The cost of debt shows up as extra work on later changes
Aliases: interest on design debt · extra work on later changes · follow-on edit cost
What it is
A rushed gap, a main-path-only mock, a control that does not match its neighbors, looks almost free on the day it ships. Design-debt interest is not the work that was skipped. It is the detouring, reconciling, tearing-out, and redoing that every later change to that screen has to pay. The skip is a one-time credit; the interest is charged per edit.
The question is where the bill actually appears, not which ticket should hold the IOU, and not whether borrowing can be deliberate. A screen nobody will touch again accrues nothing. The same gap on a densely edited main path is expensive.
Why it happens
A shortcut on the surface becomes an obstacle for the next job. The next copy change hits a layout locked to short strings and has to be unpicked first. The next extra state finds that only success was drawn, so empty, error, and partial failure are invented from scratch against a main path already in code. The next spacing pass finds two scales mixed on the page, so one edit fans out into patches for “we never unified this.”
Each later change pays hours for the original omission; the original saving happened once. Interest compounds: the workaround leaves its own local patches, and the round after that must dodge both the old gap and the new patches. Price the debt by how many planned edits will hit the screen and which extra steps each one will grow — not by how many days the original rush saved.
Where it stops holding
A campaign shell that dies when the campaign dies has no later edits, so later-cost talk is empty. A screen already scheduled for a full replace pays the old gap once, as demolition, not as rolling interest — count the teardown, not a drip. A purely visual tweak that never touches the broken structure can sit at near-zero interest until a change finally hits that structure, at which point the bill appears all at once. A new team also pays “figure out why it was done this way”; context the original group still has orally is extra hours for everyone else.
Applying it
- When judging a shortcut, list the next two or three committed edits that will hit this screen (copy, states, platform, experiment) and write the step gap between doing it right now and carrying the gap forward.
- Record that gap as hours or blockers, not as “worse experience.” Screens with no later edits are not high-interest debt.
- Prefer the main path and high-churn screens: the same gap costs more where edits are frequent.
- How to check: pick an existing gap and mark, on the last two real change logs, which hours were spent working around or filling it. If you cannot mark extra steps, you do not yet have later-cost — only a dislike.