A New Frontier in Predictive Analytics
Presented by Oliver Wyman
As insurers collect more data than ever — from telematics and IoT to satellite imagery and climate indicators — they are getting a much clearer, richer picture of their risk exposure. This creates a choice: stick with traditional, constrained models that fit regulatory and business limits, or adopt less-constrained models to push predictive performance and uncover fresh insights. The best path is blended; using unconstrained approaches to set a performance ceiling, run gap analyses against constrained models, and then target practical, compliant improvements to close that gap.
“A New Frontier In Predictive Analytics,” by Nickolas Alvarado and Brett Nunes from the Oliver Wyman Actuarial P&C team, explores how using unconstrained models can empower risk management, drive innovation, and help insurers gain a competitive edge. It examines:
- The latest advancements in the insurance analytics landscape
- Constraints and limits of traditional models
- How to strategically use unconstrained models
- The importance of strong model governance
The full report also features examples of how three insurers used unconstrained models to improve workers’ compensation profitability analysis, loyalty program redemption analysis, and personal auto telematics behavior model performance. Uncovering new data-driven insights allowed these insurers to gain a better understanding of their business, which in turn resulted in more strategic decision-making to improve their bottom lines.
More broadly, the explosion of data in recent years, combined with the limitations of traditional models and the rise of sophisticated algorithms, presents an unparalleled opportunity for insurers. While business needs and regulatory requirements usually call for constrained models, insurers that leverage unconstrained models in a strategic way can drive substantial transformation.
By comparing unconstrained and constrained models, insurers can quantify the performance loss attributed to certain limitations. This insight is essential; it reveals the true cost of constraints and highlights areas where regulatory or business constraints may be unnecessarily hindering segmentation or value creation. Maximizing predictive power does not mean abandoning traditional methods — instead, it means strategically enhancing them with advanced techniques and a robust framework for governance and interpretability.