Removing the bus-factor from a Series B fintech

It was 2:47 AM on a Tuesday when the payment reconciliation queue silently stopped processing. The founder's phone lit up with seventeen missed calls — all from the same number. Marcus, the principal engineer who had built the system three years ago, was on a flight back from his sister's wedding. No one else knew where to even start looking.

The CEO had quietly known for over a year. Every architecture diagram in the company traced back to one person. Every incident postmortem ended with 'Marcus fixed it.' Every new hire's onboarding included an unspoken disclaimer: 'when in doubt, ask Marcus.' The founder loved Marcus — and was terrified of the day he'd hand in his notice.

We didn't start by writing documentation. We started by quietly mapping who actually answered which Slack questions over the previous 90 days. The pattern was undeniable: 73% of all production-critical questions routed through one person. Once the team saw the heat map, the conversation shifted from 'we should document things' to 'we have a structural problem and here's exactly where it lives.'

• Mapped knowledge ownership across 12 critical subsystems using 90 days of Slack and PR data
• Scored architecture, documentation, and team resilience against 25 dimensions
• Co-authored runbooks and decision records for the top 5 risk areas — engineers writing alongside Marcus, not after him
• Built a 90-day cross-training plan with measurable handoff checkpoints the board could read

Within one quarter, the engineering org went from a single point of failure on payments to documented, redundant ownership across three engineers. The next investor update led with a single slide titled 'Key-person risk: resolved' — and was cleared as a green flag with no follow-up questions.