Post by Università Bocconi

What should firms do when they know almost nothing about their customers? In markets where preferences are uncertain, data are incomplete and tastes change rapidly, the instinct is often to gather more information and refine predictions. Yet new research suggests a different answer: when #uncertainty is extreme, the safest strategy may be intelligent randomness. Launching a new product can be a leap in the dark. 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? Nenad Kos (Bocconi University), together with  Kyungmin Kim (Emory University), examines how a monopolist should design and price a product when the distribution of consumer tastes is unknown. Rather than relying on optimistic assumptions, the authors adopt a worst-case approach. The firm evaluates each design by asking what profit it would earn under the most unfavorable possible distribution of preferences. In other words, it maximizes profit under the worst-case scenario.    This perspective leads to a counterintuitive result. When a firm commits to a single design, it risks missing the market entirely if consumer tastes are concentrated elsewhere. By contrast, dividing the market into equal segments and randomizing across them ensures a minimum level of demand regardless of how preferences are distributed. #Randomization becomes a form of insurance against extreme uncertainty. In this setting, #DecisionMaking is less about predicting perfectly and more about avoiding catastrophic outcomes. One of the most striking findings is that profits remain positive even under total ignorance. Unlike simpler monopoly models, where lack of information can drive profits close to zero, the ability to adjust product design protects the firm from collapse. The flexibility to reposition the product prevents any single consumer type from becoming fatal. This insight reshapes how economists think about #MarketUncertainty and product strategy.    The analysis also challenges common #modeling assumptions. Economists often assume consumers are uniformly distributed for simplicity. The authors show that, under certain conditions, profits in a uniform market are surprisingly close to profits in the worst-case scenario. The difference is modest, which suggests that standard assumptions may already incorporate a considerable degree of pessimism. In industries such as digital content, platforms, apps and early-stage startups, experimentation is relatively inexpensive and information about demand is noisy. In such contexts, hedging across multiple designs may outperform confident but fragile bets. Sometimes the most robust strategy is not to know more, but to design in a way that protects against knowing too little.