In January 2016, at the cyclocross world championships in Zolder, Belgium, officials carried a scanner through the pit area and passed it over more than a hundred bicycles. Cyclocross is the winter discipline of bicycle racing: an hour of laps on a short course of grass, mud and sand, with barriers the riders dismount for, on bikes that look like road bikes fitted with knobbed tires. One of the bicycles had a motor in it. The bike stood in the pit of Femke Van den Driessche, nineteen years old and the reigning Belgian and European champion in the women's under-23 category, and the motor sat inside its seat tube with a battery, wired to a switch under the handlebar tape. Whether it had ever been switched on did not matter. The presence of one in a bike at a race is the offense. In April the Union Cycliste Internationale banned her for six years and annulled every result she had posted since the previous October, both titles included.
Nobody needed a result to know what the motor did. It added power to the pedals, and more power is more speed. The motor was an advantage anyone could understand before the race, from the mechanism alone, with no finish to explain. Most explanations of success run the other way, offered after the result to account for it.
The motor was also unfair, and helping was not the reason. Everything a rider brings to the line helps. The competitors had agreed to a contest of bodies and skill, and a motor is neither. Suppose instead that the rules allowed it and every rider had the same one. The motors would still add power, and having one would distinguish nobody. It would be like the skinsuit every rider already wears: a tight suit is faster than a flapping jacket, and because every rider has one, it explains nothing about who wins. An advantage the rivals can match is not an advantage.
The rules forbid very little else. The rider with the biggest engine at Zolder was born with most of it, in heart and lungs and muscle, and chose their parents no more than their rivals chose to ride without a motor. An engine in the seat tube is cheating and an engine in the chest is a gift, and the difference between them is a decision the sport has made rather than a fact about engines. Sports that want closer contests do equalize. A handicap horse race loads the better horse with more weight, and a golf handicap gives the weaker player strokes, so that a duffer and a club champion can play for money. Cycling does neither, because the race exists to find out who is fastest given what each rider was born with and what they built. The people who run a contest choose where to draw the line between the advantages it allows and the ones it forbids.
In business the line is the law, and inside the line an unfair advantage is the object of the exercise. Investors ask founders what their unfair advantage is, and they ask it as encouragement. What a company wants is the equivalent of the motor, legitimately acquired: something that improves its results and that its rivals cannot pick up at the bike shop. It might be a production process, a reputation, a base of customers, or the scale to buy inputs for less than anyone else pays. A resource with both properties, one that improves performance and that rivals cannot easily get, I will call an alpha asset, and chapter 2 gives the definition its edges.
The questions the motor leaves open are the ones this book is about: how much of an advantage is given and how much built, what a result can tell us about whether anyone has one, and how a company or a person who wants one goes about getting it. The place to begin is the starting line at Zolder, with the motor removed.
The Starting Line
One rider can sustain more power than the others because of the body they were born with. Another has spent eight winters learning to carry speed through a muddy corner, a skill that is built. A third rides better equipment or has a team car and a mechanic in the pit. All of them train. They still arrive at the line with different prospects.
The body is a good place to see how an advantage can be real and still have mixed origins. A cyclist's sustainable power depends on maximal oxygen uptake, VO2 max, measured in milliliters of oxygen per kilogram of body weight per minute, together with the fraction of that maximum they can hold and the efficiency with which they turn oxygen into work at the pedals. An untrained young woman typically measures in the high thirties; an untrained young man in the mid forties. Elite endurance athletes measure in the seventies and eighties. Greg LeMond, the American who won the Tour de France three times, is usually credited with 92.5, and Bjørn Dæhlie, the Norwegian cross-country skier who won eight Olympic gold medals in the 1990s, with 96, though the published figures for famous athletes come from different labs on different protocols and are better treated as a range than as a ranking. Training raises the number, by an amount that differs greatly from one body to the next. In the HERITAGE Family Study, 481 sedentary adults from 98 families followed the same twenty-week program. The average gain in maximal oxygen uptake was about 400 milliliters per minute. Some participants gained almost nothing and others gained more than a liter, and the response ran in families, with roughly half of its variation attributable to inheritance. The rider's engine is partly a gift and partly a construction, in proportions that differ from rider to rider, and nobody can read the proportions off a result.
Then there is the day. A cyclocross course is a sequence of corners, mud, sand, and barriers, and a race is decided by a few seconds out of an hour. A puncture, a crash in front of them, a line through a corner that held on one lap and gave way on the next: these decide places. A rival's puncture is a favorable event that leaves the rider no faster than they were. Companies have weather of their own. A construction boom lifts every supplier of building materials, a customer's bankruptcy empties a supplier's order book, and a fire at a rival's plant fills it, with no change in how well anyone operates.
So the riders differ in three ways. Some differences are advantages: a capability or a position that would, in expectation, produce better results than the rivals in this contest. Some advantages were inherited and some were built, and most, like the engine, are both. And some differences in results are not advantages at all but the variation of the day, which chapter 3 will call noise. Three separate questions follow. Whether an advantage exists is a question about mechanism: does the thing add power? What explains a particular result is a question about a mixture that no single result reveals. And what credit the winner is due is a third question, which depends on the first two and on where the advantage came from. Most arguments about success run the three together. This book keeps them apart.
The same three questions apply to a company that earns more than its rivals for a decade, a fund manager who beats the market five years running, a founder whose company sells for a billion dollars, and a professor whose early papers are cited more than their classmates' papers. In each case something improved the odds, and we want to know what it was, where it came from, whether anyone else could get it, and whether it will still be there next year.
What a Result Can Tell
Michael Phelps won eight gold medals at the 2008 Olympics. The profiles written that summer gave his body as the reason: size 14 feet, a wingspan reported as 80 inches against a height of 76, and ankles said to flex farther than other swimmers' ankles. The wingspan came from a newspaper, the feet and the ankles from a blog, and the profiles repeated the figures from one another. When Scientific American put the ankle claim to an exercise physiologist, he doubted that Phelps was much more flexible than his rivals and suggested that the main difference was fitness. Phelps may have the ankles. Nobody had measured them before the explanation went to print.
Explanations of success drive decisions. A board decides whether to keep or fire a chief executive on three years of results. A limited partner commits a hundred million dollars to a venture fund on the record of its last two funds. A hiring manager chooses the candidate whose last project succeeded. A young engineer decides whether to copy the career of a founder they admire. Each of these decisions assumes that the record reveals the ability. When inheritance and the day account for most of the record, the decision rests on a record that a person with no such ability might have produced just as easily.
The problem is easier to see at a small scale. Two people sell equipment for the same company. One books two million dollars in a year, the other one million. The first may be the better salesperson. They may also have inherited a territory full of customers whose machines were due for replacement, while the second spent the year winning customers who had never bought from the company at all. Asked to explain their year, the first will describe how they qualify a prospect and how they follow up, and they will mean it. The replacement schedule of the equipment their predecessor sold will not be in the account. A company that responds by distributing their presentation slides has copied the easiest thing to copy. Whoever succeeded was a participant in an experiment run once. Knowing what they did is useful. Knowing that it caused the result is something more.
Writers on investing, on randomness, and on the stories business books tell about successful companies have made the case that a record is a poor guide to skill. The question here is different. Whether the salesperson earned their territory, or Van den Driessche her 2015 titles, is a question about origins, and it is often unanswerable. Whether either held a real advantage going into the next year is a question about the present, and it is often answerable: a reputation, a customer base, a body, a position can be observed directly, and its effects can be reasoned about, whatever produced it. A lucky early result can create a real later advantage. A board, an investor, or a person planning a career needs the second answer more than the first, and most of this book is about how to get it.
My Mother's Question
I had been working on this book, in its earlier form, for several years when my mother asked me what it was about. She is a historian. I explained the basic idea: that companies and people who sustain an advantage over rivals possess something that improves their results and that rivals cannot easily get, and that most of these advantages are built over time by mechanisms I could describe. She listened and then asked what evidence I had that any of this was true.
Ouch. My education is in science and technology. I am usually the one asking for the evidence. I did not have a good answer that afternoon. I have a better one now. The claims in this book come in four kinds. Some are definitions, true by construction, which prove nothing about the world but turn a vague goal into a question a manager can act on. Some are mechanisms argued from logic, like the loop in which a growing customer base lowers cost and lower cost attracts customers; I can give no experiment that establishes such a loop, but I can state each link so the reader can judge whether any fails. Some are findings from populations, such as the persistence of returns across thousands of firms or the effect of a randomized management intervention across dozens of plants; these are the strongest evidence the book has, and they say what is true on average, never what is true of this firm or this person. And some are stories, including Phelps's ankles and including my own, which can illustrate a mechanism and cannot establish it. When I tell a story, I will say what mechanism it illustrates, what observation supports the mechanism apart from the result, and what else might explain the result. Where I cannot, the story is decoration and I will say so.
My own career is one of the running cases in this book, alongside two companies I co-founded, Xootr and MakerStock, because I know what was decided and what was merely survived, which is more than I can know about Amazon. My own effort is vivid to me. What someone else would have made of my opportunities is not.
Unequal and Unjust
The second paragraph of the United States Declaration of Independence holds it to be self-evident that all men are created equal. In the ordinary sense of the word, nothing of the kind is evident. Boulders, snowflakes, dogs, companies, and countries differ from one another along every dimension anyone has measured, and so do people. Van den Driessche's rivals did not start the race with her engine, her years in the mud, or her team. Most of them did not start with a motor either.
A large part of who we are was settled before we could choose anything. The genes, the family, the country, the decade, the labor market in the year of the first job search. Economists following Canadian university graduates found that those who entered the labor market in a recession earned less for about ten years than those who graduated a few years earlier or later. They started at lower-paying employers and recovered only partly, by moving to better ones. The graduates did nothing to earn or to deserve the difference. Phelps did nothing to acquire his wingspan. LeMond did nothing to acquire his lungs.
Unequal is a description. Unjust is a judgment. An inherited advantage is not a wrong done to anyone. A motor in the seat tube is. In between lie the advantages that are perfectly legal and that still raise the question of what a society owes to those without them. I hold two beliefs on this. Every person and every organization should have the opportunity to prove themselves and thereby earn access to further opportunities. That is equality of opportunity, as distinct from equality of outcomes, and no society fully provides it. Norway does not give every teenager a place at its Olympic training center. It does give nearly every child a pair of cross-country skis and a local race, which can lead to the next race, and the next. The second belief is that because so much of the inequality that shapes well-being is unchosen, people acting together have an obligation to mitigate it, especially where fundamental needs are concerned.
Organizations are, in one respect, more fairly placed than people. A person cannot change their VO2 max by much. A company can change almost everything about itself, given time and money, and most of the advantages this book describes are of the kind that a company builds rather than inherits. Even so, companies are born unequal. Some start with a founder's reputation, a parent's customers, or a patent. A company can be a month old while its most useful resources are decades old, and the date of incorporation says little about how far back to look for the origins of its advantage. The accident of being first into a market that later turned out to matter is as unearned as a wingspan.
The rest of this book asks four questions, for a company and for a person. What improves your prospects? What stops others from acquiring the same thing? What does an observed result reveal about either? And given the answers, what should you build next? The thing that improves prospects and resists acquisition is the alpha asset, and chapter 2 defines it. A base of two hundred university workshops that a small materials company acquired one at a time over two years is one. So is the purchasing scale that lets the largest restaurant chain buy beef for less than any rival, and so is a name on a basketball shoe. Chapter 3 takes up what results reveal. Chapter 4 shows how an early result, lucky or not, becomes an asset by changing the opportunities that follow. The chapters after that are about building, choosing where to compete, knowing when an advantage has stopped working, and what success entitles anyone to claim.
A rider who knows which of their advantages were given and which were built still has to decide how to train before the next season. The rest of us face the same decision with less convenient finish lines, and the decision has the same form: given the advantages you have and the ones you lack, what should you build next?
The motor in the seat tube settled Van den Driessche's case. The rest of us are not scanned. We are judged on results, by boards and investors and employers and by ourselves, and the judgment assumes that results reveal what produced them. The chapters that follow are about how often that assumption holds, and what to do on the many occasions when it does not.
Notes
passed it over more than a hundred bicycles Union Cycliste Internationale, "The UCI announces Disciplinary Commission decision in the case of Femke Van den Driessche", April 26, 2016. The release states that the bike was scanned with the UCI's new magnetic resonance testing in the rider's pit area during the women's under-23 race, that the motor and battery were concealed in the seat tube with a Bluetooth switch under the handlebar tape, and that over 100 bikes were scanned at the event. The page as currently served omits the decision text; the sanction details are as reported in "Breaking News: Van den Driessche Receives Six-Year Suspension", Cyclocross Magazine, April 2016, and "Mechanical doping: six-year ban for Femke Van den Driessche", road.cc, April 2016.
The presence of one in a bike at a race is the offense The UCI charged technological fraud under articles 1.3.010 and 12.1.013bis of its regulations, which make the rider liable for the presence of a motor in a bike at a race regardless of intent or use. Cyclocross Magazine, as above.
banned her for six years Suspension from October 11, 2015 through October 10, 2021; fine of CHF 20,000; disqualification of all results from October 11, 2015 onward, including the 2015 European and 2016 Belgian under-23 titles; return of medals and prize money; payment of costs. Cyclocross Magazine and road.cc, as above.
In the HERITAGE Family Study Claude Bouchard, Ping An, Treva Rice, James S. Skinner, Jack H. Wilmore, Jacques Gagnon, Louis Pérusse, Arthur S. Leon, and D. C. Rao, "Familial aggregation of VO2max response to exercise training: results from the HERITAGE Family Study," Journal of Applied Physiology 87, no. 3 (1999): 1003–1008, doi:10.1152/jappl.1999.87.3.1003. The abstract reports 481 sedentary adult Caucasians from 98 two-generation families, a 20-week standardized program, a mean increase in VO2max of approximately 400 ml/min, individual responses ranging from little or no gain to more than 1.0 l/min, and a maximal heritability estimate of 47 percent. The gains are absolute, not weight-normalized.
An untrained young woman typically measures in the high thirties Typical values of about 38 ml/kg/min for untrained young women and about 45 for untrained young men, and the LeMond (92.5) and Dæhlie (96.0) figures, are as compiled at "World Best VO2max Scores", Topend Sports. The compilation warns that not all values are verified and that results depend on equipment, calibration, and protocol; another source gives Dæhlie's best as 90. The text treats them as a range for that reason.
won eight Olympic gold medals in the 1990s Guinness World Records, "Most Winter Olympic gold medals won in cross-country skiing (male)": eight golds for Bjørn Dæhlie of Norway from 1992 to 1998, twelve medals in all.
reported as 80 inches against a height of 76 "What Makes Michael Phelps So Good?", Scientific American, 2008 (updated August 2016). The article attributes the wingspan figure to the Toronto Sun and the shoe-size and ankle-flexibility claims to a Detroit News blog, and quotes exercise physiologist Weiner doubting that Phelps is markedly more flexible than rival swimmers and suggesting fitness as the main difference.
earned less for about ten years Philip Oreopoulos, Till von Wachter, and Andrew Heisz, "The Short- and Long-Term Career Effects of Graduating in a Recession," American Economic Journal: Applied Economics 4, no. 1 (2012): 1–29, doi:10.1257/app.4.1.1. The abstract reports persistent earnings declines lasting ten years, initial placement at lower-paying employers, and partial recovery through moves to better firms, with less advantaged graduates affected more. The Canadian sample is identified in the paper's data section.
holds it to be self-evident that all men are created equal Declaration of Independence, July 4, 1776, second paragraph.