Technical debt explained: the interest hiding in your estimates
Ward Cunningham's metaphor via Martin Fowler: cruft is principal, slower features are interest. Prudent vs reckless debt, and what buyers should watch.
"We need to pay down technical debt" is something your development team will eventually say - usually when estimates start slipping for reasons that are hard to see. The metaphor is worth understanding properly, because it is a financial metaphor: debt is not automatically bad, but the interest compounds, and nobody else can pay it for you.
This guide follows Martin Fowler's explanation of Ward Cunningham's metaphor, checked on 19 November 2026, and translates it into what a founder or buyer should watch for.
The metaphor, properly stated
Ward Cunningham coined the debt metaphor, and Martin Fowler's formulation is the standard reference: codebases accumulate "cruft" - shortcuts, duplication, outdated designs - and technical debt frames that cruft like a financial debt. The extra effort it takes to add new features is the interest paid on the debt. Fowler's example arithmetic: if a feature would take two days on a clean codebase but five on yours, the three-day difference is interest you are paying. Principal is the work of cleaning the code up.
Deliberate vs inadvertent debt
Fowler separates debt along two axes - reckless vs prudent, and deliberate vs inadvertent. The quadrant that matters for buyers: prudent, deliberate debt is a conscious trade ("we know this is rough, we ship now to learn, we schedule the cleanup") and is a legitimate business tool. Reckless, inadvertent debt is what accumulates when nobody is tracking it - and it is indistinguishable from the outside until velocity collapses. The question to ask a team is not "do we have debt?" (every codebase does) but "where is it, and what is our interest rate?"
What debt looks like from the buyer's chair
- Estimates drift upward for similar work. The third "simple form" takes longer than the first. That is interest showing up in your invoices.
- Bug fixes spawn bugs. Changes in one place break unrelated places - a sign the code's structure no longer matches how anyone thinks about it.
- "It would be faster to rewrite." Sometimes true, usually a trap: a rewrite inherits the same pressures and discards embedded knowledge. Treat rewrite proposals as a last resort with a phased plan.
- Maintenance quotes grow. Debt is a driver of the recurring cost covered in what app maintenance costs per year.
Keeping it rational
You cannot forbid debt; shipping requires trade-offs. You can make it visible: ask teams to log deliberate shortcuts when they take them, to reserve a regular slice of capacity for cleanup (a standing percentage works better than "a big refactor later"), and to report interest symptoms - slowing estimates, rising defect rates - in plain language. In contracts, debt is one reason fixed-price vs time-and-materials matters: a hard fixed price can quietly incentivize the reckless kind.
Sources
- Martin Fowler, Technical Debt (bliki): https://www.martinfowler.com/bliki/TechnicalDebt.html
Drafted with AI assistance from this source. Debt trade-offs are context-specific; the metaphor guides judgment, it does not replace an engineer's read of your actual codebase.