Silicon Valley, like its peers in other high flying industries, believes obsessively that elite people are the key to success. Venture capitalists want to invest in “great teams,” thought leaders evangelize for not settling for anything less than “A players,” and entrepreneurs commonly cite recruiting “great people” as their most pressing challenge.
This was not as relevant fifty plus years ago. A great worker in one of Henry Ford’s Model T factories was constrained by his role. The pace of the assembly line dictated his productivity. He could reliably keep pace, but not much more.
In contrast, a talented software engineer can solve challenging problems that others may not grasp or write elegant code that does not require lengthy revisits when bugs are discovered. Hence the belief, as Mark Zuckerberg expressed it, that “someone who is exceptional in their role is not just a little better than someone who is pretty good. They are 100 times better.”
(The modest expression of this idea is “10 times better.”)
Developers are not, however, the only professionals to have graduated from the diminishing returns of manufacturing work. A great salesman can completely outperform his peers, a great CEO can turn around a struggling company, and a great teacher can measurably improve the life outcomes of his or her students in a single year of teaching.
According to professors Herman Aguinis and Ernest O’Boyle Jr, the presence of “star performers” who radically outperform their peers is becoming the norm in the 21st century workplace. They argue that employee performance is best described by a “power law distribution” in which some 20% of employees are responsible for 80% of output. In other words, it is not the “necessary many” – like the thousands of line workers assembling Model T cars – who drive a company’s performance, but the “select few” most talented teachers, scientists, or executives.
The two call for management theory and practice to pay more attention to star performers in a recent paper. Noting how the move away from a manufacturing economy to a service economy – as well as empowering technologies like the Internet – allowed people to perform far above the mean, they cite examples past and present. The publication record of academics as measured by citations shows a small number of giants outperforming the rest. The same is true of the awarding of patents and National Science Foundation grants. The power law is at work in the measurable performance of certain salesmen and women, politicians, athletes, entertainers… and in the number of kills by fighter pilots in World War II.
Despite America’s zeal for celebrating individual achievement, management practices have been slow to recognize the impact of star performers. Researchers in management studies assume a normal distribution of worker productivity – a bell curve distribution in which the majority of workers perform around the mean and a marginal number of workers perform much better or much worse. Today, they go so far as to remove star performers from their data sets as “anomalies” in order to see the expected normal distribution. Their business students then take the lessons and theories based on this assumption to the workplace.
Why have managers and academics clung to the normal distribution? We asked co-author Ernest O’Boyle Jr. of the University of Iowa’s Tippie College of Business, and his answers are a cautionary tale of how an entire field can wear self-imposed blinders.
He notes that “our field relies strongly on theory and most of our theory was developed between 1950 and 1980 in manufacturing settings.” But when management theorists started approaching service organizations where productivity was more subjective and hard to measure, they looked at supervisor ratings.
Not wanting managers to rate all their employees a 9 or 10 out of ten, however, they trained managers to rate them “on the distribution of performance we thought was correct.” That distribution, of course, was a normal curve. By forcing manufacturing-based theory onto service industries, supervisors hid all the evidence of star performers’ outsize contributions. Back at the university, doctoral students trained in statistical techniques that assumed normal distributions molded performance data to fit their model’s assumptions.
The results have made academics tie themselves into knots. Researchers believe the competitive advantage of great companies lies in their human capital. Yet a normal distribution suggests that most productivity comes from a mass of average workers. “The paradox,” O’Boyle tells us, “is how competitive advantage can be derived from the group of workers that are the most plentiful and easiest to replace.”
Academics have laid out numerous “sum is greater than the parts” explanations, but the presence of star performers suggest that companies thrive thanks to their above average and difficult to replicate contributions. O’Boyle and Aguinis, like any good researchers, intend for their work to be a call for more research to understand star performers. But they also believe that managers need to adjust their policies in response to their presence.
Managers shouldn’t stick to universal HR policies but focus on recruiting, retaining, and empowering star performers. The professors encourage performance-based pay and discourage seniority policies and longevity-based promotions. They advocate for supervisors to spend less time getting underperformers up to speed and more time facilitating star performances. (A top real estate agent, for example, could get an assistant to do his or her paperwork.) Perks like increased flexibility and helping a spouse get a job (as universities often do to recruit star faculty) can woo talent. And for the love of God, don’t be the Hewlett Packard managers in the calculator division who 5 times ignored Steve Wozniak’s design for a personal computer.
In the late nineties, the prestigious consulting firm McKinsey & Company launched “The War for Talent,” an exhaustive look at the recruiting and HR policies of successful companies. They discovered star performers and the policies that O’Boyle and Aguinis advocate:
The very best companies, they concluded, had leaders who were obsessed with the talent issue. They recruited ceaselessly, finding and hiring as many top performers as possible. They singled out and segregated their stars, rewarding them disproportionately, and pushing them into ever more senior positions.
The above description comes from New Yorker journalist Malcolm Gladwell in a 2002 article. It is devoted to the company that most took McKinsey’s findings to heart: Enron.
Gladwell describes the Enron culture, where McKinsey often consulted and former McKinsey employees filled key roles up to the CEO position, as a “star system.” Executives relentlessly pursued graduates of top MBA programs who blew them out of the water in interviews. Seniority meant nothing and talented young individuals quickly assumed new responsibilities and rose in the company. Employees had incredible flexibility to work on new projects that interested them. Bonuses went overwhelmingly to the top performers.
After six consecutive years of being named “America’s most innovative company,” Enron collapsed in a massive accounting fraud scandal. Gladwell concedes that the causes of their failure were “complex,” but argues that their “star system” was a disaster waiting to happen.
Enron gave talented individuals new responsibilities and roles so fast that their performance reviews measured impressiveness and bluster instead of actual performance. He gives the example of Lou Pai, who started Enron’s power trading business:
Pai’s group began with a disaster: it lost tens of millions of dollars trying to sell electricity to residential consumers in newly deregulated markets. The problem, Hamel explains, is that the markets weren’t truly deregulated: “The states that were opening their markets to competition were still setting rules designed to give their traditional utilities big advantages.” It doesn’t seem to have occurred to anyone that Pai ought to have looked into those rules more carefully before risking millions of dollars. He was promptly given the chance to build the commercial electricity-outsourcing business, where he ran up several more years of heavy losses before cashing out of Enron last year with two hundred and seventy million dollars. Because Pai had “talent,” he was given new opportunities, and when he failed at those new opportunities he was given still more opportunities… because he had “talent.”
The flexibility afforded to stars proved to be a constant disruption as every new, exciting project stole away talented individuals, leaving their managers with holes to fill in their department. The desire to delight and empower stars ran similarly amok. One young gas trader who wanted to start an online-trading business quickly found herself with 250 Enron employees, office and server space, and legal reviews for a project she was eminently unqualified to run.
Emboldened by the mantra of A players, Enron created an undisciplined, narcissistic company that believed it was too talented to fail.
Enron serves as a lesson in how not to interpret and apply the importance of the star performer effect.
The employees at McKinsey and the companies to whom they preached their War for Talent gospel came primarily from elite universities and worked at elite firms. This made them susceptible to the twin mistakes of thinking that star performers must be people like them (ambitious, smart people from top schools) and that there is a direct line from elite talent to elite company performance.
Enron assumed that “talented” individuals with impressive educations, confidence, and intelligence that shined in interviews would be star performers. But a star performer is someone who drives firm performance by getting more out of the resources of the organization and making the entire team better. That may or may not be someone who shines individually, and empowering individuals at the sake of the team as Enron did will be a disaster in contexts where work is not a purely independent process.
Enron also failed to realize that stars are a product of their environment, with their impact in many contexts a function of their institutional resources as much as their own intelligence or personality. Star programmers “likely possess significantly higher levels of abstract reasoning than non-stars,” Dr. O’Boyle told us. But he also cites a study of Wall Street analysts. When the star analysts “transferred from one organization to another, their performance dropped dramatically and permanently when they did not bring their team with them.” Both personal traits and the firm’s resources drive star performance, and researchers are still investigating the role played by each.
Nor is the star performer effect driven by a select few “A players” that graduate from Harvard and Stanford every year or pull themselves out of obscurity by their bootstraps; it is part of the landscape in many service sectors. The star performer model holds from McKinsey to Dunder Mifflin – O’Boyle hypothesizes “that for industries where power laws have been shown to emerge, the organizations within those industries will conform to power law distributions regardless of being a great, average, or poor organization.”
He hastens to add that companies still need to vigilantly attract top talent, otherwise…
Departing “great” superstars are replaced by new “good” superstars. The good, but not great superstars still create a power law performance distribution, but overall production is diminished. A metaphor for the effect of the departure of superstars is that of an eroding sand dune. If one removes the sand at the base of the dune, sand falls forward, but not enough to entirely replace what was removed. The falling sand lowers the height of the dune, but the shape remains the same. In organizations that cannot retain their superstars, great people will be replaced with good people and good people with mediocre. The organization’s overall production lowers due to the erosion of top people, but the shape of the performance distribution is left unchanged.
In other words, top talent is definitely important, and more important than much of management theory has long recognized. But as these industries always follow the power law, the mere presence of star performers doesn’t mean that there are a select few individuals out there – the A players – who are inherently 100 times better than anyone else.
Sports are probably a useful analogy here. Clearly select athletes are responsible for the bulk of their teams’ production and publicity. Fans idolize them, and it is impossible to imagine many other athletes equaling the accomplishments of a Michael Jordan or Tom Brady. But no one is under any illusion that assembling a team of stars is enough to win a championship. We recognize that certain stars thrive under certain coaches or styles of play (but not others); that teamwork and a good culture can trump star power; and that the solid play of non-star teammates is crucial as well.
In many industries where the contributions of star performers has not been adequately recognized and awarded, firms need to do more to recruit, empower, and hold onto their star performers. In places like Silicon Valley, however, people need to restrain themselves from blindly joining the cult of A players.