located at the end of this article)

Once we'd identified these eight segments, Thomson Financial worked with an external firm, the Parthenon Group, to conduct interviews with users, consult industry analysts, and probe public reports to develop estimates of our relative market share and growth potential with each group. We also tried to estimate what percentage of our individual competitors' revenue came from each segment. Eventually, we vetted these market share estimates with several industry sources. Though this exercise was time-consuming, it helped us see more practically where we had strengths and where we needed to reassess activities and resource allocations.

The figures surprised us: The near-term addressable market for our existing products and services was slightly smaller than we had previously imagined--about $13.7 billion. But when we went further and mapped our share in each segment relative to our competitors', along with the currently unaddressed opportunity, we discovered just where the untapped potential lay. Once we understood segment-specific needs, we uncovered opportunities in fast-growing areas where we had high market penetration and strong capabilities, and areas where our competition was somewhat fragmented. We also saw that in the long term, the addressable market could actually be bigger than anyone had anticipated. If we leveraged our skills and assets the right way, we could be a player in a much larger market and enjoy significant growth.

We realized we could completely redefine our target market--in much the same way that Apple Computer reframed its market. Apple originally saw itself as a computer company serving primarily users in the designer, educational, and media segments. But once Apple understood just how strong its appeal in the media segment could be, it redefined itself as a consumer- and media-centric company. Indeed, Apple recently dropped "Computer" from its name, and iPod and iTunes have had staggering success. Reframing its addressable market to a broader media definition not only expanded Apple's long-term potential but also helped strengthen its traditional computer business (estimates are its U.S. market share increased from about 6% to more than 8% in 2007). In Thomson's case, we confirmed that while the fixed-income adviser and investment manager market segments were approximately the same size, we had a far greater presence in the latter. We also discovered that we had a relatively healthy share in the corporate user segment, but we could dramatically increase our leadership position with the purchase of investor communications provider CCBN, a business with complementary strengths within this market segment.

Step 2: Understand the Customers' Objectives and Work Flow

After estimating the size of market segments from the bottom up, we explored the needs of each segment using quantitative survey methods combined with "day in the life" ethnographic research on how end users did their jobs. In this phase, we focused on acquiring a detailed understanding of the activities of the people who relied on Thomson's products every day. For instance, our Westlaw online database is purchased by law firms' central research departments, but the users are associates at those firms. What do they do with it? To find out, Thomson interviewed the associates in depth and followed them around at their jobs. In some instances, we went as far as hiring film crews to tape associates going about their work. On occasion we even visited with the users' own customers, law firm clients, because they were the ultimate beneficiaries of the Westlaw data.

When tracing end users' activities, we follow an approach called "three minutes." We combine on-the-job observation with 25 to 50 detailed interviews to learn what end users are doing three minutes before they use a product or service and three minutes after. (In areas where we have less ability to drill down deep, we might go with an hour or a longer time interval.) We then look at what they do during the next three minutes out in both directions, determining the share of mind and the share of time that Thomson has within those intervals and gradually coming to understand the entire work flow. A small but not insignificant finding: We learned that highly paid analysts were spending valuable time manually inputting Thomson Financial data into spreadsheets. So we built in a capability that allowed them to seamlessly export information to Excel. Such a change seems obvious, but if we hadn't been watching users, we wouldn't have discovered that straightforward way to add value.

The exhibit "Studying the Customers' Work Flow" shows an example of delving deeper into the daily activities of one of the segments--in this case investment managers. We developed a high-level map to describe the activities of a typical investment manager, or buy-side analyst, who would be an end user of Thomson Financial information products. (The circles at the bottom show how well Thomson and its competitors could meet the investment managers' needs at any point in the cycle.) Thomson could then identify new opportunities for these users to interact with the company over the course of their jobs.

Sidebar Icon Studying the Customers' Work Flow (Located at the end of this article)

Step 3: Develop Products That Provide What Users Value Most

Once we had taken the company through the first two steps, we saw that the market was not as simple as we had thought. The next item on our agenda was to create products to fill gaps. We needed to reexamine the company's development priorities based on what people would be willing to pay for, which is not always what people say they want.

At this stage, it was critical to determine where there were pain points in the work flow that customers would pay to ease. This step required us to employ statistical techniques like cluster and conjoint analysis and hone our ability to interpret the results.

We set out to identify and test new product attributes that might be of value to each division's end users. Thomson Financial, for instance, needed to deeply examine the work flow of buy-side investment managers and researchers to hypothesize which critical product attributes were missing or could be improved. We asked cross-functional teams--representing product development, customer service, sales, and strategy--to come up with seven to 10 attributes to test, based on the findings from steps one and two and their own experience. The list of attributes that evolved included real-time data, exporting data to spreadsheets, and portfolio analytics.

Because the team members interacted with customers in different ways, they had complementary perspectives that gave the group a fully rounded view of which attributes were the most promising. For example, customer service personnel had heard time and time again of customers' wish for seamless integration with Excel, while some of the product development team felt that portfolio analytics were more important. We assessed the relative importance of the attributes through a quantitative survey of investment managers and another set of qualitative interviews.

The aggregate response from more than 1,200 surveys is represented in disguised form in the exhibit "Two Views of Product Development Priorities" under the heading "Overall Preferences for Product Features." Most would stop their analysis here, assuming that the attributes in green represent the most critical elements to build and sell. However, there is rarely such a thing as the aggregate average customer. Think about it: Say you are surveying people about what temperature they like their tea served at, and half prefer it served hot, and half iced. The average result would be lukewarm tea, which you wouldn't be able to sell to anybody.

Sidebar Icon Two Views of Product Development Priorities (Located at the end of this article)

So our goal was to understand how preferences for attributes varied among different types of users within the investment manager segment. To get a true picture of the demand and which kinds of users had the largest impact on the survey results, we conducted a conjoint analysis. In it we asked the survey participants which potential product enhancements they'd be willing to trade off for others, in order to get a clear sense of which attributes they ranked the highest. The results varied among the different types of users, who valued some but not all of the same things. For instance, the ability to export data to spreadsheets was an important attribute for all users in the segment, but the ability to get real-time data really mattered to only about a third of that group.

Within the investment management group, we identified three clusters of customers--users who had only basic needs, users who wanted advanced functionality, and high-end users who needed the best real-time information. In the exhibit below you can see how the survey results varied among these clusters. The implication was that there should be three versions of the offering that Thomson Financial was trying to develop for investment managers: one for each cluster. That insight into how preferences differed was absolutely critical to us when we reset our product development priorities. It also led us to do differential pricing--to charge more for additional highly valued features. And it made clear to us that we needed to move faster in the real-time data cluster; previously, Thomson had mostly prioritized serving the basic and advanced clusters. Ultimately, we developed value propositions for each of the three clusters.

Once we had defined our priorities, we instituted an eight-quarter plan to create and roll out the new offerings. Their development took place in parallel, but they were launched in three phases. In the first phase we went after the low-hanging fruit, a basic product that efficiently met the needs of the low-end user. In phase two we added more features and created a product bundle for the advanced user, which was priced accordingly. In the third phase, we rolled out a new real-time product, which built upon some of the elements of the other two offerings. The overall result was a modular solution, much like Microsoft Office in the compatibility of its components, that flexibly addressed the majority of the market.

Sidebar Icon When Front-End Customer Strategy Makes Sense (Located at the end of this article)

Ongoing Implementation

Front-end customer strategies must be continually reevaluated and refined, because markets and competition can change so quickly. To execute an evolving strategy, we needed to have a flexible go-to-market plan and a well-considered approach for rolling the strategy out across segments and businesses. Most important, we needed to have customer feedback loops that were exceptionally effective. They played a critical role in helping us adjust course as the market has shifted.

Sales and go-to-market plans.

Inviting sales and product development people to participate in the research turned out to be the key to a successful rollout at Thomson. We were able to develop go-to-market plans that were practical to implement as well as strategically sound. For example, Thomson Financial's pricing strategy entailed a shift from selling products to selling tailored solutions of information products and services. Putting sales leaders on the planning teams from day one allowed us to accelerate the sales training process and ensure early buy-in from the people who would eventually represent Thomson on the street. Another core tenet of our go-to-market plans has been trial first and then rollout. The offering of each new product is staged and used as an opportunity to gather yet another level of end-user feedback, which informs subsequent rounds of product refinements.

Feedback mechanisms.

Perhaps the most important purpose of gaining feedback is to see if there are any gaps between your theory about what customers want and reality. Collecting data in real time and acting upon user feedback at each stage of trial are also vital. When Thomson Financial was launching an investment management offering called Thomson One Analytics as part of the pricing structure it implemented in 2002, it began with a trial of the product that involved 30 purchasers at investment houses, to ensure that it didn't lose touch with the gatekeepers who would eventually buy the offering. As Thomson One was officially rolled out to the market at large, the division monitored user feedback closely for the first six months. A year later Thomson Financial again collected detailed information on whether customers understood the pricing strategy, and how it affected their use of the product. The feedback gathered throughout the process confirmed that customers indeed wanted more tailored offerings and differential pricing based on the features provided, and helped the division migrate its customers to Thomson One's more customized solution.

Or consider another example: When Thomson Financial wanted to refine its newly acquired TradeWeb fixed-income trading platform, it solicited feedback not only from traders but also from the people in their back offices who supported them. The ability to integrate TradeWeb easily with back-office systems proved to be a key sales point.

The company ensures that it gets continual input into product development by giving customers incentives to fill out annual surveys and making frequent use of customer advisory groups. One side benefit of these tactics is that when customers are invited to offer feedback, their loyalty goes up--especially when they see their suggestions incorporated into product improvements.

Scaling up the process.

Every front-end initiative begins with a targeted group of customers and is closely monitored until it's ready to be rolled out to a large customer group or across product lines and business units. Transforming a front-end customer approach from a corporate initiative into an integral part of the company culture required more than that, however. At Thomson, it involved evangelism from the top down and bottom up. A key success factor was strong personal support from the CEO, which included spending time with frontline teams to develop and implement the process. It was a learning experience for all of us.

Equally important were peer testimonials from internal advocates. When the pioneers at Thomson Financial began reporting improved performance after implementing our new front-end strategy, other Thomson businesses took note. The cross-functional nature of the implementation teams helped ensure that the word spread quickly along and across functional lines as well. As the front-end strategy started to be seen as a competitive advantage, people within the company began to compete to adopt it first and benefit from its advantage most.

Once a critical mass of staff members had been involved in the front-end initiatives, these employees became the faculty for formal training courses. Led by successful practitioners, sessions on front-end customer strategy became a core component of Thomson's executive training and other learning programs. Within 18 months, the top 200 leaders had been exposed to front-end customer strategy. By year three, the top 500 had been trained in the methodology.

• o o

We have said that our process was initially not as thought-out as our framework here suggests. Indeed, our front-end customer strategy essentially developed organically. Thomson identified a pattern of successes in individual units that were focusing intensely on customers' work flow and segmentation by end user. Once we recognized the pattern, we began refining our research approaches and tied them together into a framework that we then replicated. We implemented the end-user strategy on a systematic basis, updating it regularly as the market and Thomson's own capabilities evolved.

Since our first front-end customer strategy exercise, in 2001, we've put every part of the organization that has any interaction with a customer through Thomson's internal university to learn the process. That includes employees in product development, sales and marketing, strategy and business development, customer service, and content and data acquisition, and the list continues to expand to other areas of the company. The goal is to have as many people as possible intimately understand the needs of Thomson's end users. Today a majority of Thomson's 32,000-plus employees have been taught the principles of front-end customer strategy. We estimate that nearly 70% of the products and services Thomson's businesses now offer are "nonlegacy" and have been developed through front-end strategies. In 2007 such offerings had the highest growth rates (in the double digits) and were powering the organic growth of the company.

Currently, the front-end customer framework is in its second or arguably even third iteration at Thomson. Markets keep changing; the maps and segments we illustrate in this article, which are from 2002, have evolved significantly through the years. Just as important, the more we learn about customer segments, the more gaps we uncover in our knowledge. We continually add new tools. Our early front-end strategies focused on incremental innovation, but we're now looking at opportunities for big, game-changing innovation. As Thomson embarks on another new phase with the acquisition of Reuters, we will inevitably find other lenses through which to view strategy and growth. Yet the need for Thomson--for any organization, especially a B2B one--to get as close as possible to customers and end users will always be essential to realizing our full growth potential. Getting to Know Users

Creating a front-end customer strategy at Thomson has been an ongoing process. Here's what it looked like at our Thomson Financial division:

Step 1: Map out your real market.

As recently as 2001, we were using third-party reports to estimate market size, as most firms do. The conventional wisdom split an approximately $15 billion financial-information market into three categories: firms on the buy side, firms on the sell side, and corporate clients. This framing was far too vague, so we decided to break the market down into segments of users. Identifying eight segments, we dug deeply into competitor reports, interviewed customers, and consulted analysts, and then mapped out our share in each relative to our competitors, along with the currently unaddressed opportunity, to get a clear picture of just where untapped potential lay.

Step 2: Understand the customers' objectives and work flow.

The next step was to find out exactly how our products were being used--which meant gathering information not on the activities of the bank's head of research, who bought the product, but on the behavior of analysts doing research for their clients. We used a combination of traditional survey methods and less traditional methods such as "day in the life" observations of customers to chart users' activities. Key to this research was an approach called "three minutes." What were end users of a product or service doing three minutes before they used it and three minutes after? What were they doing for the next three minutes? We kept asking that until we got a view of the full day. We wanted Thomson products to be a part of as much of that day as possible.

Step 3: Develop products that provide what users value most.

Once we had a picture of users' needs, we could start to add new features that would address those we didn't already meet. But first we had to discover the biggest pain points for end users--which aspects of their jobs were so problematic that customers would pay to make them better? When we surveyed more than 1,200 investment managers, for instance, we saw the features those users valued most in the aggregate. Then we went a step further, doing a conjoint analysis in which we asked investment managers to make trade-offs among attributes that might enhance the product. This gave us a truer picture of their preferences. We saw that within the investment manager group there were three distinct clusters of needs: basic users, advanced users, and real-time-focused users. The three clusters valued some but not all of the same things. We then concentrated our development efforts on creating three versions of our solution, each aimed at meeting the needs of one cluster.

Keep the focus on users.

At Thomson we are continually evaluating and refining our customer strategy. Implementing it requires a flexible go-to-market plan, which we enable by including sales and product development people right up front in the research; by employing effective customer feedback loops that are built into a periodic review process; and by gradually scaling up the strategy across segments and businesses. A Better Way to Map the Market

When Thomson Financial reframed its market, breaking it down by end users rather than purchasers, it saw segments where it had market penetration and opportunities for growth.

Studying the Customers' Work Flow

By tracking the activities of users in its investment manager segment, Thomson Financial was able to develop a clear picture of their work flow and their information needs at each stage. Thomson then examined how well it met those needs, relative to competitors.

Two Views of Product Development Priorities

When Thomson Financial surveyed its investment manager customers about the product features that mattered most to them, eight attributes--including real-time prices, Excel integration, and portfolio analytics--scored relatively high. Then Thomson Financial looked more closely to see what was driving each attribute's score and discovered three patterns of behavior among the investment managers. There were three distinct clusters: basic users, advanced users, and users who required real-time information. Each group had different needs and valued a different set of attributes (circled at right).

*Relative importance of the attribute to the end user

Source: Parthenon Investment Management Survey When Front-End Customer Strategy Makes Sense

If your market is experiencing discontinuity.

Regulatory changes, new technology, or extraordinary events all can transform industries and are clear signals to reevaluate the addressable market, customer needs, and company offerings. Ask whether any macroeconomic or special events might change the way customers interact with your product.

If you lack clear value propositions.

In its earlier days of developing front-end customer strategy, Thomson asked the employees of a key division to articulate its value proposition, and the results came back even more disparate and varied than anticipated. Ask a sample of 10 key employees or even 10 sales executives to write down your core value proposition, and look for inconsistencies.

If you rely too heavily on channel segmentation.

There is no single right way to segment a company's revenue base, but too often companies confuse sales channel segmentation with end-user segmentation. Segmenting sales by channels like corporate and government buyers won't uncover similarities and differences in the behavior of users in companies or government agencies--telling you, say, which are basic reference users and which do heavy analytics. Ask if you have a segmentation scheme that helps you better understand users' behavior with your products.

If you sense that you face new customer demands and competition.

Whether fact-based or gut-based, any sense that customer demand patterns are significantly changing should be a call to action. Look especially for shifts in the makeup of your total sales and in growth segments, which can often be related to new and nontraditional competition. Ask not if your growth rate is the same but if the causes of growth are the same. Ask not just if new competition is better but whether it offers a "good enough" alternative.

Talent Management for the Twenty-First Century

Every talent management process in use today was developed half a century ago. It's time for a new model.

by Peter Cappelli

Failures in talent management are an ongoing source of pain for executives in modern organizations. Over the past generation, talent management practices, especially in the United States, have by and large been dysfunctional, leading corporations to lurch from surpluses of talent to shortfalls to surpluses and back again.

At its heart, talent management is simply a matter of anticipating the need for human capital and then setting out a plan to meet it. Current responses to this challenge largely fall into two distinct--and equally ineffective--camps. The first, and by far the most common, is to do nothing: anticipate no needs at all; make no plans for addressing them (rendering the term "talent management" meaningless). This reactive approach relies overwhelmingly on outside hiring and has faltered now that the surplus of management talent has eroded. The second, common only among large, older companies, relies on complex and bureaucratic models from the 1950s for forecasting and succession planning--legacy systems that grew up in an era when business was highly predictable and that fail now because they are inaccurate and costly in a more volatile environment.

It's time for a fundamentally new approach to talent management that takes into account the great uncertainty businesses face today. Fortunately, companies already have such a model, one that has been well honed over decades to anticipate and meet demand in uncertain environments--supply chain management. By borrowing lessons from operations and supply chain research, firms can forge a new model of talent management better suited to today's realities. Before getting into the details, let's look at the context in which talent management has evolved over the past few decades and its current state.

How We Got Here

Internal development was the norm back in the 1950s, and every management development practice that seems novel today was commonplace in those years--from executive coaching to 360-degree feedback to job rotation to high-potential programs.

Except at a few very large firms, internal talent development collapsed in the 1970s because it could not address the increasing uncertainties of the marketplace. Business forecasting had failed to predict the economic downturn in that decade, and talent pipelines continued to churn under outdated assumptions of growth. The excess supply of managers, combined with no-layoff policies for white-collar workers, fed corporate bloat. The steep recession of the early 1980s then led to white-collar layoffs and the demise of lifetime employment, as restructuring cut layers of hierarchy and eliminated many practices and staffs that developed talent. After all, if the priority was to cut positions, particularly in middle management, why maintain the programs designed to fill the ranks?

The older companies like PepsiCo and GE that still invested in development became known as "academy companies": breeding grounds for talent simply by maintaining some of the practices that nearly all corporations had followed in the past. A number of such companies managed to ride out the restructurings of the 1980s with their programs intact only to succumb to cost-cutting pressures later on.

The problems faced by Unilever's Indian operations after 2000 are a case in point. Known as a model employer and talent developer since the 1950s, the organization suddenly found itself top-heavy and stuck when business declined after the 2001 recession. Its well-oiled pipeline saddled the company with 1,400 well-trained managers in 2004, up 27% from 2000, despite the fact that the demand for managers had fallen. Unilever's implicit promise to avoid layoffs meant the company had to find places for them in its other international operations or buy them out.

The alternative to traditional development, outside hiring, worked like a charm through the early 1990s, in large measure because organizations were drawing on the big pool of laid-off talent. As the economy continued to grow, however, companies increasingly recruited talent away from their competitors, creating retention problems. Watching the fruits of their labors walk out the door, employers backed even further away from investments in development. I remember a conversation with a CEO in the medical device industry about a management development program proposed by his head of human resources. The CEO dismissed the proposal by saying, "Why should we develop people when our competitors are willing to do it for us?" By the mid-1990s, virtually every major corporation asserted the goal of getting better at recruiting talent away from competitors while also getting better at retaining its own talent--a hopeful dream at the individual level, an impossibility in the aggregate.

Outside hiring hit its inevitable limit by the end of the 1990s, after the longest economic expansion in U.S. history absorbed the supply of available talent. Companies found they were attracting experienced candidates and losing experienced employees to competitors at the same rate. Outside searches became increasingly expensive, particularly when they involved headhunters, and the newcomers blocked prospects for internal promotions, aggravating retention problems. The challenge of attracting and retaining the right people went to the very top of the list of executives' business concerns, where it remains today.

The good news is that most companies are facing the challenge with a pretty clean slate: Little in the way of talent management is actually going on in them. One recent study, for example, reports that two-thirds of U.S. employers are doing no workforce planning of any kind. The bad news is that the advice companies are getting is to return to the practices of the 1950s and create long-term succession plans that attempt to map out careers years into the future--even though the stable business environment and talent pipelines in which such practices were born no longer exist.

That simply won't work. Traditional approaches to succession planning assume a multiyear development process, yet during that period, strategies, org charts, and management teams will certainly change, and the groomed successors may well leave anyway. When an important vacancy occurs, it's not unusual for companies to conclude that the candidates identified by the succession plan no longer meet the needs of the job, and they look outside. Such an outcome is worse in several ways than having no plan. First, the candidates feel betrayed--succession plans create an implicit promise. Second, investments in developing these candidates are essentially wasted. Third, most companies now have to update their succession plans every year as jobs change and individuals leave, wasting tremendous amounts of time and energy. As a practical matter, how useful is a "plan" if it has to be changed every year?

Talent management is not an end in itself. It is not about developing employees or creating succession plans, nor is it about achieving specific turnover rates or any other tactical outcome. It exists to support the organization's overall objectives, which in business essentially amount to making money. Making money requires an understanding of the costs as well as the benefits associated with talent management choices. The costs inherent to the organization-man development model were largely irrelevant in the 1950s because, in an era of lifetime employment and a culture in which job-hopping was considered a sign of failure, companies that did not develop talent in-house would not have any at all. Development practices, such as rotational job assignments, were so deeply embedded that their costs were rarely questioned (though internal accounting systems were so poor that it would have been difficult to assess the costs in any case).

That's no longer true. Today's rapid-fire changes in customers' demands and competitors' offerings, executive turnover that can easily run to 10%, and increased pressure to show a financial return for every set of business practices make the develop-from-within approach too slow and risky. And yet the hire-from-without models are too expensive and disruptive to the organization.

A New Way to Think About Talent Management

Unlike talent development, models of supply chain management have improved radically since the 1950s. No longer do companies own huge warehouses where they stockpile the components needed to assemble years' worth of products they can sell with confidence because competition is muted and demand eminently predictable. Since the 1980s, companies have instituted, and continually refined, just-in-time manufacturing processes and other supply chain innovations that allow them to anticipate shifts in demand and adapt products ever more accurately and quickly. What I am proposing is something akin to just-in-time manufacturing for the development realm: a talent-on-demand framework. If you consider for a moment, you will see how suited this model might be to talent development.

Forecasting product demand is comparable to forecasting talent needs; estimating the cheapest and fastest ways to manufacture products is the equivalent of cost-effectively developing talent; outsourcing certain aspects of manufacturing processes is like hiring outside; ensuring timely delivery relates to planning for succession events. The issues and challenges in managing an internal talent pipeline--how employees advance through development jobs and experiences--are remarkably similar to how products move through a supply chain: reducing bottlenecks that block advancement, speeding up processing time, improving forecasts to avoid mismatches.

The most innovative approaches to managing talent use four particular principles drawn from operations and supply chain management. Two of them address uncertainty on the demand side: how to balance make-versus-buy decisions and how to reduce the risks in forecasting the demand for talent. The other two address uncertainty on the supply side: how to improve the return on investment in development efforts and how to protect that investment by generating internal opportunities that encourage newly trained managers to stick with the firm.

Sidebar Icon Operations Principles Applied to Talent Management (Located at the end of this article)

Principle 1: Make and Buy to Manage Risk

Just as a lack of parts was the major concern of midcentury manufacturers, a shortfall of talent was the greatest concern of traditional management development systems of the 1950s and 1960s, when all leaders had to be homegrown. If a company did not produce enough skilled project managers, it had to push inexperienced people into new roles or give up on projects and forgo their revenue. Though forecasting was easier than it is today, it wasn't perfect, so the only way to avoid a shortfall was to deliberately overshoot talent demand projections. If the process produced an excess of talent, it was relatively easy to park people on a bench, just as one might put spare parts in a warehouse, until opportunities became available. It may sound absurd to suggest that an organization would maintain the equivalent of a human-capital supply closet, but that was extremely common in the organization-man period.

Today, a deep bench of talent has become expensive inventory. What's more, it's inventory that can walk out the door. Ambitious executives don't want to, and don't have to, sit on the bench. Worse, studies by the consulting firm Watson Wyatt show that people who have recently received training are the most likely to decamp, as they leave for opportunities to make better use of those new skills.

It still makes sense to develop talent internally where we can because it is cheaper and less disruptive. But outside hiring can be faster and more responsive. So an optimal approach would be to use a combination of the two. The challenge is to figure out how much of each to use.

To begin, we should give up on the idea that we can predict talent demand with certainty and instead own up to the fact that our forecasts, especially the long-range ones, will almost never be perfect. With the error rate on a one-year forecast of demand for an individual product hovering around 33%, and with nonstop organizational restructurings and changes in corporate strategy, the idea that we can accurately predict talent demand for an entire company several years out is a myth. Leading corporations like Capital One and Dow Chemical have abandoned long-term talent forecasts and moved toward short-term simulations: Operating executives give talent planners their best guess as to what business demands will be over the next few years; the planners use sophisticated simulation software to tell them what that will require in terms of new talent. Then they repeat the process with different assumptions to get a sense of how robust the talent predictions are. The executives often decide to adjust their business plans if the associated talent requirements are too great.

Operations managers know that an integral part of managing demand uncertainty is understanding the costs involved in over- or underestimation. But what are the costs of developing too much talent versus too little? Traditionally, workforce planners have implicitly assumed that both the costs and the risks even out: that is, if we forecast we'll need 100 computer programmers in our division next year and we end up with 10 too many or 10 too few, the downsides are the same either way.

In practice, however, that's rarely the case. And, contrary to the situation in the 1950s, the risks of overshooting are greater than those of undershooting, now that workers can leave so easily. If we undershoot, we can always hire on the outside market to make up the difference. The cost per hire will be greater, and so will the uncertainty about employees' abilities, but those costs pale in comparison to retention costs. So, given that the big costs are from overshooting, we will want to develop fewer than 100 programmers and expect to fall somewhat short, hiring on the outside market to make up the difference. If we think our estimate of 100 is reasonably accurate, then perhaps we will want to develop only 90 internally, just to make sure we don't overshoot actual demand, and then plan to hire about 10. If we think our estimate is closer to a guess, we will want to develop fewer, say 60 or so, and plan on hiring the rest outside.

Assessing the trade-offs between making and buying include an educated estimation of the following:

• How long will you need the talent? The longer the talent is needed, the easier it is to make investments in internal development pay off.
• How accurate is your forecast of the length of time you will need the talent? The less certainty about the forecast, the greater the risk and cost of internal development--and the greater the appeal of outside hires.
• Is there a hierarchy of skills and jobs that can make it possible for candidates who do not have the requisite competencies to learn them on the job, without resorting to specialized development roles or other costly investments? This is particularly likely in functional areas. The more it is so, the easier it will be to develop talent internally.
• How important is it to maintain the organization's current culture? Especially at the senior level, outside hires introduce different norms and values, changing the culture. If it is important to change the culture, then outside hiring will do that, though sometimes in unpredictable ways.

The answers to these questions may very well be different for different functional areas and jobs within the same company. For instance, lower-level jobs may be easily and cheaply filled by outsiders because the required competencies are readily available, making the costs of undershooting demand relatively modest. For more highly skilled jobs, the costs of undershooting are much higher--requiring the firm to pay for an outside search, a market premium, and perhaps also the costs related to integrating the new hires and absorbing associated risks, such as misfits.

Principle 2: Adapt to the Uncertainty in Talent Demand

If you buy all of your components in bulk and store them away in the warehouse, you are probably buying enough material to produce years of product and therefore have to forecast demand years in advance. But if you bring in small batches of components more often, you don't have to predict demand so far out. The same principle can be applied to shortening the time horizon for talent forecasts in some interesting, and surprisingly simple, ways.

Consider the problem of bringing a new class of candidates into an organization. At companies that hire directly out of college, the entire pool of candidates comes in all at once, typically in June. Let's assume they go through an orientation, spend some time in training classes, and then move into developmental roles. If the new cohort has 100 people, then the organization has to find 100 developmental roles all at once, which can be a challenge for a company under pressure, say, to cut costs or restructure.

But in fact many college graduates don't want to go directly to work after graduation. It's not that difficult to split the new group in half, taking 50 in June and the other 50 in September. Now the program only needs to find 50 roles in June and rotate the new hires through them in three months. The June cohort steps out of those roles when the September cohort steps into them. Then the organization need find only 50 permanent assignments in September for the June hires. More important, having smaller groups of candidates coming through more frequently means that forecasts of demand for these individuals can be made over shorter periods throughout their careers. Not only will those estimates be more accurate but it will be possible to better coordinate the first developmental assignments with subsequent assignments--for instance, from test engineer to engineer to senior engineer to lead engineer.

A different way to take advantage of shorter, more responsive forecasts would be to break up a long training program into discrete parts, each with its own forecast. A good place to start would be with the functionally based internal development programs that some companies still offer. These programs often address common subjects, such as general management or interpersonal skills, along with function-specific material. There is no reason that employees in all the functions couldn't go through the general training together and then specialize. What used to be a three-year functional program could become two 18-month courses. After everyone completed the first course, the organization could reforecast the demand for each functional area and allocate the candidates accordingly. Because the functional programs would be half as long, each forecast would only have to go out half as far and would be correspondingly more accurate. An added advantage is that teaching everyone the general skills together reduces redundancy in training investments.

Another risk reduction strategy that talent managers can borrow from supply chain managers is an application of the principle of portfolios. In finance, the problem with holding only one asset is that its value can fluctuate a great deal, and one's wealth varies a lot as a result, so investment advisers remind us to hold several stocks in the same portfolio. Similarly, in supply chain management it can be risky to rely on just one supplier.

For a talent-management application, consider the situation in many large and especially decentralized organizations where each division is accountable for its own profit and loss, and each maintains its own development programs. The odds that any one division will prepare the right number of managers to meet actual demand are very poor. Some will end up with a surplus, others a shortfall. If, however, all of these separate programs were consolidated into a single program, the unanticipated demand in one part of the company and an unanticipated shortfall in another would simply cancel out, just as a stock portfolio reduces the volatility of holding individual stocks. Given this, as well as the duplication of tasks and infrastructure required in decentralized programs, it is a mystery why large organizations continue to operate decentralized development programs. Some companies are in fact creating talent pools that span divisions, developing employees with broad and general competencies that could be applied to a range of jobs. The fit may be less than perfect, but these firms are finding that a little just-in-time training and coaching can help close any gaps.

Principle 3: Improve the Return on Investment in Developing Employees

When internal development was the only way to produce management talent, companies might have been forgiven for paying less attention than they should have to its costs. They may even have been right to consider their expensive development programs as an unavoidable cost of doing business. But the same dynamics that are making today's talent pool less loyal are presenting opportunities for companies to lower the costs of training employees and thereby improve the return on their investment of development dollars, as they might from any R&D effort.

Perhaps the most novel approach to this challenge is to get employees to share in the costs. Since they can cash in on their experience on the open market, employees are the main beneficiaries of their development, so it's reasonable to ask them to contribute. In the United States, legislation prevents hourly workers from having to share in the costs of any training required for their current job. There are no restrictions, however, even for hourly workers, on contributing to the costs of developmental experiences that help prepare employees for future roles.

People might share the costs by taking on learning projects voluntarily, which means doing them in addition to their normal work. Assuming that the candidates are more or less contributing their usual amount to their regular job and their pay hasn't increased, they are essentially doing these development projects for free, no small investment on their part. Pittsburgh-based PNC Financial Services is one of several companies that now offer promising employees the opportunity to volunteer for projects done with the leadership team, sometimes restricting them to ones outside their current functional area. They get access to company leaders, a broadening experience, and good professional contacts, all of which will surely help them later. But they pay for it, with their valuable time.

Employers have been more inclined to experiment with ways to improve the payoff from their development investments by retaining employees longer, or at least for some predictable period. About 20% of U.S. employers ask employees who are about to receive training or development experiences to sign a contract specifying that if they leave the business before a certain time, they will have to pay back the cost. As in the market for carbon credits, this has the effect of putting a monetary value on a previously unaccounted for cost. This practice is especially common in countries like Singapore and Malaysia: Employees often leave anyway, but typically the new employer pays off the old one.

A more interesting practice is to attempt to hang on to employees even after they leave, making relatively small investments in maintaining ties. Deloitte, for example, informs qualified former employees of important developments in the firm and pays the cost of keeping their accounting credentials up-to-date. Should these individuals want to switch jobs agai