Business Model Innovation in Data Competitions: Insights from the 2024 ADRENALIN Load Disaggregation Challenge

Data competitions have become an essential mechanism for fostering open innovation in domains such as computer vision, natural language processing, and energy informatics. Despite their growing prominence, the underpinning business models that sustain these competitions remain underexplored, leading to uncertain long-term adoption and limited value capture. This paper addresses these gaps by examining the 2024 ADRENALIN Load Disaggregation Competition as a case study of both technical and business model innovation within energy informatics. The research adopts an embedded single-case design, supported by multiple data sources and a comparative analysis with the digital gaming industry. Findings indicate that integrating multi-sided platform principles—derived from gaming ecosystems—can catalyze scalable Non-Intrusive Load Monitoring (NILM) solutions, enhance stakeholder engagement, and establish viable revenue pathways. Empirical results show that unsupervised NILM algorithms developed under zero ground truth constraints achieved NMAE values below 0.30, demonstrating high performance across diverse building types and climates. The study concludes that a hybrid revenue approach, combining sponsorships, IP licensing, and post-competition collaborations, enables sustainable operations beyond the typical prize-based model. These insights contribute to both academic literature and practical guidelines, illustrating how data competitions can evolve into robust, multi-stakeholder ecosystems with enduring economic and societal benefits.