Firms Can Still Make Money When They Know Nothing About Consumers

Launching a new product can be a leap in the dark. From video games to streaming platforms, from apps to consumer electronics, companies are constantly asked to commit to design decisions long before they truly understand who their customers are or what they want. Surveys are incomplete, data are noisy, and tastes change fast. What happens when firms must design and price a product while knowing almost nothing about consumer preferences?

A recent paper in the Journal of Economic Theory tackles this question, offering a striking and counterintuitive answer: when uncertainty is extreme, the best strategy is not to guess better, but to randomize intelligently.

When ignorance is the rule, not the exception

The paper, “Robust Product Design and Pricing” by Nenad Kos (Department of Economics, Bocconi University; IGIER, CEPR, and PERICLES) and Kyungmin Kim (Emory University), studies a monopolist who must decide both what product to design and how much to charge — without knowing the distribution of consumer tastes.

This situation is far from hypothetical. The paper opens with a real-world example: Nintendo’s decision in 1989 to bundle the Game Boy either with Super Mario Land or Tetris. With little information about who would buy a handheld console, Nintendo made a design choice under deep uncertainty. The gamble paid off — but could economic theory explain why some design strategies are safer than others?

To do so, the authors rely on a classic framework in economics: the Hotelling model, where consumers differ in tastes along a line (think sweetness of cider, or style of a product), and products that are further from a consumer’s ideal point are less appealing.

But unlike standard models, here the firm has no information at all about how consumers are distributed along that line.

Designing against the worst case

Instead of trying to predict what consumers might want, the authors take a radically different approach: they assume everything that can go wrong, will go wrong. In their model, the firm acts as if an enemy force — “nature” — is free to choose the most unfavorable distribution of consumer tastes after the firm commits to a design and a price.

This does not mean that consumers are hostile, but that the firm refuses to rely on optimistic assumptions. Any concentration of demand, any imbalance in tastes, any unexpected clustering of consumers is treated as a real possibility. The firm therefore evaluates each strategy by asking a brutal question: what is the minimum profit I would earn in the worst conceivable market?

As the authors put it, the seller “maximizes her profit under the worst-case scenario” (p. 1). This mindset turns product design into a form of insurance. The goal is no longer to do exceptionally well in some markets, but to avoid doing disastrously badly in any of them.

Once the problem is framed this way, the logic of the optimal strategy becomes clearer. A firm that commits to a single design risks missing the market entirely if consumers’ tastes happen to lie elsewhere. By contrast, a firm that divides the market into several equal segments and randomly chooses which one to serve guarantees itself a minimum level of demand no matter how tastes are distributed. Randomization is not a sign of confusion — it is a deliberate way to neutralize the worst-case scenario. 

Why don’t profits collapse

One of the paper’s most important insights is that profits remain positive, even under total ignorance. This is remarkable because in simpler monopoly models without product design, profits can be driven arbitrarily close to zero when the seller lacks information. Here, design flexibility changes everything.

As the authors note, “The seller’s profit is bounded away from 0. This is in stark contrast to the monopoly model without product design” (p. 2). The reason is intuitive: when a firm can adjust where it positions the product, no single consumer type is intrinsically the worst one. Nature cannot concentrate all demand in a single disastrous location.

Is the uniform market assumption that bad?

Economists often assume that consumers are uniformly distributed, mostly for convenience. But when the loss in value from a poor match between the product and consumers’ tastes grows in a simple, linear way, the authors show that profits under a uniform market are never more than 12.5% higher than in the worst case — and often much closer. In some cases, the difference shrinks to about 4%.

This means that, surprisingly, the uniform distribution is almost as pessimistic as the worst possible one.

More products, same logic

What if the firm can design multiple product varieties instead of just one? The logic remains robust. If the firm can produce several versions, the optimal strategy is still to divide the market into equal segments. When the number of products is small, the firm randomly selects which segments to serve. Once it has enough products, it covers the entire market.

There’s a real world beyond the theory

The paper can be particularly appealing to industries where experimentation is cheap and information is scarce: digital content, online platforms, early-stage startups, and creative industries. When data are unreliable, hedging across designs can outperform confident but fragile bets. In an age obsessed with personalization and precision, Nenad Kos and Kyungmin Kim provide a sober reminder: sometimes the safest strategy is not to know more, but to insure yourself against knowing too little.