Case Study
Nine months stalled, ninety days to production
U.S. healthcare provider — risk-transfer financial platform · Senior Project Manager · Jan-Apr 2026
At a glance
Role
Senior Project Manager (Level IV)
Timeline
Estimated at 5 months, 9 months in at handover, shipped ~3 months later
Team
9+ core, ~15 across DevOps, Finance, QA
Domain
Financial calculation and reporting, healthcare
The situation
I inherited this project from my supervisor on her last day before maternity leave. Her assessment: it was essentially done, a few bug fixes short of a production go-live that had already slipped from a 5-month estimate to 9 months.
The challenge
Within the first week it was clear the handover assessment didn't hold. The team could execute the financial formulas but couldn't explain why they worked the way they did, and none of it was documented. The client kept asking exactly that question, and nobody could answer. Then, partway in, I found an entire missing module. Not a defect — a scope gap that had gone unnoticed for nine months. And the release itself had to replace an already-live system, using a restored database that had accumulated manual changes needing reconciliation.
What I did
Month 1 — reconstruct the logic.
I went to every developer who had touched the financial code and reverse-engineered the calculation logic from scratch, then documented it. This wasn't a deliverable anyone had asked for. But you cannot get a client to sign off on financial calculations they don't understand, and you cannot maintain a system whose reasoning lives in three people's heads.
Month 2 — close the scope gap.
Built the missing module plus the additional features that surfaced once the documentation exposed what was actually incomplete. Re-baselined the plan rather than pretending the original one still applied.
Month 3 — earn the go-live.
Ran an intensive UAT and regression cycle with engineering and QA. Because the logic was now documented, client questions had answers, and the go-live approval came on evidence rather than on trust.
Release weekend.
Orchestrated the cutover including reconciliation of the manually-modified restored database against the new system.
The outcome
- A project stalled at 9 months delivered to production within ~3 months of takeover.
- Financial logic documented from zero — the artifact outlived the project.
- Client greenlit go-live after UAT, on the strength of being able to interrogate the calculations.
- Delivered via the last weekend of April 2026, replacing a live system.
What I took from it
The instinct on a late project is to go faster. Spending the first month writing documentation on a project already four months overdue felt wrong and was the reason it shipped. The delay was never a velocity problem — it was that nobody could explain the product, so nobody could approve it.