Modeling Demand Dynamics for Optimal Design of Evolving Products

Demand for modern products is inherently dynamic, driven by evolving customer preferences and continuous updating of software-enabled products. Traditional demand models, however, treat product design as a static optimization problem, identifying feature sets which maximize instantaneous demand without capturing transient customer behaviors and update-induced market fluctuations. This approach creates a gap for modern product designers and decision-makers seeking to strategically time product updates. To bridge this gap, we introduce a controls-based approach to modeling dynamic demand through two complementary frameworks. First, we establish a networked agent-based demand model leveraging decision field theory (DFT), a cognitive decision-making model, to capture individual-level deliberation processes. Second, recognizing that detailed individual-level data is often unavailable, we propose a hybrid group-level model combining epidemic spreading processes with discrete impulse events. We validate both modeling approaches using daily peak user-count data from two prominent live-service video games, demonstrating the capability for both models to capture and predict real-world demand dynamics. These models enable product developers and managers to strategically optimize the timing and content of updates, better anticipate engagement fluctuations, and sustain long-term user retention. Ultimately, this work establishes a foundation for decision-support systems that can dynamically adapt product design and update strategies in response to evolving market behaviors.