The photo above was taken by Robert Capa, the only civilian photographer to follow the Allied troops landing at Omaha Beach on June 6, 1944. Capa took 106 photos that day. But then something very unfortunate happened. Most of the pictures were destroyed in a processing accident. Only eleven survived. They are known as “The Magnificent Eleven.”
These photos — mostly blurred, surreal shots — convey the intensity and chaos experienced by soldiers on the front line as the invasion unfolded. In hindsight, it turns out that the invasion was a great success, but it certainly didn’t feel that way for the soldier when he first landed. That soldier stepped right into a fog of war, and into a dangerous and uncertain fate.
The generals that orchestrated this invasion also faced a high degree of uncertainty. The question that is on the mind of any strategist trying to succeed in this world, is what leads to higher performance? In the world of business, this question has spawned a plethora of books selling us formulas for success, filled with insider “secrets” along with invitations to imitate companies with stellar track records. A very notable exception to this is Phil Rosenzweig’s book, The Halo Effect, effectively an antidote to the prevailing nonsense in management literature, as the author uncovers the problems underlying explanations of success in business.
The Halo effect
During World War I, the American psychologist Edward Thorndike carried out research on how army officers rate their subordinates. “In one study, he asked army officers to rate their soldiers on a variety of features: intelligence, physique, leadership, character, and so on. He was struck by the results. Some men were thought to be “superior soldiers” and rated highly at just about everything, while others were thought subpar across the board. It was as if officers figured that a soldier who was handsome and had good posture should also be able to shoot straight, polish his shoes well, and play harmonica, too.”
Thorndike called this the Halo effect.
A well-documented setting for the Halo effect is the job interview. When a recruiting manager picks up a CV for the first time, some of the first tangible information he will see is the university the candidate went to, his grade point average, and other distinctions. “With this information clearly in mind – relevant, tangible, and seemingly objective – interviewers tend to shade their evaluations about other things that are less tangible, such as a candidate’s personal manner or the quality of answers to general questions.” The business world is marred by this kind of reliance on labels and credentials as indicators of competence.
For Rosenzweig, however, nothing lends itself more to the Halo effect than leadership: “Good leaders are often said to have a handful of important qualities: clear vision, effective communication skills, self-confidence, personal charm. Most people would agree these are elements of good leadership. But defining them is a different matter altogether, since several of these qualities tend to be in the eye of the beholder – which is affected by company performance.”
A serious scholar of leadership, James Meindl, “concluded after a series of insightful studies that we have no satisfactory theory of effective leadership independent of performance. We think we know what good leadership is all about – clarity of vision, communication skills, good judgment – but in fact a wide range of behaviors can be said to fit these criteria.” It might be that we can describe leadership in terms of general virtues, but it’s very hard to pinpoint exactly how a good leader should act in a specific situation.
Rosenzweig argues that “for all the books written about leadership, most people don’t recognize leadership when they see it unless they also have clues about company performance from other things that can be assessed more clearly – namely, financial performance.”
Another study, by Marianne Bertrand at the University of Chicago and Antoinette Schoar at MIT, investigated “whether company performance is affected by the chief executive’s personal managerial style. It’s easy to find anecdotal evidence that CEOs of successful companies have effective personal styles. Instead, the authors defined ‘managerial style’ in terms of two specific policies: investment policies (capital expenditure, frequency of M&As) and financial policies (level of debt and dividends). Both were objective and measurable. They controlled for a number of other variables. They found that individual managers do have preferred personal styles when it comes to investment and financial policies, but these preferences explain 4% of the variance of company performance.” Other research similarly indicates that managerial policies explain no more than 10% of variance in performance. This is useful to know, but hardly a panacea.
The delusion of causality
If all we have is a correlation, there is no way to really know if the phenomena that we see are in fact explained by an underlying causal mechanism. Psychologist Edwin Locke made the point emphatically: “While the method of correlation may be useful for the purposes of suggesting causal hypotheses, it is not a method of scientific proof. A correlation, by itself, explains nothing.” One way to eliminate some of the uncertainty regarding correlations is to analyze them over an extended period of time — to see if a regular, consistent and predictable pattern emerges.
Rosenzweig cites a study performed by Benjamin Schneider and colleagues at the University of Maryland, which “used a longitudinal design to examine the question of employee satisfaction and company performance, to try to find out which causes which. They gathered data over several years so they could watch both changes in satisfaction and changes in company performance.” Their conclusion was that “financial performance, measured in return on assets and earnings per share, has a more powerful effect on employee satisfaction than the reverse. It seems that being on a winning team is a stronger cause of employee satisfaction.” The fact that employees are satisfied just doesn’t have any effect on the company’s financial performance; but the reverse seems to be true.
How did Wal-Mart become such a success? How did it grow so big? Rosenzweig observes that “there’s no shortage of theories: Perhaps it was a strategy of everyday low prices, or a relentless obsession with detail, or a culture that gets ordinary people to do their best, or a sophisticated use of information technology in supply chain management, or maybe a bare-knuckled approach to squeezing its suppliers. Are some explanations right? Are all of them right? Which are most important? Do some only work in combination with others? The fact is, it’s hard to be sure.”
Many people look at a single explanation for company performance and leave everything else aside. That wouldn’t be an issue if there were no correlation among different factors, but in reality there will likely be a multitude of interrelated factors that are likely to be found in the same company. This points out the importance of having a multivariate analysis – making sure that one analyzes the various possible effects of as many potentially relevant factors as possible, before determining the effect that a single one of them might have.
Some may claim that “since business will never be understood with the precision of the natural sciences, it’s best understood as a sort of humanity, a realm where the logic of scientific enquiry doesn’t apply. Well, yes and no. It may be true that business cannot be studied with the rigor of chemistry or geology, but that doesn’t mean that all we have is intuition and gut feel.”
We can’t buy 100 companies and run a laboratory experiment under perfectly controlled conditions. The only possible approach available to social science is one of quasi-experimentation, where we analyze events that have already taken place, and then try to find out what affects a particular variable, say a company’s bottom line or its number of acquisitions.
Rosenzweig argues that much of social science, “precisely because it’s done carefully and is circumspect in its findings, tends not to provide clear and definitive guidelines for action. It’s just not very appealing to read that a given action has a measurable but small impact on company success. Managers don’t usually care to wade through discussions about data validity and methodology and statistical models and probabilities. We prefer explanations that are definitive and offer clear implications for action. We like stories.”
Another problem that arises when authors interview leaders to ask them about the source of their success, is that the explanations are very likely going to be filled with rationalizations, as well as hindsight and survival bias. “Retrospective self-reporting is commonly biased by performance (…) Ask managers why their companies are successful, and we’re likely going to get the kind of attributions of the sort we’ve seen over and over.” Yet most popular books from business literature rely on exactly this method: gather as much information as you can from the most successful companies out there, find out what they have in common, and there you have it – the secret formula for success.
Many management books make a typical “sample selection directly based on the dependent variable – that is, based on outcomes. It’s a classic error.” Suppose, for instance, that we want to find out what causes an outcome such as high blood pressure. We’ll never find out if we only examine patients with high blood pressure. We have to compare them to people who don’t have the condition. It’s the only way to find out what really differentiates people with high blood pressure. “The same applies to companies: by only looking at companies that perform well, we can never hope to show what makes them different from companies that perform less well.” A few examples of pseudo-scientific bestsellers that conveniently forget to do this include In Search of Excellence, Built to Last, and From Good to Great. Don’t buy these books: they are full of shit.
Why the appeal? Because they work wonderfully as stories. Many of the authors behind these best-selling works masquerade under the pretense of rigorous research. Collins and Porras, the authors of Good to Great, describe how they devised a “systematic and comprehensive” framework for data collection. “They read more than 100 books, including company histories and autobiographies. They consulted more than 3,000 documents, ranging from articles to company publications to video footage. They read Harvard and Stanford case studies. They performed “extensive literature searches” from sources including Forbes, Fortune, Business Week, The Wall Street Journal, Nation’s Business, The New York Times, and so on. The impression was clear: we were very, very thorough.”
Yet the authors fail to address a basic problem of social science: how to draw a valid inference from a limited statistical sample. As Rosenzweig points out, “the quantity of data is entirely beside the point if the data aren’t of good quality. If your data sources are corrupted by the Halo effect, it doesn’t matter how much you’ve gathered.”
Is there any value to be gained from a good story? Perhaps. Rosenzweig believes that “the test of a good story is not whether it is entirely, fully, scientifically accurate. Rather, the test of a good story is whether it leads us toward valuable insights, and if it inspires us toward useful action, at least most of the time.”
Most management books return again and again to the same fundamentals, proclaiming that “companies do well when managers live by deeply held values, pursue a clear vision, care about their employees, focus on their customers, and strive for excellence.” One can probably find a lot of good in these basic principles. But the problem is that there are also many ways in which stories that promise success can be harmful. In her book The Confidence Game, Maria Konnikova uncovers how con artists lure their victims with convincing stories of guaranteed success. Somehow, it seems, we are programmed to believe in good stories.
The delusion of lasting success
“Lasting business success,” Rosenzweig argues, “is largely a delusion.” It’s very hard to see this when one picks a limited sample of successful companies. If these companies have performed extremely well over the last sixty years, should we not infer that they are on the right track, that they just got it right? To see how this type of reasoning can be misleading, we must shift our perspective and examine the performance of a large number of companies over a longer period of time:
How many successful companies on the S&P 500 in 1957 were still on the S&P 500 in 1997, forty years later? Only 74. The other 426 were gone – nudged aside by other companies, or acquired, or bankrupt. And of the 74 survivors, guess how many outperformed the S&P 500 over that time period? Only 12 out of 74. The other 62 survived, yes, but they didn’t thrive.
This indicates that markets are much more dynamic that one would think, and that yesterday’s performance is often a very poor indicator of tomorrow’s performance. Successful companies come and go. They do not remain at the top for long. As Rosenzweig observes, “companies that last longest usually aren’t the best performers. Enduring greatness is neither very likely, nor, when we find it, does it tend to be associated with high performance.”
When looking for determinants of success, one cannot simply pick companies based on their track record. That would be “an exercise in ex post facto selection,” and it is certainly not a good way to understand the business world. “What’s missing is the flux and the dynamism of performance, the ebb and flow,” or what Schumpeter called creative destruction.
In fact, “if we start with the full data set and look objectively at many years of company performance, we find the dominant pattern is not one of enduring performance at all, but one of rise and fall, of growth and decline.” One such long-term study by McKinsey concludes that “the corporate equivalent of El Dorado, the golden company that continually performs better than the market, has never existed. It is a myth. Managing for survival, even among the best and most revered corporations, does not guarantee strong long-term performance for shareholders. In fact, just the opposite is true. In the long run, the markets always win.”
“Looking for those few golden companies that succeed decade after decade may be a delusion, but it’s one that managers are eager to grasp. After all, showing how companies tend to rise and fall over time doesn’t make for a very compelling story. We prefer to read about Excellent and Visionary companies; we want to know the secrets of their success so we can do likewise. It’s a far more appealing story than the one suggested by the facts: that success is largely transitory and that most companies that have done well in the past won’t outperform the average in the future.”
Success is not necessarily random – but it is certainly fleeting. “It’s entirely normal and very predictable that companies fall back after outstanding performance.” This phenomenon is commonly referred to as regression to the mean.
There is a simple reason as to why high performance is so difficult to maintain: in a market system, high profits tend to decline due to the “erosive forces of imitation, competition, and expropriation.” Rivals copy the leader, patents expire, new companies enter the market, “consulting companies spread best practices, and employees move from company to company.” In such a dynamic environment, whatever initial advantage one has will eventually dissipate.
On risk and uncertainty
In his essay called The Hedgehog and the Fox, the British philosopher Isaiah Berlin entertains the idea that people can be categorized as Hedgehogs and Foxes. Berlin was inspired by the ancient Greek poet Archilochus, who wrote that “the fox knows many things, while the hedgehog knows one important thing.” (πόλλ’ οἶδ’ ἀλώπηξ, ἀλλ’ ἐχῖνος ἓν μέγα)
Berlin argues that Hedgehogs are heavily invested in a big central idea, and relate everything to a single defining vision or system. Foxes draw on all sorts of dispersed information, and rely on a multitude of strategies for different problems. They are comfortable with nuance, they can live with contradictions.
Rosenzweig borrows this analogy to differentiate risk-takers: those who put everything they have on a big bet (Hedgehogs) and those who diversify their investment into several, smaller bets (Foxes):
Imagine that a thousand people spend the day betting at the racetrack, and at the end of the day we select the ten bettors with the highest winnings – we’ll call them Great bettors. When we look closely at these most successful bettors, we’re likely to find that all of them placed big bets on long shots – that’s how they came out ahead of the other 990. They were Hedgehogs, focusing on a few big things. Very few Foxes will be among the top ten, because Foxes tend to diversify their positions. Yet even if the top ten bettors were all Hedgehogs, it does not follow that Hedgehogs, on average, outperformed Foxes, because some Hedgehogs may have done very well but many more have gone home broke. In fact, overall Foxes probably did better than Hedgehogs – they took more prudent risks and avoided big losses.
As Rosenzweig astutely remarks, it is entirely possible “that a Hedgehog-like focus is a risky but necessary gamble when striving for Greatness.”
The economist Robert H. Frank makes a similar observation when he asserts that “we sense, correctly, that performance depends far more strongly on ability and effort than on small random occurrences.” However, our intuition often fails “because even things that are highly improbable in any specific instance become likely if there are enough opportunities for them to occur” in a larger population and over a longer period of time. Some people just get wildly lucky.
When it comes to companies operating in a market, things really aren’t that different. “A strategy always involves risk because we don’t know for sure how our choices will turn out.” Here also, some lucky Hedgehogs get disproportionate wins. Rosenzweig cites four reasons as to why strategic decisions are so uncertain:
The first has to do with customers and market demand. “Many major initiatives, like the launch of a new product or a new business model, don’t easily lend themselves to experiments. There are, in fact, legendary examples where mountains of market research didn’t help at all.” One such example was the planned expansion of American retail store Target, which decided to make an aggressive move into the Canadian market by simultaneously opening more than a hundred new stores in that country. After widespread consumer disappointment and massive losses, Target eventually closed all of its Canadian stores.
A second source of risk has to do with competitors. As Robert H. Frank points out, “chance events are more likely to be decisive in any competition as the number of contestants increases. That’s because winning a competition with a large number of contestants requires that almost everything go right (…) The most skilled contestant is no more likely to be lucky than someone else. With a large number of contestants, there are bound to be many with close to the maximum skill level, and among those some will also happen to be very lucky.”
In a similar vein, Rosenzweig points out that competitors may “have made equally good predictions about customers and may pursue a similar set of choices – or may leapfrog entirely with a revolutionary product or service. Predicting a rival’s moves is hardly an exact science – especially when that rival is also trying to predict our behavior. Expand the game to include multiple players, each with somewhat different resources and capabilities and risk preferences, and the complexity of the game grows in exponential leaps.”
A third source of risk comes from technological change. “Some industries are stable, with products that don’t change much, and customer demand that stay steady for a long period of time. But in other industries, technology changes rapidly and strategic choices can come one after the another with life and death consequences.” People working in venture capital know how extremely hard it is to predict technological breakthroughs. Modern technologies have also given rise to winner-take-all markets, where a small initial advantage can make a small challenger grow rapidly into a giant with a dominant market share.
A final source of risk comes from internal capabilities. “The fact is, managers can’t tell exactly how their company – with its particular people and skills and experiences – will respond to a new course of action. The many and subtle interrelationships inside a company make it hard to know what will be the outcome of a given set of actions.” For Rosenzweig, “an organization isn’t a system of mechanical parts, interchangeable and replaceable. It’s better understood as a sociotechnical system, a combination of men and machines, of people and things, of hardware and software, but also of ideas and attitudes.”
All of this indicates that performance is a very relative thing — it depends on a myriad of factors both inside and outside of the company itself. Rosenzweig points out that this is “an uncomfortable truth, because it admits that some elements of business performance are outside of our control.”
“Despite our best efforts, the ways that people and processes work together in complex organizations are very hard to untangle and even harder to transplant elsewhere with the same results. None of this suggests that some practices aren’t, on balance, better than others, nor does it mean that they won’t be generally useful for most companies much of the time. It means only that execution, like strategy, doesn’t lend itself to predictable cause-and-effect relationships. Our best efforts to isolate and understand the inner workings of organizations will be moderately successful at best.”
“If we’re not careful, any successful company can be said to execute well, and any failure can be explained, after the fact, as a failure to execute,” due to the Halo Effect. “It’s always easier to bang the drum about execution than to address fundamental questions of strategy. It’s always easier to insist we’re going in the right direction but just need to run a little faster; it’s far more painful to admit that the direction may be flawed, because the remedies are much more consequential. Managers quite naturally find it easier to keep the attention on execution, which everyone will always agree can be done better.” But the fact remains that if you’re on the wrong train, every stop is the wrong stop.
In a complex and unpredictable world, Rosenzweig argues that “wise managers know that business is about finding ways to improve the odds of success.” As we will see, some risk-taking may also required.
Learning to operate in a fog
Once we accept that success is far more uncertain than we would think, what does that leave for decision-makers? What kind of guidance should they follow? According to Rosenzweig, “we have to take a look at the decision-making process itself, setting aside the eventual outcome.”
Regardless of what happened after decisions were made, managers should ask themselves, “had the right information been gathered, or had some important data been overlooked? Were the assumptions reasonable, or had they been flawed? Were calculations accurate, or had there been errors? Had the full set of eventualities been identified and their impact estimated?”
“This sort of rigorous analysis, with outcomes separated from inputs, isn’t natural to many people. It requires an extra mental step, judging actions by their merits rather than simply making ex post facto attributions, be they favorable or unfavorable. It may not be an easy task, but it’s essential.”
Rosenzweig brings up the example of Robert Rubin, former Goldman Sachs executive and secretary of the treasury under Bill Clinton. For Rubin, this sort of probabilistic thinking was natural. His view of the world was based on a deep understanding of uncertainty:
Some people I’ve encountered in life seem more certain about everything than I am about anything. That kind of certainty isn’t just a personality trait I lack. It’s an attitude that seems to me to misunderstand the very nature of reality – its complexity and ambiguity – and thereby to provide a rather poor basis for working through decisions in a way that is likely to lead to the best results.
In an environment filled with uncertainty, it is essential to decide “what not to do as much as what to do.” As Nassim Nicholas Taleb points out with his concept of via negativa, it is far easier to rule out and avoid bad strategies than to spend effort on finding the one winning formula. Time spent on identifying and avoiding very costly mistakes can be extremely valuable to ensure continued survival. And yet, precisely because a tragedy is avoided, nobody notices, and the manager that helped prevent a very costly failure gets no bonus at the end of the year.
Survival is one thing, outstanding success is another. According to Rosenzweig, “it’s not enough to do something well; you have to do it better than others – and that means you have to take chances. Andy Grove’s 1996 book, Only the Paranoid Survive, is a thoughtful primer for managers about strategic inflection points – moments of extreme risk when the company’s life is at stake.”
In fact, “a willingness to take risks is essential – and not for the faint of heart. It’s stimulated in part by fear. Grove made a point that fear plays a major role in creating and maintaining a passion for winning in the marketplace: fear of competition, fear of bankruptcy, fear of being wrong, and fear of losing all can be powerful motivators. Business bestsellers don’t normally talk about fear. Fear has no place in the rosy world of rags to riches.”
For Rosenzweig, once a company takes a calculated risk – when it commits to follow a new and uncertain road – “the best managers act as if chance is irrelevant: persistence and tenacity are everything.”