Organizations & Markets

As generative AI systems rapidly improve, a key question emerges: how do users adapt to these changes, and when does such adaptation matter for realizing performance gains? This paper studies prompt adaptation—how users adjust their inputs in response to evolving model behavior—using a common experimental design applied to two preregistered tasks with 3,750 total participants who submitted nearly 37,000 prompts. We show that the importance of prompt adaptation depends critically on task structure.

The book gives evidence and advice for leveraging values as a concrete way to improve outcomes in leadership and life.  The first part of the book is about leveraging values as an individual, the second half is about organizational values.  The audience is thoughtful students of business, leaders, and scholars.

There has been growing interest in applying reinforcement learning (RL) to inventory management, either by optimizing over temporal transitions or by learning directly from full historical demand trajectories. This contrasts sharply with classical data-driven approaches, which first estimate demand distributions from past data and then compute well-structured optimal policies via dynamic programming.

Organizations emphasize specific production practices to deal with authenticity pressures, but the practices that signal authenticity to audiences must be continually adapted when production environments are unstable. Changes in the environment can make production practices suddenly infeasible, compelling organizations to perform in different ways the highly visible practices that audiences have come to associate with authenticity.

Entering new markets is crucial for technology startups to scale, yet these ventures often face high uncertainty about demand in these markets. This study examines how the composition of initial users shapes startups’ new market growth amid such uncertainty. It theorizes that startups face a learning tradeoff when targeting a foreign market: Local initial users, who are more familiar to the startups, provide clearer signals due to shared language and norms; however, more representative foreign users provide more transferable insights about the target market.

We study the dynamic pooling of multiple orders into a single trip, a strategy widely adopted by online delivery platforms. When an order has to be dispatched, the platform must determine which (if any) of the available orders to pool it with, weighing the immediate efficiency gains against the uncertain, differential benefits of holding each order for future pooling opportunities. In this paper, we demonstrate the effectiveness of using the delivery distance as a proxy for opportunity cost via a potential-based greedy algorithm (PB).

We construct a new quality-adjusted commercial real estate rent index for U.S. office, retail, and industrial markets using more than one million CompStak lease transactions from 2010-2025. A hierarchical hedonic framework with building-, block-, and ZIP-level fixed effects allows us to control for both observable and unobserved quality, producing quality-adjusted rent indices.

We introduce a novel measure of corporate hierarchies for over 3,100 U.S. public firms. This measure is obtained from online resumes of 7 million employees and a network estimation technique that allows us to identify hierarchical layers. Equipped with this measure, we document several facts about corporate hierarchies. Firms have on average ten hierarchical layers and a pyramidal organizational structure. More hierarchical firms have a more educated workforce, higher internal promotion rates, and longer employee tenure.

Deep Reinforcement Learning (DRL) provides a general-purpose methodology for training inventory policies that can leverage big data and compute. However, off-the-shelf implementations of DRL have seen mixed success, often plagued by high sensitivity to the hyperparameters used during training. In this paper, we show that by imposing policy regularizations, grounded in classical inventory concepts such as "Base Stock", we can significantly accelerate hyperparameter tuning and improve the final performance of several DRL methods.