Your team's monolithic authentication service has grown to 80k lines of code over 5 years. It takes 4 hours to deploy, has no test coverage, and has caused 3 incidents in 6 months — each taking 4+ hours to resolve. Your PM wants to ship 3 new auth features this quarter. How do you frame the refactor vs. feature build decision?
Quantify debt in dollars and velocity loss; avoid purity arguments. Propose incremental strangler work tied to features the PM cares about, with measurable exit criteria like deploy time and incident rate.
Ready to practice?
Write your structured answer, then compare to a strong model answer.