
How I Decide Which Technical Debt to Fix
Every engineering team keeps a list of code it hates: the deployment script nobody dares touch, the service with the four-thousand-line class, the test suite that fails on Tuesdays. Dislike is a signal worth hearing, but it is not a measurement, and if I fixed everything my teams complained about we would spend years rewriting working systems while dangerous debt waited in the queue. In this guide, I share the five factors I score – operational risk, security exposure, delivery friction, maintenance cost, and probability of change – how I sort every item into fix now, defer with a trigger, or accept in writing, and why the debt a team hates most is sometimes the debt worth keeping.






