Technical Debt Is a Business Decision
Engineers didn't create most of your technical debt. Deadlines did. Treating it as an engineering problem keeps it unfixable.
Debt has an author
Behind almost every shortcut in a codebase is a date someone chose: the launch that couldn't slip, the demo for the board, the quarter that needed saving. Those were often correct decisions — speed genuinely bought something. The dysfunction begins afterwards, when the borrowing is forgotten but the interest keeps compounding, and engineers are left explaining 'slowness' that was purchased deliberately by people no longer in the room.
The borrowing is forgotten but the interest keeps compounding.
Put it on the ledger
The fix is bookkeeping, not heroics. When a shortcut is taken, record what was borrowed, what it bought, and what repayment looks like — in business language, one paragraph. Review the ledger quarterly beside the roadmap. Debt that blocks where the company is going gets paid; debt in code that's stable and rarely touched gets explicitly tolerated. Both are fine. Unexamined is the only wrong state.
Framed this way, the conversation stops being engineers asking permission to clean up and becomes leadership managing a portfolio of borrowings it chose. That's what it always was — the ledger just makes it honest.
If this sounds like a conversation your team is having, we should have it together.
