How much of your roadmap goes to technical debt, and how do you justify it to the business?

Reserve a standing share of capacity, commonly fifteen to twenty percent, so it never has to win an argument against a feature. Justify it in business terms: cycle time, incident rate, on call load, and the cost of the last outage. Debt framed as engineering hygiene loses every time; framed as the reason the next three features ship in weeks instead of quarters, it holds.

I do not let it compete. If debt has to win an argument against a feature every sprint, it loses every sprint, so we hold a standing share of capacity, around twenty percent, that engineering directs. Then the job is making the value legible. I stopped using the phrase technical debt in front of the business entirely, because it sounds like housekeeping. Instead I report cycle time and incident count. When our deploy time went from forty minutes to eight after a build overhaul, that was a very easy story to tell, because everyone in the company had felt the forty minutes. I also ask the engineering lead to keep a short ranked list of the debt that is actually hurting rather than a wish list, so when we do have a bigger window we spend it on the item that unblocks the roadmap instead of the one that annoys people most.