Which Opportunities to Pursue
A rare-disease focused company needed to evaluate multiple preclinical assets across a wide range of potential indications. The key question was not how to build a complex ranking model, but how to help leadership confidently prioritize the asset–indication combinations worth further investment, outside their existing pipeline.
The work was intended to support senior leadership and portfolio governance discussions ahead of upcoming preclinical pipeline decisions. Starting from a nearly blank slate, most relevant disease- and asset-level data had to be assembled from scratch. As is typical in rare disease, evidence was uneven, requiring structured judgment alongside limited epidemiological and outcomes data.
The engagement was framed as an opportunity assessment, aimed at comparing relative attractiveness without overstating the certainty of the inputs.
A purpose-built dataset was created spanning dozens of rare diseases and linked assets, capturing epidemiology, disease burden, treatment effectiveness, mortality, and competitive context. Client-generated inputs on expected clinical benefit and rationale strength for each asset–disease pair were incorporated into the same structure. Where data gaps existed, the project generated transparent, well-described, assumption-based estimates to support the assessment.
To reduce complexity, machine-learning clustering grouped similar asset–indication combinations into a small number of “opportunity types.” These groupings were based on decision-relevant variables such as clinical benefit potential, unmet need, disease severity, and competitive dynamics.
The outputs were translated into clear, decision-friendly visuals and narratives, allowing leadership to compare a manageable set of opportunity types rather than reviewing numerous individual options.
By making the underlying assumptions explicit, the work gave leadership a clearer basis for weighing trade-offs and conducting a more disciplined, evidence-based discussion.
The engagement delivered both immediate insights and a durable asset. All inputs were consolidated into a client-owned database, enabling ongoing updates as new data emerges. Analytical outputs can be refreshed through repeatable methods, providing a stable foundation for future portfolio decisions.
This is the kind of NPP work that reflects my approach: make assumptions explicit and bring structure to uncertainty so commercial recommendations are stronger without overreaching the evidence.