Case Study 1: Business Success and Military History -- The Graveyards of Strategy

"We tend to think of history as an account of what happened. But history is also an account of what didn't happen -- and we almost never tell that story." -- Adapted from a principle in counterfactual historical analysis

Two Graveyards, One Structure

This case study examines survivorship bias operating in two domains that shape how billions of people think about strategy, leadership, and success: business advice literature and military history. The two domains appear unrelated. One fills airport bookshops; the other fills university libraries. But they share the same structural anatomy: the selection process that determines which cases become visible (success, victory) simultaneously destroys the evidence that would reveal whether the visible cases are actually informative (the failures, the defeated, the abandoned). In both domains, the result is a body of knowledge that systematically overstates the reliability of the strategies employed by the survivors and understates the role of circumstance, timing, and luck.

Part I: The Business Book Problem

How Success Gets Studied

The business success literature is one of the largest and most lucrative genres in publishing. Each year, dozens of books promise to reveal the secrets of successful companies, successful leaders, or successful entrepreneurs. The genre's structure is remarkably consistent: identify successful cases, analyze what they did, extract the common traits, and present those traits as prescriptive advice. Do what the winners did, and you will win too.

The most influential examples are familiar. Tom Peters and Robert Waterman's In Search of Excellence (1982) studied forty-three companies judged to be excellent performers and identified eight attributes they shared. Jim Collins's Good to Great (2001) studied eleven companies that transitioned from good to great performance and identified six factors that distinguished them. Collins's later work, Built to Last (1994, coauthored with Jerry Porras), studied eighteen companies that had endured for decades and identified the principles that explained their longevity.

Each of these books was a bestseller. Each was treated, at the time of publication and for years afterward, as something approaching scientific evidence about the causes of business success. And each was built on a methodology that made survivorship bias not merely possible but inevitable.

The Methodological Problem

The problem is not that these authors were sloppy. The problem is structural. To study what makes companies succeed, you must first identify successful companies. But the only companies available for study are the ones that have already succeeded -- the survivors. The companies that attempted the same strategies and failed are, by definition, not in the sample. They are bankrupt, dissolved, acquired, or forgotten. Their records are scattered or destroyed. Their founders are not giving interviews. Their case studies are not being taught in business schools.

This means that any trait identified as common among the survivors might be equally common among the non-survivors. If all the "great" companies had bold, visionary leaders, that tells you nothing unless you also know how many bold, visionary leaders ran companies that failed. If all the "excellent" companies were close to their customers, that tells you nothing unless you also know how many customer-focused companies went bankrupt. The trait that appears to distinguish winners from the general population may not distinguish winners from losers at all -- because the losers are invisible.

Consider an analogy. Suppose you studied everyone who won the lottery and discovered that they all bought lottery tickets. You might conclude that buying lottery tickets causes wealth. The conclusion is technically supported by the data: one hundred percent of winners bought tickets. But the conclusion is absurd, because it ignores the millions of people who also bought tickets and did not win. The ticket is necessary but not sufficient. The "success trait" is visible; the failure rate is hidden.

The business success literature commits a more subtle version of this error. The traits it identifies -- visionary leadership, disciplined culture, customer focus, willingness to pivot -- are probably present in many failed companies too. But we do not know, because failed companies are not studied with the same rigor, frequency, or cultural attention.

The Predictive Failure

The most compelling evidence that business success books suffer from survivorship bias is their predictive track record. If the traits identified by Peters, Collins, and others were genuinely causal -- if they reliably produced success -- then the companies profiled should have continued to succeed after the books were published. They did not.

Of the forty-three "excellent" companies profiled in In Search of Excellence in 1982, roughly a third were in serious financial difficulty within two years. By the early 2000s, several had ceased to exist as independent entities. The "excellent" traits did not predict continued excellence.

Of the eleven "good to great" companies profiled by Collins in 2001, the post-publication record is similarly discouraging. Fannie Mae required a government bailout. Circuit City went bankrupt. Wells Fargo became the center of a massive fraud scandal. Nucor and Abbott Laboratories performed respectably, but the overall portfolio performed no better than the broad market -- which is what you would expect if the selection criteria identified past luck rather than future-predicting traits.

The pattern has repeated with sufficient regularity to constitute a stylized fact about the genre: the companies profiled in business success books do not, on average, outperform the market after publication. This is consistent with the hypothesis that the books identified companies that had already been lucky and then attributed their luck to skill. The survivorship bias in the sample ensured that the "secrets of success" were not actually secrets of success. They were common traits of survivors, which is a different thing entirely.

What the Graveyard Would Tell You

If you could study the graveyard -- the thousands of companies that pursued the same strategies as the "great" companies and failed -- what would you learn?

You would almost certainly learn that the traits identified by success literature are necessary but not sufficient. Disciplined culture, customer focus, and visionary leadership probably do improve a company's odds somewhat. But they do not guarantee success, and many companies that embodied these traits failed anyway -- because of bad timing, unfavorable market conditions, regulatory changes, technological disruption, a single catastrophic decision, or simple bad luck.

You would also learn that the role of luck is far larger than success literature acknowledges. Successful companies were often in the right market at the right time with the right product -- circumstances that they may have positioned themselves to exploit but did not create. The failed companies in the graveyard may have had equally good strategies deployed in slightly less favorable conditions.

Most importantly, you would learn that the variance in outcomes -- the spread between the best and worst results for companies with similar traits -- is enormous. Success literature, by studying only the right tail of the distribution, creates the impression that the strategies it recommends reliably produce success. The graveyard would reveal that the same strategies also produced a great deal of failure. The strategies are not reliable. They are probabilistic, with wide confidence intervals. But you cannot see the confidence intervals if you can only see the survivors.

Part II: Military History and the Vanishing Losers

The Winner's Archive

Military history, as a discipline, has a survivor problem that is even more severe than business literature's, because the selection process in military conflict is far more destructive. When a company goes bankrupt, its records may be scattered but they are not systematically destroyed. When a civilization loses a war -- particularly in the pre-modern era -- its records, its archives, its libraries, its oral traditions, and its intellectual culture may be physically annihilated.

The result is a historical record that is systematically skewed toward the victors' perspective. We know Roman military strategy because Rome won, survived for centuries, and produced a literate elite that documented its methods. We know Greek military strategy because Greek city-states produced writers -- Thucydides, Xenophon, Polybius -- who analyzed warfare with remarkable sophistication. We know the strategies of the European colonial powers because those powers dominated the production of written historical records for centuries.

What we do not know, in most cases, is what the losers knew. The military doctrines of the peoples Rome conquered and absorbed -- their tactical innovations, their strategic debates, their logistical systems -- are almost entirely lost. We know Hannibal's strategy at Cannae because Roman historians documented it (it was, after all, one of the most devastating defeats Rome ever suffered, and Romans had reason to understand how it happened). But Carthaginian military thought more broadly -- the accumulated doctrine and strategic culture of one of the ancient world's most successful commercial and naval powers -- is largely gone, destroyed when Rome razed Carthage in 146 BCE.

The Strategy Attribution Error

The survivorship bias in military history produces what we might call the strategy attribution error: the tendency to conclude that the winning side's strategy was superior because it won. This reasoning is circular. The strategy won; therefore it was good. It was good; therefore it won. The reasoning excludes the possibility that the strategy succeeded because of circumstances unrelated to its inherent quality -- terrain, weather, logistics, disease, morale, intelligence failures on the other side, or simple luck.

Consider the Allied victory in World War II. The standard narrative attributes the victory to the Allies' strategic superiority: the coordination of the D-Day invasion, the encirclement strategy on the Eastern Front, the island-hopping campaign in the Pacific, the economic mobilization of the United States. These were genuine achievements. But the narrative inevitably emphasizes the strategies that worked and de-emphasizes the strategies that failed -- the disastrous Allied campaigns in North Africa in 1942, the catastrophic Battle of the Hurtgen Forest, the intelligence failures that allowed the Battle of the Bulge, the costly and arguably unnecessary bombing campaign against German cities.

The Axis powers, by contrast, are studied primarily through the lens of their failure. German military strategy in World War II was, by many measures, tactically innovative and operationally brilliant. The blitzkrieg campaigns of 1939-1941 represented a genuine revolution in combined-arms warfare. But because Germany lost, its strategic innovations are studied primarily as a cautionary tale about overreach, hubris, and moral corruption. The same innovations, had they been employed by the winning side, would be studied as exemplars of strategic genius.

The bias is not merely a matter of perspective. It affects the actual body of strategic knowledge that is transmitted across generations. Military academies teach the campaigns that won. They study the tactics that succeeded. The strategies of the losers -- which may have been equally sound but deployed under less favorable conditions -- receive far less attention. The result is a strategic canon that overrepresents success and underrepresents the full range of strategic possibility.

The Total Destruction Problem

The most severe form of survivorship bias in military history occurs when the losing civilization is destroyed so thoroughly that its entire knowledge base is eliminated. The Mongol conquests of the thirteenth century destroyed cities, libraries, irrigation systems, and administrative infrastructures across Central Asia, Persia, and the Middle East. The sack of Baghdad in 1258 -- in which the Mongols destroyed the House of Wisdom, one of the greatest libraries in the medieval world -- eliminated not just military records but the accumulated scientific, philosophical, and literary knowledge of the Abbasid civilization.

The Spanish conquest of the Americas destroyed the record-keeping systems of the Aztec and Inca empires. The Maya codices -- the written records of one of the most sophisticated civilizations in the pre-Columbian Americas -- were systematically burned by Spanish missionaries, with only four surviving. The military, administrative, astronomical, and historical knowledge contained in those codices is permanently lost.

In each case, the destruction was not incidental to the military victory. It was a consequence of it. The selection process -- military conquest -- destroyed the evidence that would have provided the losers' perspective on the conflict, their own assessment of their strategies, and their understanding of what went wrong. The evidence was not merely hidden or overlooked (as in the streetlight effect). It was annihilated by the same process that determined who won and who lost.

What the Graveyard Would Tell You

If the military graveyard could speak, what would it say?

It would almost certainly say that winning strategies are far less reliable than military history suggests. Many strategies that won in specific circumstances would have failed under slightly different conditions. Many strategies that failed could have succeeded with minor adjustments or different opposition. The variance -- the range of possible outcomes for any given strategy -- is far wider than the clean narratives of military history imply.

It would say that luck, circumstance, and contingency play a much larger role in military outcomes than the victory narratives acknowledge. Battles turn on weather, disease, timing, morale, a single commander's decision at a single moment. The strategy is one variable among many, and often not the decisive one.

And it would say that the knowledge of the defeated -- their innovations, their insights, their hard-won understanding of what works and what does not -- is worth preserving, not because it led to victory but because it represents a different perspective on the same strategic problems. The losing general may understand the battlefield better than the winning general, precisely because the losing general was forced to grapple with circumstances that the winning general's luck allowed him to avoid.

The Structural Parallel

Business success literature and military history suffer from the same structural bias, operating through the same mechanism:

In both domains, the fundamental error is the same: studying only the survivors and concluding that the survivors' traits caused their survival. The correction is also the same: seek out the dead.

Discussion Questions

1. If the traits identified by business success literature (visionary leadership, customer focus, etc.) are present in both successful and failed companies, what does that tell us about the actual causes of business success? What additional evidence would you need to identify genuine causal factors?

2. The chapter argues that military history is "written by the winners." Can you identify an exception -- a case where the losing side's perspective has been preserved and has changed our understanding of a conflict? What made preservation possible in that case?

3. Jim Collins's methodology in Good to Great included comparison companies -- firms in the same industries that did not make the good-to-great transition. Does this comparison adequately address survivorship bias? Why or why not?

4. The Mongol destruction of Baghdad's House of Wisdom in 1258 destroyed centuries of accumulated knowledge. Can you identify a modern equivalent -- a case where the destruction of knowledge has distorted our understanding of a field, a technology, or a conflict?

5. If you were designing a business school curriculum that was resistant to survivorship bias, what would you change about how business strategy is taught? What cases would you add? What questions would you require students to ask?

6. Connect the strategy attribution error in military history to the narrative capture discussed in Chapter 36. How do victory narratives amplify survivorship bias? How does the structure of a good story reinforce the tendency to attribute outcomes to strategy rather than luck?