Exact Sciences
Launching a Cancer Diagnostic Platform in Europe
Data integration strategy for a greenfield lab expansion across three enterprise systems
Role
Manager, Technical Business Analysis
Team
Led 5 business analysts
$140M
Projected revenue tied to European expansion
150+
Fields mapped across systems
3
Enterprise systems integrated end-to-end
5
Business analysts led across workstreams
Background
Exact Sciences was expanding its flagship cancer diagnostic test into Europe by opening a new lab in Germany — the company's first international lab operation. The lab runs on Starlims, a laboratory information management system. Customer and order data lives in Salesforce. At the end of the process, patients receive a PDF report delivered through Salesforce. MuleSoft is the integration layer connecting everything.
None of these three systems were connected at the start of the project. The work was greenfield: define what data needs to flow between them, map every field, resolve every constraint, and get all three systems talking end-to-end before the lab went live.
My Role
I led a team of five business analysts responsible for defining and documenting data requirements across all three systems. My team owned the source-to-target mapping document — the single artifact that every integration decision flowed through.
Beyond the technical mapping work, I drove stakeholder alignment across three distinct teams with different vocabularies and priorities: the lab team working in Starlims, the CRM and patient experience team in Salesforce, and the engineering team building the MuleSoft integration layer. I also worked directly with data engineering on technical constraints, and coordinated with an internal lab software team who had institutional knowledge of Starlims.
Approach
The core artifact was a source-to-target mapping document covering 150+ fields. This document was the single source of truth for every integration decision — if it wasn't in the mapping doc, it wasn't getting built.
The key challenge was operating at the intersection of technical constraints and business needs. When the lab team requested data in a format MuleSoft couldn't support, when Salesforce had storage constraints on a field, or when GDPR prohibited certain data from being transmitted across borders, the answer wasn't just "no." I worked with each team to understand the underlying need and find an alternative approach that satisfied both the business requirement and the constraint.
GDPR was a consistent thread throughout the mapping work. The Germany lab meant patient data was crossing regulatory environments, and every field decision had to be weighed against what was legally permissible to send, store, and retain — not just what was technically convenient.
The internal Starlims team was a critical partner. They understood the lab system's data model and its limits in ways that weren't documented anywhere, and pulling that knowledge into the mapping process early prevented late-stage surprises.
Outcome
The lab launched successfully in Germany. All three systems — Starlims, MuleSoft, and Salesforce — were integrated end-to-end, and patient reports were delivering correctly through Salesforce at launch.
I traveled to Germany for the hypercare period after go-live to support the team on the ground, troubleshoot issues in real time, and ensure the integration held up under live conditions. The European expansion tied to this work carried $140M in projected revenue.
Reflection
The source-to-target mapping document was the most important artifact on this project. Not because data mapping is glamorous work — it isn't — but because it forced every team to be explicit about what they needed and gave everyone a shared reference point when disagreements came up. When someone said "that's not what we agreed," the answer was always in the doc.
The other thing that mattered most was being able to translate between technical constraints and business terms. Telling a stakeholder "MuleSoft can't do that" doesn't help anyone. Saying "MuleSoft can't transform the data in that format at this step, but here's what it can do, and here's how we can restructure the requirement to get you the same outcome" — that's the work.