← Back to case studies Case Study

Building a HealthTech Data & Analytics Function

A zero-to-one data function for a HealthTech scale-up: modern data stack, self-serve analytics, automation, and board-level visibility.

Project: HealthTech data platform and analytics operating model
Tech: AWS, Snowflake, dbt, Fivetran, Omni, Python
Role: Strategy, delivery, and hands-on implementation

The Results

The work turned data from a reactive reporting service into a business operating layer used across operations, leadership, product, and customer support.

7 FTE
Manual operational effort replaced through automation
1 → 3
Data team grown while scaling platform coverage
24h → 10m
Customer-service response SLA improvement from AI-enabled workflows

The Challenge

A fast-growing HealthTech business needed trusted reporting and faster operational feedback while scaling. Data lived across operational systems, definitions varied by team, and analysts were pulled into repeat manual work instead of improving the business.

What mattered

The problem was not only technical. The platform had to earn trust with stakeholders, support board-level decisions, and give operators confidence that the numbers matched the reality of the business.

The Solution

I designed and implemented the modern data stack, established modelling patterns in dbt, introduced governed self-serve analytics in Omni, and built automation around high-friction operational workflows.

Implementation Highlights

Platform foundations

Implemented reliable ingestion, warehouse modelling, and analytics layers across AWS, Snowflake, Fivetran, dbt, and Omni.

Operating cadence

Built board-ready reporting, core business KPIs, and reusable definitions so teams were not arguing from different numbers.

Automation and AI

Delivered automation and AI-powered customer-service tooling that reduced manual work and accelerated response times.

AWS Snowflake dbt Fivetran Omni Python

Impact

The data function became a practical multiplier: fewer manual reports, clearer leadership decisions, faster operational response, and a team structure capable of scaling beyond a single analyst.

Before

Reactive
Manual reporting and inconsistent definitions
→

After

Operational
Governed metrics, automation, and self-serve analytics

Key Takeaways

Senior data work is part product, part platform

A useful data function needs infrastructure, judgement, stakeholder trust, and a bias toward removing operational friction.

AI works best when tied to a real workflow

The customer-service tooling mattered because it changed response times and workload, not because it was AI for its own sake.

Need a Data Function That Actually Changes Operations?

I help teams build the platform, metrics, automation, and operating rhythm that make data useful every day.

Get in Touch