How do you plan tech debt and feature delivery together?

A strong answer treats technical debt as a product-delivery input, not a competing backlog owned only by engineering. Make it concrete: estimate incident risk, security exposure, developer time lost, blocked features, and customer impact. Classify urgent reliability or security items separately from ordinary cleanup, then reserve a predictable portion of capacity for the latter while allowing threshold-based interruption for the former. For each debt item, state the smallest safe investment, expected outcome, owner, and a metric such as deploy lead time, error rate, or support volume. Explain trade-offs plainly with product partners and revisit the allocation each planning cycle. Avoid a vague “20 percent always” rule if evidence shows a different risk profile. A good example includes paying down a bottleneck before a feature depended on it, then measuring the avoided delay or reliability gain.