In 2018 I pulled from a financial database the records of every publicly traded restaurant company in the United States. For each one I computed its average return on invested capital over the ten years from 2006 to 2015, less its cost of capital. The list has familiar names on it: McDonald's, Starbucks, Yum Brands (which owns KFC and Taco Bell), Chipotle, Domino's, and several dozen smaller chains. As a group they earned, by my calculation, about six percentage points a year more than their capital cost them. But the spread was not evenly shared. The companies at the top of the list earned several times the group average, while those at the bottom destroyed value for a decade and survived anyway.
I am an engineer, not an economist, and my perhaps naive view of economic theory is that competition should erase such differences. A business that earns too much attracts imitators and entrants, supply rises, prices fall, and the excess profit disappears. On a trip to Istanbul I watched this view hold exactly. Fish bread is a street food sold from a small charcoal grill set on a couple of crates at the edge of a market. The vendor grills two pieces of fish, toasts a split loaf, and assembles a sandwich with olive oil, herbs and salt, all for the equivalent of about four dollars. I ate a lot of fish bread, and to my palate it was identical from one vendor to the next, uniformly excellent, and priced the same everywhere. If a vendor raised prices by ten percent, a new grill would appear on the next set of crates within the week. You can be pretty sure that no fish bread vendor in Istanbul earns much more than any other. (I concede that a tourist's reading of the fish bread economy may be incomplete. As a young man I spent a year in Naples watching teenagers sell cigarettes from folding tables. It took me most of that year to learn about the supply chain and the management hierarchy behind what looked like a free market.)
Very few markets in a modern economy look like the fish bread market. Across thousands of public companies, the spread between the return on capital and the cost of capital persists for decades. When the spread is decomposed statistically, a large part of it belongs to the individual firm rather than to its industry. Why does it persist? The marginal competitor, the firm on the sidelines that would enter if the business looked attractive enough, cannot get what the incumbent has. Yum Brands has something that a smaller chain like Potbelly cannot easily acquire.
Two Tests
By an advantage I mean a capability or a position that improves expected performance relative to the relevant rivals, in a particular setting, over a particular period. Expected, because a lucky result is not an advantage (chapter 1 was about the difference). Relative to rivals, because a capability that every competitor shares improves nobody's standing. In a setting, because the same capability can be decisive in one market and ordinary in another; a strong hand at bridge is nothing at poker. Over a period, because nearly every advantage erodes.
The thing that confers an advantage I call an alpha asset: a resource that an entity owns or controls that is (a) performance enhancing for the job the entity does and (b) hard for its rivals to acquire. I borrowed the term from finance. In the capital asset pricing model, a security's expected return is decomposed into beta, the part explained by exposure to the market as a whole, and alpha, the part specific to the security. Alpha assets are the underlying explanation for financial alpha, the reason a firm earns more than its exposure to the economy would predict.
The concept is not mine. It is a reframing of the dominant idea in the academic field of competitive strategy, the resource-based view of the firm, which Jay Barney set out in 1991. Jay, now at the University of Utah, is an affable and enthusiastic man with whom I once worked on a consulting project in healthcare. The resource-based view (RBV, in the alphabet soup of acronyms that specialized fields love) holds that a firm sustains an advantage when it controls VRIN resources. (Got it?) VRIN stands for valuable, rare, inimitable and non-substitutable. (Got it now?) I made a good-faith effort to use VRIN in an earlier book, but it is just too wonky for anyone outside a doctoral seminar. The resource-based view is one of the most powerful ideas in social science and one of the most poorly marketed. It needs a makeover, so let's relabel VRIN resources as alpha assets and simplify the four conditions to two: valuable and non-substitutable collapse into performance enhancing; rare and inimitable collapse into hard to acquire.
A resource is performance enhancing when its presence causes a significant improvement in results, with no good substitute, for the competitors who do the job. Three phrases in that definition need defending.
No good substitute. In a solo time trial, a race against the clock, there is no substitute for a light, stiff bicycle; a rider who turns up on a Dutch utility bike finishes last. There is no substitute for electricity in a factory. But there is a substitute for the Advil brand of anti-inflammatory, since Motrin, Nurofen and the generic bottle contain the same ibuprofen, and so the molecule is closer to being uniquely performance enhancing than the brand is.
Significant. The question is whether the resource accounts for much of the difference between winning and losing, not merely whether it helps. Water is critical to every rider; a dehydrated cyclist does not finish. An earpiece with a radio link to the team car is a small improvement, decisive on the right day, but a rider can win without it. Ibuprofen explains none of the variation among the top finishers.
For the competitors who do the job. In my 2022 draft of this definition I wrote "for most competitors," meaning to exclude idiosyncratic needs such as insulin, which is performance enhancing for a diabetic cyclist and irrelevant to the rest of the field. That exclusion was too strong, especially for people, because a resource can be performance enhancing only in combination with something else the holder has. A tough general counsel may be essential to a company run by a mercurial founder and surplus to a company run by a cautious one. In such cases the asset is the combination, and the combination can be rare even when each part is ordinary. So a better statement of the test is that the resource improves results for anyone doing this job who holds the complements it needs, with the complements counted as part of the asset.
Performance enhancement is necessary but not sufficient. A restaurant needs reliable electricity, accurate books, and compliance with labor law. None of these is a source of advantage; every competitor has them.
A resource is hard to acquire when a rival would need a large amount of time or money to obtain it, or could not obtain it at all. I don't have a precise threshold, but my rule of thumb is that an investment above ten percent of the rival's revenue, or a wait of five years or more, makes a resource hard to acquire.
McDonald's scale is hard to acquire; a smaller chain like In-N-Out could not match it with any plausible investment in any plausible time. Because scale is also performance enhancing, through purchasing, advertising and the cost of running a kitchen at volume, McDonald's scale is an alpha asset. The Nike name is hard to acquire, since no rival can use it, and put on a basketball shoe it would lift the sales of almost any shoe. It is an alpha asset. A cyclist's cardiovascular system, as chapter 1 showed, is partly inherited and only partly trainable, which makes it hard to acquire. Since it is also performance enhancing, it is an alpha asset too.
Some resources pass the second test and fail the first. Starbucks is headquartered in Seattle; it would take Luckin Coffee, a chain that opened in China, years and a great deal of money to move its headquarters to Seattle, and the move would not improve a single cup of coffee it sells. Or consider the name of Bjørn Dæhlie, the Norwegian cross-country skier. Because no other skier can use it, the name is hard to acquire; it did nothing for his speed on the trail. After he retired, he put the name on a line of sportswear, and in the Scandinavian market the name became an alpha asset for a company whose job is selling clothes.
Who Pays, and Who Gets Paid
The two tests leave out a third consideration: when a scarce, performance-enhancing asset is for sale at a price that fully reflects its value, its buyer earns nothing above the normal return. If everyone can see that a brand, a location or a patent is worth a hundred million dollars, the seller charges a hundred million, and the buyer's return on the purchase is ordinary. (An asset that can be bought at a fair price is, almost by definition, not hard to acquire.) Above-normal returns from buying assets go to the buyer who expects more from the asset than the market does, who already holds the complements that make it worth more in their hands, or who is lucky. Barney made this argument in 1986, five years before his paper on the resource-based view, under the heading of strategic factor markets. For that reason, most of what follows is about building alpha assets rather than buying them.
A related question is who collects the return. When a firm's advantage rests on a resource that a person controls, the person can bargain for the value. Consider a software company whose results depend on one engineer: the company has an alpha asset only in the sense that the engineer has one; it pays for the asset in compensation and in the risk that they leave. Russell Coff made this argument in 1999: value created by hard-to-imitate resources often accrues to employees and other stakeholders rather than appearing as unusual returns to shareholders. If you work for someone else, you are in the engineer's position. Your scarce capability is your alpha asset, and who captures its value depends on where you work and on what terms.
Important Is Not Alpha
Many things are important without conferring advantage. An enterprise software company cannot succeed without an effective sales process, a restaurant without competent operations, or a bank without timely and accurate accounting. None of these is a source of sustained advantage. They are the price of admission: every serious competitor does them competently, and the differences among competent performers are small.
Product falls into this category more often than product people like to hear. Although a good product is important for every company, it is an alpha asset only in a few settings. The first is the short window after a disequilibrium, when the new product is dramatically better than anything else available. The second is an industry where a patent or a regulatory approval makes the product itself exclusive, as in pharmaceuticals, and the third is the small number of businesses where making the product is so hard that the capability to make it is the asset, as in leading-edge semiconductor fabrication. Elsewhere, every competitor must make a good product and none makes a decisively better one. Chapter 7 is about using the window while it is open.
With the two tests in hand, a manager or a founder can audit a business. The appendix contains the worksheet I use in class. It lists the categories of resource that most often turn out to be alpha assets: the solution itself and the know-how behind it, the processes that deliver it, the brand, scale, the customer network, the organization's relationships upstream and downstream, its culture, and a residual category for location, balance sheet and the founder. For each resource you assign two scores from one to ten, rarity (how hard would it be for a rival to acquire?) and necessity (how much does the resource improve results, with how few substitutes?), and multiply them. In a worked example for McKinsey, the consulting firm, I give the firm's frameworks a rarity of three, since they are published and walk out the door with every departing consultant; its client relationships get a rarity of eight.
The scores are prompts for an argument, not measurements. A number on the worksheet means that someone has asserted a position about a resource that someone else can dispute with evidence. (That is more than I can say for most strategy discussions.) The worksheet also has a failure mode, the subject of chapter 3: if you fill it in after a company has succeeded, you will find alpha assets everywhere, because you already know the result. Two habits reduce the risk. First, score the resource by what a rival would have to spend to get it, a question about the rival rather than about the firm's history. Second, ask whether a venture investor who knew the firm's resources and not its results would fund a plan that depended on them.
Endowments
Chapter 1 said that organizations are born unequal; the two tests say how. A new company can start with an alpha asset it did nothing to build. In 2016 Microsoft decided to build its own business messaging product, Teams, rather than buy Slack. Because Microsoft had attached Teams to the installed base of Office 365, Teams reached twenty million daily users within three years, despite wide agreement (which I share) that Slack was the better product. The customer network was an endowment from the parent. When the singer John Legend co-founded a wine label, LVE, with Raymond Vineyards in Napa, Raymond made the wine; the alpha asset was Legend's name, which carried the label into mainstream distribution that a tiny fraction of new wineries ever reach. If you have that kind of name, congratulations. The rest of us start with speed, a better product for a while, and not much else.
A person's endowments are the same in kind and harder to change. By the same two tests, a capability or a position is an alpha asset for you if it improves your results in the work you want to do and if the people competing for that work cannot easily acquire it. For most of us the candidates are a skill that took a decade to build, a reputation, a set of relationships, and a position inside an organization that gives access to problems worth solving. You can score each one on the same worksheet. The difference from the corporate case is the objective. Whereas a firm's objective is, to a first approximation, return on invested capital, a person's objective is not obvious. A capability that would win a contest the person does not want to enter does them no good. (I am a competent administrator, and I dislike meetings.) Chapter 9 returns to the question of what an advantage is for.
Assets Stop Being Alpha
The 1989 Tour de France came down to a final time trial of 24.5 kilometers from Versailles into Paris. The race leader, Laurent Fignon, started the day fifty seconds ahead of the American rider Greg LeMond. LeMond rode with a teardrop aerodynamic helmet and a pair of clip-on handlebar extensions, recently introduced by Scott, that let him tuck his arms into a narrow position in front of the bike. Fignon rode bareheaded with his ponytail in the wind. LeMond beat him by fifty-eight seconds on the day and won the Tour by eight. Engineers who have since analyzed the ride agree that the equipment accounts for more than eight seconds. On that afternoon the helmet and the bars were alpha assets, but by the mid-1990s every professional rode a time trial with both.
How long does an asset remain hard to acquire? Almost never forever. Luckin Coffee opened in China with a mobile-first ordering system that had no counterpart at Starbucks. I once stood in a Luckin shop in Shanghai while a barista tried and failed to get the app onto my phone and finally gave me the coffee for free. For several years the system was an alpha asset; today every coffee chain in China takes orders by phone. (Luckin's ordering system outlasted its accounting; in 2020 the company admitted that a large part of its reported sales had been fabricated.)
The central tenet of microeconomics is that competition drives markets toward equilibrium. An equally serious discipline could have been built on the opposite observation: at the time scales on which people and companies live, the world is never in equilibrium, because friction of every kind keeps it out. Customers are slow to switch, know-how is slow to spread even inside one company, and rivals are slow to notice, slower to copy, and slowest to copy well. Advantage exists in the gap between a new asset and its imitation, and friction sets the length of the gap. Chapter 8 asks what happens when the gap closes.
Notes
which Jay Barney set out in 1991 Jay B. Barney, "Firm Resources and Sustained Competitive Advantage," Journal of Management 17, no. 1 (1991): 99–120. The VRIN conditions (valuable, rare, imperfectly imitable, non-substitutable) are developed there. The reframing into two tests is mine.
a large part of it belongs to the individual firm rather than to its industry Gabriel Hawawini, Venkat Subramanian, and Paul Verdin, "Is Performance Driven by Industry- or Firm-Specific Factors? A New Look at the Evidence," Strategic Management Journal 24, no. 1 (2003): 1–16. Their finding that firm effects dominate, and that a few exceptional firms account for much of the firm effect, was disputed by Gerry McNamara, Federico Aime, and Paul M. Vaaler, "Is Performance Driven by Industry- or Firm-Specific Factors? A Response to Hawawini, Subramanian, and Verdin," Strategic Management Journal 26, no. 11 (2005): 1075–1081, with a reply by Hawawini, Subramanian and Verdin in the same issue. The decompositions are statistical, not causal; the text claims only that persistent firm-level differences exist.
Barney made this argument in 1986 Jay B. Barney, "Strategic Factor Markets: Expectations, Luck, and Business Strategy," Management Science 32, no. 10 (1986): 1231–1241.
Russell Coff made this argument in 1999 Russell W. Coff, "When Competitive Advantage Doesn't Lead to Performance: The Resource-Based View and Stakeholder Bargaining Power," Organization Science 10, no. 2 (1999): 119–133.