Chapter 1: What a Rival Lets You Discover
Prologue: A Customer Asks for Another Quote
The customer has not complained. She has asked when the bicycle could be ready and whether the estimate includes everything she is likely to need. Then she says she will ask another shop. Sabine watches her leave with the bicycle. There is no sale to enter in the till, and no reliable way of telling whether the customer will return.
Aster Cycle Works is a hypothetical repair business. Its owner, Sabine, has a workshop, a reputation to maintain, and limited working hours. Customers arrive with needs that do not fit neatly around work already booked. Taking this customer's bicycle would require her to commit some of that capacity before knowing what the next person will ask for. The estimate carries those obligations into a comparison Sabine cannot control: another shop may have a different workload and be able to make a different promise.
Sabine might conclude that her estimate was too high. That is possible. The customer might instead need the bicycle sooner, dislike uncertainty about the final bill, or want an explanation she can understand without knowing much about repairs. A competing shop could be cheaper and less suitable. It could also have found a way to provide a service that Sabine has underestimated. She cannot discover which explanation fits simply by defending her costs.
Her costs are real. So are the customer's alternatives.
The encounter creates a problem of interpretation before it creates a problem of strategy. A lost sale contains remarkably little information by itself. Sabine does not see the other shop's workload, the promises it makes, or the customer's eventual experience. Cutting every estimate would answer a question she has not yet resolved. Maintaining every practice would be equally presumptuous. Competition exposes a decision to correction without supplying a complete explanation of the error.
This is why a market cannot be understood merely by counting the sellers in it. Another workshop may be close enough to appear in a search result but have no capacity for new work. Its opening hours may be unusable for this customer. A distant business may be a credible alternative if collection is straightforward. A customer who depends on the bicycle for daily travel faces a different comparison from someone who can leave it unused. The relevant choice takes shape around an actual need, rather than the boundary of a trade directory.
Suppose the customer returns. Sabine has won the work, but the return alone does not establish that hers was the best offer. The other business might have been closed. The customer might have been unable to make sense of its estimate. Habit might have settled the decision. Sabine should welcome the work without treating it as a certificate of excellence. A sale can reward good service; it can also reflect the difficulty of finding anything else.
From the customer's side, comparison is work. She must make time, explain the problem again, and decide whether differences between estimates reflect different services or different prices for the same service. Technical uncertainty leaves room for misunderstanding even where both shops act honestly. Making the quoted number smaller is therefore only one way a rival can improve the customer's position. Explaining its scope, keeping an appointment, or making an uncertain completion date less uncertain may matter more.
None of this makes price secondary in every case. A customer with little spare income may care about a difference that Sabine regards as modest. A business that cannot cover its costs cannot indefinitely provide the service. Those pressures meet at the counter without cancelling one another. Calling the result competitive does not tell us whether a particular bargain is affordable, sustainable, or well understood.
Sabine also buys things. She compares suppliers, chooses services for her business, and decides when the disruption of changing an arrangement is worth enduring. In one transaction she wants loyalty; in another she wants an easy exit. Her preferences need not be inconsistent. They reveal why an account of competition built entirely from the seller's viewpoint is incomplete. The freedom that disciplines her business may protect it elsewhere.
The chapters that follow move out from such encounters to more complicated arrangements: an entrant raising money before earning a reputation, a platform deciding whose offer becomes visible, a merger promising savings while removing a choice, software reacting to prices faster than its users can follow. The scale changes. So does the evidence needed to understand what is happening. A theory of discovery cannot settle the effects of a particular merger. A historical cartel cannot tell us that every similar price movement reflects an agreement.
Throughout, the question is what makes an alternative usable. Somebody must be able to develop it, finance it, bring it to customers, and survive long enough to be judged. Customers must be able to recognize the offer and bear the cost of choosing it. Rules can help establish these conditions. They can also obstruct them. A safeguard deserves examination in terms of the problem it addresses and the alternatives it excludes, rather than approval or rejection because it is a rule.
Back at Aster Cycle Works, Sabine can do something smaller than solve the market. She can review whether her estimate explains the work, whether her promised timing reflects available capacity, and whether customers hear about changes early enough to respond. These actions may earn more business. They may reveal that some work belongs elsewhere. The latter possibility is uncomfortable for an owner whose rent still falls due, yet it is part of what customers gain from having a choice.
For now, the customer has her bicycle and Sabine has an unaccepted estimate. The comparison could bring the customer back, take her elsewhere, or lead her to postpone the work. Sabine cannot ask her own accounts which explanation fits. Before the next estimate, she has to decide what she knows about the service people want and what she has merely grown accustomed to offering. Another workshop has given the customer somewhere to test the difference.
Chapter 1: What a Rival Lets You Discover
A customer can be dissatisfied without knowing what would satisfy her. The portion feels wrong, the wait consumes too much of her break, or the seller treats a modest request as a nuisance. She may continue buying because she assumes the alternatives would be much the same. Another offer changes more than the destination of her money. It can make her old dissatisfaction intelligible. She discovers that a shorter wait was possible, that she wanted a smaller portion rather than a cheaper large one, or that a familiar inconvenience belonged to one seller rather than to the service itself.
Rivalry supplies occasions for learning. A disappointing meal can clarify why the established vendor deserves a higher charge; an unfamiliar collection arrangement can reveal that convenience mattered more than the customer realized. The information arrives through use, and its value depends on what someone can do with it afterward. If the customer can return to her previous vendor or try a different combination next time, a modest mistake need not become an enduring commitment. If the apparent trial conceals a continuing obligation, the same experiment becomes more costly. Discovery has terms, just as the service does.
Consider a hypothetical street food market. Several vendors offer lunches to people working nearby. One vendor decides to prepare smaller meals that can be collected quickly. Nothing in this example establishes that demand exists. The vendor has formed a conjecture: some customers would exchange variety and quantity for a dependable return to work. Preparing the offer requires a commitment before the conjecture can be tested. Ingredients, space, labor, and the owner's attention are put at risk. Customers then decide whether the proposed exchange fits their lives.
If they buy, the owner still has several explanations to separate. The attraction might be speed, novelty, location, or an introductory concession. A busy first week cannot identify the lasting source of demand. Nor can a quiet first week establish that the underlying idea is worthless. Customers may not yet know the stall exists, may distrust a new vendor, or may encounter the offer at the wrong hour. Discovery is not a clean vote in which each transaction reports a single reason. It requires interpretation by people who have interests in the answer.
An established vendor watching the new stall faces a different learning problem. A lost customer provides a warning without providing a complete diagnosis. Cutting the price of a large meal will not necessarily recover a customer who needs to leave sooner. Extending the menu may worsen the wait. The incumbent can react vigorously and still misunderstand the threat. Effective rivalry therefore depends on more than an incentive to respond. It depends on whether the response addresses the reason another offer became attractive.
Knowledge before and after an offer
In his essay on competition, Friedrich Hayek challenged analysis that starts with knowledge already adjusted among market participants. He treated competition as a process through which people learn about costs, wants, and possible suppliers, rather than merely a state in which informed sellers face informed buyers. The essay is a theoretical and normative argument, originally a lecture, not an experiment showing that any particular market will discover the best solution.[1-1]
The practical question opened by that argument is straightforward. What did the participants know before another offer could be made, and what could they learn afterward? At the lunch market, the established vendors may know their own preparation routines well. They may know regular customers by sight. Yet they may have little reason to discover how many potential customers stay away because the routine is inconvenient. Those absent customers cannot express their preference through purchases that never occur. A different offer gives some of them a reason to enter the market.
The discovery is consequently not confined to transferring existing buyers. A vendor may learn that a group previously treated as uninterested was unwilling to buy on the available terms. Demand becomes visible when someone changes those terms. This does not mean every proposed product creates a valuable market. The willingness to try something can be brief, and a transaction may displace another activity whose value is difficult to see. It does mean that yesterday's sales are an incomplete description of tomorrow's possibilities.
Costs also contain unknowns. An owner knows what the current routine consumes but may not know the least expensive reliable way of producing the service customers actually want. The smaller meal could reduce preparation complexity while increasing packaging work. Faster collection could require a worker to arrive earlier, moving costs out of the customer's view rather than eliminating them. A practice that looks efficient during a quiet period may fail when demand becomes uneven. An honest discovery process includes these disappointments; otherwise the apparent improvement rests on unfinished work someone else must absorb.
A static comparison remains useful. Holding the offer and circumstances reasonably constant can help a customer identify a price difference or help an owner understand a cost. The mistake is to let the comparison erase the decisions that made those circumstances possible. If a new routine changes quality, availability, or the kinds of customers served, comparing prices alone can misdescribe the change. Conversely, invoking a dynamic process cannot excuse every current injury. A promise of future discovery is weak comfort to a customer who paid for a service that was never delivered.
There are two different questions here: whether an offer is acceptable now and whether people remain able to propose a better one. A market can perform tolerably on the first question while closing the second. Customers receive a familiar service, but no one can test an alternative. It can also remain open to experimentation while treating current customers badly. Entry by itself does not ensure truthful descriptions, dependable delivery, or accessible remedies. The capacity to learn needs a workable relationship between a claim and what happens when someone relies on it.
The customer's contribution
The customer does not simply wait for entrepreneurs to reveal the world. Comparing offers requires work. She must notice the alternative, decide which features matter, and judge how much uncertainty to accept. At the street market she can observe a line, read a menu, or try a meal. Some features are immediately available; others emerge only through repeated use. A single satisfactory purchase says little about whether the vendor can maintain the service during a difficult week. Familiarity therefore has a legitimate value even when another offer appears cheaper.
This value complicates the usual portrait of loyalty. A returning customer may be complacent, but she may also be conserving scarce attention. If lunch occupies a small part of a demanding day, searching for a marginally better option may cost more than it saves. Her decision to remain is not proof that all offers are equivalent. Nor does it prove that the incumbent has coerced her. It may reflect an entirely sensible limit on the amount of comparison she wants to undertake.
Other customers can indirectly benefit from the few who do compare. If some buyers are willing and able to leave, an incumbent may improve terms broadly to retain them. But that benefit is conditional. A seller who can identify attentive customers might reserve better terms for them while leaving everyone else with an inferior offer. A demanding customer may also obtain a personal exception rather than a general change. The existence of active comparison should therefore prompt a second inquiry: who receives its gains?
Even an ordinary comparison can be difficult to carry across sellers. One menu describes the main item; another includes a side and a drink. One vendor promises readiness at a particular time; another gives an estimate. One accepts an order only after payment; another lets a customer inspect the available food first. Treating these offers as identical quantities with different prices would discard much of what the customer is choosing. Yet treating every detail as incomparable would protect sellers from scrutiny. A useful comparison preserves the differences that affect the customer's decision while exposing terms that merely obscure it.
The vendor can help by making a commitment legible. The size of the meal, the collection arrangement, and the limits of a promise can be stated clearly. This is not a demand that every seller offer the same service. It makes diversity easier to evaluate. When descriptions become deliberately elusive, the customer may spend her effort decoding the offer rather than deciding whether she wants it. More apparent choice then produces more work without a corresponding improvement in control.
Customers also make mistakes. Someone may choose the faster meal and discover that eating less leaves the afternoon unpleasant. Another may attribute a delay to poor organization when an unusual interruption caused it. A fair account of competition cannot convert each purchase into a verdict about quality. What gives repeated choice its force is the opportunity to correct a judgment. That opportunity weakens when an initial trial commits the buyer for too long, when the relevant outcome arrives after the seller has disappeared, or when the same disappointment is difficult to describe to anyone else.
Mistakes that teach, mistakes that persist
The hypothetical new vendor has a reason to distinguish enthusiasm from repeat demand. A tempting response to early uncertainty is to expand immediately, taking the first signs of interest as confirmation. Expansion changes the experiment. The owner must coordinate more work, employ people who did not devise the routine, and serve customers who heard a simplified account of the promise. A useful small operation can become an unreliable larger one without the original idea having been foolish. Scale tests a different set of capabilities.
Suppose the owner instead keeps the operation narrow while learning which parts matter. This reduces some risks but creates others. Limited capacity may turn customers away and conceal the size of potential demand. The incumbent may copy the visible feature before the entrant can recover the cost of finding it. A patient experimenter can lose to a quick adapter. The competitive process does not distribute rewards in exact proportion to intellectual contribution. Timing, capital, location, and the ability to execute affect who keeps the proceeds of a discovery.
That unfairness in private reward does not establish that customers gained nothing. It does establish that a theory of learning must also explain why people will keep paying to learn. If useful experiments are easy to copy and expensive to conduct, too few may be attempted. If early success grants protection far beyond what investment requires, later improvements may be excluded. The relevant terms vary with the kind of knowledge involved. A preparation routine, a reputation, and a readily visible service feature do not have the same defenses or the same costs of imitation.
Failure can produce useful information, but the usefulness should not be exaggerated. An owner who loses savings may have learned something personally devastating that others cannot interpret. Observers rarely see all the choices that led to failure. They may conclude that demand was absent when the actual problem was an inconvenient opening time. They may imitate a successful-looking feature while missing a hidden cost. For learning to spread, people need some way to connect outcomes to decisions. A procession of openings and closures is not, by itself, evidence that mistakes are being corrected efficiently.
There is also a selection problem in the stories people tell afterward. The surviving vendor can describe persistence as insight because the outcome makes the earlier risk look purposeful. A failed vendor's similar persistence can be described as denial. Neither description tells us what could reasonably have been known when the decision was made. Judging an experiment requires reconstructing the alternatives available at that point, including the cost of waiting for more evidence. Certainty acquired afterward is not a resource the original decision maker possessed.
Repeated mistakes may persist for reasons unrelated to ignorance. An owner can understand that a routine inconveniences customers yet retain it because changing the routine would require paying workers differently or surrendering a personal preference. Another seller may understand a neglected demand but lack access to an appropriate site. A customer may know a better option but be unable to reach it during a short break. Information is necessary for many improvements; it is not a substitute for the capacity to act on what has been learned.
When the incumbent does the experimenting
The established vendor should not be assigned the role of obstruction by definition. Existing sales can fund experiments. Familiar staff can identify recurring problems, and a reputation may persuade customers to try an unfamiliar variation. The owner may know enough about demand to distinguish a passing request from a recurring need. A new entrant, meanwhile, may underestimate the difficulty of a service the incumbent has learned to provide reliably. Competition can bring a welcome challenge without giving the challenger superior judgment.
Imagine that the incumbent introduces a clear collection option before the smaller-meal stall opens. The change could reflect independent learning from customers. It could reflect the prospect of entry. It could be a mixture of the two. From outside, identical behavior supports several explanations. Calling the improvement evidence of entrepreneurial virtue or evidence of fear would go beyond what the example establishes. The customer's gain does not depend on settling that motive, although understanding it would matter when judging whether future improvement requires a credible rival.
The incumbent's ability to experiment can also affect which experiments survive. Suppose the established vendor has several profitable offerings and can sustain an unprofitable quick-collection service while the entrant depends on that service alone. A price cut may reflect an efficient use of existing resources. It may instead make a promising new operation unable to continue long enough for customers to judge it. To distinguish those possibilities, an observer would need to follow the relevant costs, the scope and duration of the response, and the conditions under which customers could still encounter the entrant. Sympathy for the smaller business supplies none of that information. Nor does the incumbent's ability to describe its response as vigorous competition. The question concerns the opportunity to test a service, including whether the test has been defeated by a better offer or by an obstruction to reaching customers.
The customer may be unable to judge those questions while choosing lunch. She can compare the service she receives more readily than the reasons a vendor charges a particular amount. That difference places limits on what her purchase can establish about the wider contest. If the entrant disappears, her next choice comes from a smaller set of offers regardless of whether she welcomed the earlier discount. Understanding the change requires evidence about what prevented the entrant from continuing. The individual buyer's sensible decision at the counter cannot supply an answer on its own.
An incumbent can improve without losing many customers if the prospect of loss becomes credible. But credibility has a material basis. Someone must be able to enter, serve, and remain long enough to be considered. A vague statement that dissatisfied customers could always go elsewhere is inadequate when elsewhere cannot receive them. The alternative must survive the customer's actual constraints. A service offered at an unreachable time is not an effective substitute for a person whose schedule cannot change.
This makes available capacity part of information. Discovering an attractive seller is of little immediate use if that seller cannot accept the order. A full stall may discipline its rival only after it expands, after customers adjust their plans, or after another vendor copies the idea. Until then, visible superiority and effective competitive pressure can diverge. The disappointed buyer knows what she would prefer but still has to choose from what can actually be supplied.
Learning under unequal circumstances
People arrive at a market with different resources for bearing a mistake. A customer with flexible time can try an unfamiliar stall and recover from a disappointing purchase. Someone with a rigid schedule may need reliability more urgently and be less able to test it. The same new offer can expand one person's choices while remaining unusable for another. A general increase in variety therefore says less than it appears to say about the distribution of practical alternatives.
The seller also learns from an uneven sample. Customers who speak readily may shape the owner's understanding more than those who quietly leave. An offer designed around visible complaints can neglect people who never approached the stall. Complaints can reveal a defect, but silence cannot be read as consent. The owner has to distinguish the loudness of a request from the willingness and ability to sustain the service it would require. No amount of attention eliminates this uncertainty; it changes which uncertainties are noticed.
A seller might make the offer increasingly elaborate to serve every expressed preference. More choice can then interfere with the original attraction. The quick collection service becomes difficult to prepare, and customers who valued simplicity bear the consequences. Learning sometimes means refusing a request because satisfying it would destroy another commitment. The refusal is not evidence that competition has ceased. Another seller may choose a different combination, allowing customers to decide which compromise suits them.
The market is especially informative when these compromises remain visible. A vendor who explains that a narrow menu supports dependable collection gives the customer something to evaluate. A vendor who promises speed, variety, low cost, and unlimited adjustment without acknowledging the tension invites a comparison on impossible terms. If that promise wins orders before its failure becomes visible, a careful seller can be disadvantaged by candor. The ability to punish broken promises, and the time over which reputation forms, then influence what kinds of offers are selected.
There is no reason to assume that all information produced by rivalry is truthful. Sellers can emphasize attractive features and conceal inconvenient ones. Customers can repeat impressions that are sincere but mistaken. A competitor can identify a real weakness while exaggerating its significance. More messages will not resolve a missing connection between claim and result. Where that connection is fragile, rivalry may reward skill in presentation more readily than improvement in service.
At the hypothetical market, even a simple description of waiting time can illustrate the difficulty. A vendor may describe the interval after food is ready, while a customer understands the interval after placing an order. Both can use the same ordinary word and believe they have been clear. When a dispute follows, it supplies little useful learning unless someone identifies the difference. The vendor who clarifies when the promise begins gives customers a better basis for choosing. The incumbent can then respond to the actual commitment rather than to an impressive but ambiguous claim. What has improved is the comparability of the offer. No additional stall was required, and no one had to dictate which meal should be sold. A small repair to the information available can make existing rivalry more consequential than another layer of promotion.
There are limits to how much clarification an individual customer can demand. She cannot inspect the entire routine during a short break. Useful information has to fit the decision she faces.
The owner must choose what to explain without hiding the terms that would change that decision. A concise description can save the buyer time while preserving the essential uncertainty: the service aims to be dependable, but unexpected variation remains possible. An overconfident guarantee may attract customers initially and make later disappointment harder to interpret. A description so qualified that it promises nothing is equally unhelpful. The commercial task is to make a commitment the vendor can stand behind and the customer can recognize when using the service. Competition learns through these commitments, including the occasions on which a promise has to be revised because the original conjecture was wrong.
Nor can every valuable outcome be inferred from the customer's willingness to pay. A lunch stall's operation affects workers, neighbors, and the use of shared space. The transaction records what buyer and seller agree to exchange; it does not automatically include every consequence. A competitive discovery may be commercially successful while imposing costs outside the transaction. Recognizing this does not make the information inside the transaction useless. It sets a boundary around what that information can establish.
Keeping a discovery usable
The most consequential discovery may concern the customer's own standard. Before encountering a credible alternative, she may regard a long wait as unavoidable. Afterward, she can ask why the wait occurs and whether it is worth accepting. The incumbent must now explain something previously taken for granted. A rival has changed the content of the conversation even if she returns to her usual stall. The old offer can survive, but it survives a more informed comparison.
That benefit can decay. The successful new vendor may become less attentive, the incumbent may stop responding, or the customer may lose the time needed to compare. A discovery is not an institutional achievement that remains secure once someone has made it. It must be carried through descriptions, skills, available capacity, and continuing opportunities to act. Information that cannot travel with the customer leaves her dependent on the place where she first encountered it.
For the hypothetical market, the decisive outcome is therefore not which stall deserves to win forever. It is whether customers can learn something consequential from another offer and use that knowledge without bearing unreasonable uncertainty. Vendors need room to test a different service, customers need room to revise a judgment, and promises need enough substance to survive contact with use. The work begins with a rival's conjecture. It becomes valuable when someone else can understand what was discovered and make a choice that was previously unavailable.
Notes and Sources
[1-1] Friedrich A. Hayek, “The Meaning of Competition,” originally the Stafford Little Lecture, May 20, 1946; published in Individualism and Economic Order (1948), pp. 92–106. Read in the authorized Econ Journal Watch reprint, 13(2), May 2016, pp. 360–371. This theoretical and normative argument supplies no causal estimate. https://competitionandappropriation.econ.ucla.edu/wp-content/uploads/sites/95/2018/06/HayekMeaningOfCompetition.pdf