Skip to content
← Back to work
Case Study / GTM Healthcare Intelligence

GTM Healthcare Intelligence

Specialty-EHR go-to-market analytics architecture organized across four maturity layers — governance, descriptive, diagnostic, prescriptive — so Sales, Marketing, and Finance can share a source of truth without sharing a surface.

7 sub-projects · 4 maturity layersCMS NPPES · Medicare PUF · CensusL2 logistic regressionReact · Next.js · Python · Claude APIOpen ↗
Translation artifact
2024Solo build
01Problem
Situation: Specialty-EHR vendors run go-to-market with Sales, Marketing, and Finance each maintaining their own pipeline definitions, conversion math, and revenue forecasts — built around the cadences and decision rituals each function actually trusts.
Complication: When the three functions disagree on inputs, downstream forecasts and territory decisions become negotiations rather than analyses. Centralized BI rollouts under-perform precisely because they ask each function to abandon the surface it knows; standardization stalls when no team will adopt a definition it didn't author.
Question: Can a single analytics layer serve all three functions without flattening the differences in how each one actually decides?
02Requirements
  • Sales leadership

    Funnel diagnostics by specialty and metro — what's dropping off and where, named — surfaced inside the views the team already uses for pipeline review.

    ConvertPath used in mock pre-QBR sessions

  • Marketing

    Account scoring grounded in fit + intent signals from public data, with conversion math that survives quarterly review.

    Anchored to MGMA/HFMA benchmarks and CMS NPPES

  • Finance

    A metric registry with conflict resolution and anomaly detection so close-the-books forecasts trace back to a defined source.

  • Cross-function

    A natural-language layer where any function can ask 'why are we losing X' and get a structured answer with cited sources — not a black-box recommendation.

    AskGTM returns sourced reasoning, not bare numbers

03Decision

Per-function surfaces over a shared metric registry

chosen
  • meets criterion: Multi-stakeholder fit
  • meets criterion: Adoption friction
  • meets criterion: Source-of-truth integrity
  • partially meets criterion: Build effort

Function-specific surfaces match the rituals each team actually uses — Sales lives in funnel views, Finance in close-the-books cycles, Marketing in scoring loops. The registry keeps definitions reconcilable without forcing a uniform interface. The 4× build cost (four maturity layers, multiple consumer surfaces) is justified by the documented failure rate of single-dashboard rollouts in cross-functional GTM contexts.

Centralized BI dashboard (one surface for all three)

  • partially meets criterion: Multi-stakeholder fit
  • does not meet criterion: Adoption friction
  • meets criterion: Source-of-truth integrity
  • meets criterion: Build effort

Embedded analytics in the CRM (Salesforce-native only)

  • does not meet criterion: Multi-stakeholder fit
  • partially meets criterion: Adoption friction
  • does not meet criterion: Source-of-truth integrity
  • meets criterion: Build effort
04Solution

A four-layer analytics architecture organized by decision maturity — each function enters at the layer that matches its rituals and crosses into others when needed.

Governance
NorthStar — metric registry with conflict resolution, reporting adoption tracker, shadow-spreadsheet monitor, and anomaly detection. The single source of truth for what "pipeline" means.
Descriptive
PracticeFlow + SpecialtyPulse — benchmarking against MGMA/HFMA standards and trend monitoring at practice and specialty level, anchored on CMS NPPES, Medicare PUF, and Census demographics.
Diagnostic
ConvertPath + SpectrumIQ — funnel-decay analysis showing where deals drop off by specialty, plus metro-level opportunity scoring via L2 logistic regression.
Prescriptive
AskGTM + AskPractice — natural-language query layer over the same governed data; "why are we losing Ortho deals?" returns a structured, sourced answer rather than a hidden inference.
05Outcome
  • Architecture

    4 maturity layers

    Governance · Descriptive · Diagnostic · Prescriptive

  • Sub-projects

    7 mapped

    Each tied to a specific decision moment

  • Public data anchors

    5 sources

    CMS NPPES · Medicare PUF · Census · MGMA · HFMA

  • Build status

    End-to-end demo

    Synthetic + public data only · no proprietary or patient data