Knowledge@Wharton Human Resources Research Article

Share this Article

Failing to manage your company's talent needs, says Wharton management professor Peter Cappelli, "is the equivalent of failing to manage your supply chain." And yet the majority of employers have abysmal track records when it comes to the age-old problem of finding and retaining talent.

Supply chain managers "ask questions like, 'Do we have the right parts in stock?' 'Do we know where to get these parts when we need them?' and 'Does it cost a lot of money to carry inventory?' These questions are just as relevant to companies that are trying to manage their talent needs," he says. In other words, the principles of supply chain management, with its emphasis on just-in-time manufacturing, can be applied to talent management.

"This is a fundamentally different paradigm in terms of thinking about talent," according to Cappelli, the author of a book coming out in April titled, Talent on Demand: Managing Talent in an Age of Uncertainty. His theory, he suggests, addresses a major complaint about the field of human resources -- that it is "touchy-feely, squishy stuff with little applicability to business problems. HR practices have typically been about meeting individuals' needs, figuring out what psychological profile they fit and what should be done to help them grow and advance. But if you're an employer who is worried about issues like the finances of the company, you would like HR to think about personnel from the perspective of money and costs, and what happens if you don't have the right people in place to do the necessary jobs."

Those who study supply chain management tackle these kinds of questions all the time, notes Cappelli. "Managing supply chains is about managing uncertainty and variability. This same uncertainty exists inside companies with regard to talent development. Companies rarely know what they will be building five years out and what skills they will need to make that happen; they also don't know if the people they have in their pipelines are going to be around."

Part of the problem is that many companies are locked into an older paradigm based on the assumption that they can accurately meet their talent needs through static forecasting and planning models, even though the global marketplace is an increasingly unpredictable, unforgiving environment. "The idea that we can achieve certainty through planning is no longer true," Cappelli states. "Instead, we have to deal with uncertainty by being more responsive and adaptable."

Sitting on the Shelf

The term "talent management" simply means "trying to forecast what we are going to need, and then planning to meet that need," Cappelli notes. The definition of supply chain management is essentially the same: "We think that demand for our products next year is going to be 'X'. How do we organize internally to meet that demand?"

Underlying supply chain questions is the issue of inventory, which in talent management terms often comes up when employers talk about having a "deep bench" of talent. "You hear that phrase a lot -- 'we have a deep bench,' or 'we have a big talent pipeline' -- and it is said with pride," Cappelli says. "Yet if you think about it in supply chain terms, a deep bench is the equivalent of lots of inventory, which sounds terrible when we think of products. In fact, it is worse when we talk about talent. That's because an inventory of talent is much more costly than an inventory of widgets. Talent doesn't sit on the shelf like widgets do. You have to keep paying talent. And the best way to have a piece of talent walk away is to tell it to sit on the shelf and wait for opportunity. Anyone who is ambitious will leave, and then you will lose the big upfront investment you made in that person."

Avoiding inventory buildup directly relates to companies' efforts to manage the uncertainty around their talent needs. "Suppose a company forecasts that it will need 100 new engineers this year," says Cappelli. "No one ever asks the question: 'How accurate is that forecast?' As it turns out, that forecast is almost always wrong because business needs are so hard to predict. So the way to proceed is to ask the next question: 'What happens if we are wrong?' You can be wrong in one of two ways: You can end up needing more engineers than you thought and have to either carry them or lay them off, or fewer engineers than you thought and have to scramble to find extras. Next question: 'What does that cost us in each case? Does it cost us more if we have too many, or if we have too few?' It's almost always the case that it is much worse in one context than in the other."

If companies start thinking about what the odds are of being wrong, and what the associated costs are, "then they know which way to bet and they greatly reduce their likelihood of losing a lot of money," Cappelli says.

From there, the challenge is to reduce the odds of being wrong. That points to another technique from operations research -- the portfolio approach, whose goal is to minimize the variability that occurs when different markets are headed in different directions. In the financial world, investors create a portfolio of diverse investments where some are likely to be up when others are down in order to reduce their overall risk exposure. Applied to talent management, the concept means balancing out the kinds of errors that might occur, for example, when different divisions in a large, highly decentralized organization try to predict the number of sales people (or general managers, engineers, etc.) each division thinks it will need. "Some divisions will end up with too many sales people, and some with too few, but if you pool these different divisions with respect to hiring, it's likely the variations will cancel out rather than multiply," Cappelli says. "The problem has been that companies have decentralized so much that they stopped even thinking about how to coordinate talent questions across divisions."

As he writes in his book: "In the language of operations research and supply chain management, the problems of undersupply and oversupply are collectively known as 'mismatch costs.'" The portfolio solution addresses the mismatch problem by encouraging companies to coordinate the different talent development efforts into one common program. When some divisions overshoot demand and others undershoot it, "the company can offset the mismatch by moving candidates around."

Reducing bottlenecks is another supply chain concept relevant to talent-on-demand. The CIA had this problem when it faced a two-year waiting list to get people through security clearances, according to Cappelli. "New hires were stacked up with nothing to do, exactly the way goods can get stacked up in an assembly line. It's important to remember that the assembly line can move only as fast as the slowest part."

In the CIA's case, because it wasn't able to increase the flow of people through security, the question becomes: Why is it hiring so many people, knowing they can't get through the bottleneck? "The organization shouldn't make that many hires at once," Cappelli says. "You see this in many companies, including those that hire people only once a year, like college grads. Say they hire 50 graduates in June into training slots. At the end of the year, they have 50 people expecting to move from the training program into more permanent positions. Why doesn't the employer stagger the process and hire people twice a year instead of once? Not all college grads prefer to start work in June; some want to travel and start later in the year." The advantage of staggering the hires is that the company then needs only half the number of training positions and, more important, can adjust the amount of hiring in the latter period to changes in demand.

Other operations research practices that Cappelli relates to talent development include shortening the forecasting cycle, reorganizing the delivery of development programs to improve responsiveness, and working out "queueing problems." Queueing problems occur in situations where, for example, employees are waiting for rotational assignments but can't get them because the incumbents have no vacancies to move into -- the result of a business downturn, change in assignment length or a product redesign, for example. "The analogy in manufacturing is an 'unbalanced [assembly] line,' in which inventory builds up behind the slower-moving station, or in this case, the assignment that takes longer to complete."

Bust Instead of Boom

Many of the so-called new ideas in talent management now -- like 360-degree feedback, assessment centers, job rotation and especially long-term succession planning -- were common in the heyday of big corporations and stable growth that followed World War II, says Cappelli, who is head of Wharton's Center for Human Resources. "The current business environment bears little resemblance to the post-World War II period -- which explains why the planning-based approach no longer makes sense."

The 1970s were a case in point. Companies carefully crafted long-term plans for developing talent that turned out to be totally wrong, Cappelli says. "Everyone expected the economy to boom, and in fact it was flat. Employers turned out lots of talent they couldn't use, which led to the general abandonment of internal talent development." The early recession and re-engineering wave of the 1980s reversed that. "Companies got rid of huge numbers of employees, which meant there was no institutional memory left in the ranks. That's when people started to reinvent practices that were common in the 1950s."

In the 1990s, Cappelli says, companies turned to outside hiring, inspired in part by the large number of laid-off employees that were a hangover from the 1980s and in part by the ability to hire workers "just in time." But employers also began to hire people away from their competitors -- a game that created retention problems and usually ended in an expensive draw once more companies started to play it. In addition, organizations found that outside hiring did nothing to improve morale for those inside the company who saw new people being brought in over them.

Companies are left facing a dilemma: On the one hand, it's hard to get a payback from investing in talent development when priorities suddenly shift and employees change jobs every few years instead of once or twice a career. But doing no internal development and relying only on outside hiring is also problematic since it leaves the employer vulnerable to the whims of the labor market. "What we need is a way to deal with the uncertainty of business needs and the uncertainty of internal talent pipelines," says Cappelli, noting that the choice is not between developing talent internally or hiring from the outside. The better option is to do some of both: Use more adaptable models of internal development that include getting employees to share the costs, and then use outside hires to fill in the shortfalls when forecasts inevitably prove wrong.

Companies that are moving in this new direction include startups, which have a clean slate as far as talent management practices go, and professional services firms, where getting the right talent mix is especially critical. "Consulting firms, auditing firms, law firms and so forth started to make the effort to calculate the costs of poor talent management back in 1999 when the labor market was so tight," Cappelli says. "Because of the need to constantly hire new people, they know that their ability to compete can be severely compromised by high turnover." Indian firms, he adds, may be the leaders in new ways to think about managing talent because the talent crunch in that country is so severe.

Cisco's 'Voluntary Sabbatical'

In his book, Cappelli cites the talent management processes at a number of companies, including Unilever, IBM, General Electric, EDS, Dow, Capital One, Citibank, Corning, Johnson & Johnson and Bear Stearns, to name a few.

He describes the sophisticated forecasting model at Dow, which incorporates traditional statistical-based forecasting with such factors as the political and business climate in each of Dow's countries of operation, changes in labor and employment legislation, and business plans for the operating units. "Standardized systems make it possible to aggregate the individual estimates up to an overall projection for the company," Cappelli says.

At Capital One, where the challenge was to help the company plan its workforce -- which had gone from 20,000 employees in 2001 to 14,000 in 2005 to 30,000 in 2007 after a series of acquisitions -- the company assembled a team with experts in marketing and operations research, but none from the traditional HR function. The group used data mining techniques, manufacturing models and information from its PeopleSoft system to generate talent planning models for each business unit. "Rather than just predict the number of people required in each role, they also modeled outcomes such as attrition rates, employee morale, rates of promotion and outside hires." The big innovation at both companies is that these models have moved past traditional forecasting and toward simulations in order to deal with uncertainty in business. "Rather than generating a static estimate of how many workers will be needed two years out," says Cappelli, "they say to operating managers: 'Tell us the assumptions you have about your business, and we'll give you a talent estimate. Better yet, give us a range of different assumptions, and we'll give you a range of talent estimates within which the reality will most likely lie.'"

Cappelli recounts efforts by some companies to give employees more control over the career development process and thus make them more likely to stay with the company. Duke Power, for example, allows employees, under certain conditions, to post and swap jobs with other employees at their same job and salary level. Coca-cola run job fairs for its junior auditors; Gap operates an internal headhunting office where employees with at least two years' experience can look for other positions within the company.

Chubb "opened up its internal labor market by eliminating both job tenure and supervisor approval as requirements for changing jobs within the company." McKinsey posts all the projects that use associates (the level of employee below partner) on a worldwide system along with information on the relevant industry, the client, the in-house team and the type of project work. It then encourages associates to rank their preferences.

During the IT downturn in 2001 and after, Cisco offered a "voluntary sabbatical" to its employees in which the company agreed to pay one-third of their salaries while they spent time working at nonprofit organizations." Deloitte tries to keep former employees of Deloitte & Touche plugged in to the company for as long as five years after they leave (often for family reasons), provided they don't take a new job. Its Personal Pursuits program covers certification and skills programs fees to help them stay current, and offers access to company career and work-life programs, among other things. The idea in both cases is to keep employees "on the hook" with the company so that they can be brought back to work quickly should demand pick up.  

Cappelli acknowledges that uncertainty about whether skills will be needed in the future and whether employees will stick around makes it difficult for employers to recoup investments in those employees. One of the best ways to deal with that problem is to get employees to share the costs of development. "Rather than trying to guess who is ready for advancement,' Cappelli notes, "many companies have moved toward self-nomination, where individuals volunteer or apply for development experiences. The employers usually require that the candidates keep doing their regular jobs and maintain good performance in them. So the developmental experiences, which are typically work-based, are essentially free to the company."  

Selling Old Ideas as New Concepts

For the most part, however, companies are not yet going in new directions when it comes to adopting more efficient talent management techniques. This is especially so at the bigger, older employers like General Electric, Procter & Gamble, IBM, PepsiCo and members of the oil industry. Many of these are known as "academy companies," referring to their reputations as places where employees go to learn management skills and then are hired away by other firms.

GE "is doing the same things it did in the 1950s," says Cappelli. "It laid out a model and that model is not being questioned. Some parts of this model work well and are quite consistent with what I describe -- especially the ability to make matches between people and opportunities -- but other components aren't as efficient, such as the goal of having deep benches of talent. IBM no longer guarantees people lifetime employment and they do some amount of outside hiring, but they still direct the careers of their managers from headquarters.

So many HR people were laid off during the 1980s that HR personnel don't know that the planning models many are embracing are decades old, says Cappelli. "HR people are all drinking the same Kool-Aid. They are selling these practices like they were new ideas. At the same time, internal accounting is so bad that they don't even know the costs of their inefficient talent management efforts. Companies don't realize that they need a change."

A new approach to talent management is needed for two key reasons, according to Cappelli: On the public policy side, companies are not developing the talent the U.S. needs to stay competitive. On the employer side, most of the companies aren't doing talent planning, or their planning is wrong, even as their ability to hire on a just-in-time basis is eroding. Employers can't easily find people out there to poach; it's an expensive and time-consuming process to even look.

When planning practices were first initiated, Cappelli says, markets were stable enough to make long-term planning possible. "IBM, for one, had 15-year business plans that were pretty accurate. Companies in the defense industry had 10-year plans. You didn't have to make year-end adjustments back then. But these days, demand can change within a year. Authority and accountability are pushed onto individuals and not systems, and career mobility across companies is high. Employers must adapt to that reality."

Knowledge@Wharton Categories

• Finance and Investment
• Leadership and Change
• Executive Education
• Marketing
• Insurance and Pensions
• Health Economics
• Strategic Management
• Real Estate
• Law and Public Policy
• Human Resources
• Business Ethics
• Innovation & Entrepreneurs
• Operations Management
• Managing Technology