Physician Segmentation

Case Study


  • A small molecule franchise observed a surprisingly wide variety of treatment patterns across physicians in high-level market research
  • Franchise growth was stalling short of forecast expectations
  • The client product was apparently not being used in many appropriate patients for which it should arguably have been considered the best option
  • Putnam was tasked with executing a physician segmentation to help better understand the physician landscape and product positioning and messaging that would better resonate with physicians

Putnam Approach:

  • Review recent clinical and epidemiological literature to confirm prevalence and sketch out profiles for archetypal patients
  • Review map of the patient accrual process and formulate hypothesis on key steps where appropriate patients might be “falling through the cracks”
  • Claims data analysis to further quantify the treatment landscape and identify behavioral factors that appeared to distinguish physician sets
  • Small-N qualitative research to further explore usage patterns, archetypal patient profiles, refine hypotheses, and directionally validate questions for quantitative research
  • Large-N quantitative research to capture demographic, psychographic (attitudinal), and behavioral (treatment preferences, prescribing) across a representative set of physicians
  • Sophisticated regression analysis of quant data to identify segments
  • Integrated analysis of claims and quant data to refine and validate segments
  • Segment-specific recommendations on achieving product usage growth


  • Major sampling and question-design flaws identified in prior market research and claims analysis methodologies, and fixed for forward-going analyses
  • More than $80M in upside opportunity identified, with $35M of this considered “low-hanging fruit” that might be achieved by concrete, prioritized initiatives within 12 months
  • Revamping of promotional aids commenced, following insights in the quant research


Strategy for the Life Sciences