Five months of structured market research and NPPES data analysis — six segments, each independently sized from primary sources.
Step 1 — Segment Architecture
Rejected the single-TAM approach. Built a six-segment framework where each segment is defined by organisational structure and buying behaviour — not industry vertical. A multi-location dental group and a multi-location physician group have the same buyer profile (standardisation across sites, multi-stakeholder approval, long cycle) even though the clinical workflows differ. Segmentation by structure, not specialty, allowed us to size each segment independently with appropriate ARPU assumptions and win rate context.
Step 2 — Source Validation Across 55+ Claims
Validated 55+ individual claims across six research domains: global market sizing, small practice data, multi-location and hospital counts, multi-specialty and startup projections, competitor pricing, and SaaS benchmarks. Material findings: EHR adoption was 83.6% for certified EHRs but 95% overall (different things, different implications for SAM); the US share of the global market was not 60% as assumed — healthcare IT analogues from Grand View Research, Polaris, and Fortune consistently show North America at 40–50%; the CAGR upper bound of 16% was unsupported — two sources confirmed 13.7–14.2%.
Step 3 — NPPES Full Dataset Analysis
Downloaded and analysed the CMS NPPES February 2026 dataset — 9.4M NPI records covering all US healthcare providers. For the multi-location segment specifically: extracted all Entity Type 2 organisations (physician group NPIs) with 2–9 registered practice addresses. Result: 15,047 confirmed organisations — not the 30,000–50,000 previously assumed. The prior estimate had conflated multi-physician practices (multiple doctors at one address) with true multi-location practices (same organisation, multiple physical sites). The AHRQ figure of 35,448 that the estimate was based on measures multi-physician groups, not multi-location practices. The $180M TAM overstatement flowed directly from this single definitional error. Expanded the NPPES analysis to all 16 provider types (not just physicians) — total active provider NPIs: 5,431,245 (3.2x the physician-only count), surfacing a separate expansion segment of 162,825 multi-physician single-location addresses and a multi-location all-provider upper bound of 34,839 orgs.
Step 4 — ARPU Benchmarking & Practice Count Calibration
The multi-specialty ARPU had been estimated at $7,500/yr — competitor pricing validation against NexHealth, IntakeQ, and Phreesia benchmarks confirmed the correct range was $12,000/yr, a 60% uplift that increased the multi-specialty SOM from $4.05M to $6.48M and made it the highest-SOM segment. The small practice count was revised upward from 130,000 to 150,000 after analysis found the prior figure implied an implausibly high average practice size of 2.54 physicians per practice for the sub-5-physician band. Census Bureau Business Dynamics Statistics cross-validation confirmed 128,948 physician office firms and 198,477 establishments — compatible with a 150,000 central estimate.
Step 5 — Final Source Reconciliation
Final validation against 25+ remaining claims. US market share confirmed at 42–45% based on North America share × US fraction of North America. The $90B DPC market reference corrected to a global projection, not a US figure. Hospital clinic averages replaced with tiered ranges by bed count — a single average concealed a 10x range in IT procurement budgets between 50-bed and 500-bed facilities. The multi-specialty NPPES count was upgraded from an AMA-derived estimate to a direct NPPES count — 28,042–32,392 confirmed addresses with 4+ physicians across 3+ specialties.
Step 6 — Stripe Data Calibration
Validated ARPU assumptions against the client’s own Stripe revenue data. Analysis of ~1,587 fixed-price active subscriptions showed a weighted blended ARPU of $700–$1,100/yr — materially below the model’s $1,500/yr assumption. The gap is explained by the composition of the installed base (skewed toward entry-tier plans) vs. new deals closed in sales calls (which close at higher tiers). Both figures were preserved in the final model with explicit notes: $700–$1,100 as the current population reality; $1,500–$2,000 as the aspirational new-deal target. This distinction changed the interpretation of ARPU growth from a background assumption to an explicit strategic lever.
