Optimizing What We Reveal: Information Design in Operations

Information is a resource, and companies and social service systems use this resource to support user and consumer decision-making. ISyE Associate Professor Krishnamurthy Iyer studies how to share information in ways that improve overall welfare and outcomes.

In brick-and-mortar stores, consumers make decisions based on what they can physically see—quality, quantity, and other visible cues. Online platforms, by contrast—ridesharing, e-commerce, and even hospital systems—can decide how much information to reveal and how to present it. Signals such as whether an item is a “best seller” or “top choice,” or estimates of wait times for a service, can meaningfully shape the decisions users make.

“Online platforms control and consciously reveal what information they want their users to see, and this information is often all users have to rely on to make an active decision,” Iyer explains.

Many hospital systems, for example, have implemented dashboards that show patients potential wait times, helping them decide which care center to visit and when.

“Without such a dashboard in place, people decide on their own, and they may over-subscribe to services, which makes wait times increase. Through a dashboard, a hospital can provide information to influence behavior so people with less urgent needs can find an alternative location or choose to arrive at a less busy time. This can curb load on the hospital system, improving welfare for the hospital staff and patients.”

How does the information provided influence user behavior, particularly in service-system settings? And how much information is too much—or too little?

To study these questions, Iyer develops new models that capture realistic aspects of these systems. These include structural constraints on how information is shared (centralized vs. decentralized), as well as limitations on what is known and what is learned over time. And because people often deviate from perfectly rational decision-making, his models also account for behaviors such as ambiguity aversion.

Wait time information, for instance, can be communicated in different ways: a point estimate, a range, or more abstract “signals” such as low versus high congestion. The format affects how users respond. At the same time, the information must remain truthful; otherwise, users will notice inconsistencies over time and adjust their behavior accordingly.

By combining theoretical tools from queueing theory, game theory, and economics, Iyer aims to characterize the “right” amount of information to provide in different operational settings.

“Game theory is one of my main research interests, and this particular research question is a practically informed application of game theoretic modeling. The challenge is that the novel models we create have to be realistic, and at the same time not overly complicated to allow for analysis.”

As society continues to shift toward virtual-first platforms, decisions about what information to share—about product availability, service quality, or wait times—can influence both individual behavior and broader outcomes like system performance, welfare, and satisfaction.