The Rewrite Viability Audit
by Camille Fournier • Author & Tech Executive at Various (formerly Two Sigma, Rent the Runway, Goldman Sachs)
Camille is one of the most respected technology executives in the industry and the author of 'The Manager's Path', the definitive guide for engineering management. She is a former CTO of Rent The Runway and has held senior leadership roles at Two Sigma, JP Morgan Chase, and Goldman Sachs.
🎙️ Episode Context
Camille Fournier dissects the friction points between Product Managers and Engineers, offering tactical advice on how to build trust, avoid 'hoarding' credit, and prevent over-engineering. She also provides a contrarian perspective on software rewrites, arguing they are often traps due to underestimated migration costs, and lays out a blueprint for modern Platform Engineering teams treating internal infrastructure as a product.
Problem It Solves
Prevents teams from falling into the 'Rewrite Trap,' where years are wasted rebuilding systems with negative ROI and stalled feature development.
Framework Overview
A decision-making heuristic to evaluate if a system rewrite is necessary. It shifts the focus from 'how hard is it to build the new thing' to 'how hard is it to migrate the old thing.'
🧠 Framework Structure
The Migration Multiplier: Recognize t...
The 'Do Nothing' Test: Ask, 'If we di...
The Hidden Logic Assessment: Acknowle...
Evolution over Revolution: Instead of...
When to Use
When an engineering team proposes a 'ground-up' rewrite of a legacy system to move faster or use modern tech.
Common Mistakes
Assuming the new system will automatically be faster, or forgetting that you must support the old system simultaneously during the multi-year build.
Real World Example
Camille discusses how teams often convince themselves that a new framework will solve support woes, only to realize the migration halts all feature work for 6-24 months.
Engineers notoriously, notoriously, notoriously, massively underestimate the migration time for old system to new system.
— Camille Fournier