You need budget for a year of technical debt work. Leadership has repeatedly declined. How do you make the case?

What the interviewer is testing

Whether you can translate an engineering concern into business terms, and whether you recognise why the previous attempts failed.

Why the previous attempts failed

"Technical debt" is a term that has lost force with non-technical audiences. It sounds like engineering wanting to tidy up, competing against revenue features, and losing — correctly, from the business's point of view, because no consequence was quantified.

The case that works

Express the interest being paid, not the principal. Not "the code is bad" but:

• How much slower each change is now. "A change that took three days two years ago takes eleven." Measure it — lead time by component is available.
• How much of the incident load it causes. "Sixty per cent of production incidents in the last year originated in this component. Average incident cost is X."
• How much capacity it consumes. "Two engineers' full time is spent on maintenance that would not exist."
• What it prevents. "We cannot deliver the feature you asked for last quarter without this work first, and here is why."

That last one is usually the most effective, because it connects to something leadership already wants.

The framing

Attach it to a business objective rather than requesting it as a separate initiative. If the roadmap includes expanding to a new market, the platform work that expansion requires is part of that programme's cost, not a competing request.

Propose a proportion rather than a project. A standing allocation — say 20% of capacity — is easier to approve than a year-long programme, does not require a large single decision, and is far less likely to be cut when priorities shift.

Show the trend. Lead time and change failure rate degrading over quarters is more persuasive than a snapshot, because it implies a trajectory.

What a strong answer adds

Being selective. "A year of debt work" is unlikely to be approved and probably should not be — most debt is not worth repaying. Identify the specific components where the interest is highest, quantify those, and ask for that. A focused request with numbers behind it succeeds where a general one does not.

Common weak answers

Arguing from code quality. Requesting a freeze on features.