"Deconstructing the Sunk Cost Fallacy, Sunk Architecture Syndrome, and Not-Invented-Here syndrome when designing large-scale software systems."
Introduction
Software engineers pride themselves on purely logical decision-making. Yet architectural choices—from selecting databases to choosing framework rewrites—are often deeply clouded by unconscious cognitive biases.
The Sunk Cost Fallacy in Legacy Codebases
Teams frequently pour millions of dollars into patching aging monolithic systems simply because of the historical time and capital invested. Recognizing when an architecture has hit an efficiency dead-end is a critical leadership skill.
“The hardest thing in software engineering is not writing complex code; it is having the intellectual honesty to delete what is no longer serving the mission.”
Combating Not-Invented-Here (NIH) Syndrome with Standardized RFCs
Standardized Architecture Decision Records (ADRs) and pre-mortem analysis meetings ensure that teams evaluate existing open-source solutions objectively before committing to building custom internal tools.
Key Takeaways
• Sunk cost bias leads teams to maintain obsolete architectures past their economic viability.
• Not-Invented-Here syndrome causes wasteful internal duplication of battle-tested open source tooling.
• Pre-mortem exercises identify systemic failure modes before code is written.
