Kefiw

Archived noindex page. Kefiw's public focus is Property decision help.

Archived page

This older Kefiw page is kept for reference, marked noindex, and removed from the primary sitemap. The current Kefiw experience is focused on property decisions: cost, quotes, damage, buying, selling, owning, and packets.

Go to Property

When to Run Tech Debt Interest

Five engineering moments when the compound cost number actually changes scheduling.

Tech debt is invisible until you make it visible. These five moments mean run the math and attach it to the ticket.

Engineering teams defer tech debt because it has no deadline. Attaching a compounded cost number gives it one. Run the math at these five moments and debt repayment stops losing to feature work by default.

Quick answer

Tech debt is invisible until you make it visible. These five moments mean run the math and attach it to the ticket.

What you are trying to do
Five engineering moments when the compound cost number actually changes scheduling.
Best next step
Tech Debt Interest
Limit to remember
Treat this as a practical aid for the task, not a replacement for professional judgment.

Key points

  • Before accepting any hack or shortcut: estimate the clean-fix hours and the postponed payoff. The gap is the "interest rate" you're agreeing to.
  • When estimating a refactor: compounded cost at current month tells you what the fix is already worth. Compare to the cost at wait+6 to argue for scheduling.
  • Quarterly planning: run on the top 5 known debts. Sort by payoff-at-end-of-quarter. Prioritize the ones compounding fastest.
  • After a major feature lands on legacy code: velocity just spiked. Debt in that area is compounding faster — rerun and reschedule.
  • Post-incident review: if tech debt contributed to an incident, quantify it. The payoff hours plus the incident cost tell the real story.
  • Before handoff to a new team or contractor: unfixed debt compounds faster under unfamiliar hands. Either fix before handoff or explicitly budget for it.

Examples

  • Accept-hack trigger
    Proposal: 2hr hack vs 8hr clean fix. 12 months later at 8%/mo: hack is now 20hr payoff. Net time saved by hack: 2 - 20 = -18 hours. Unless velocity is low, pay the 8 now.
  • Quarterly-planning trigger
    Top 5 debts compounded at end-of-quarter: 8, 22, 45, 60, 140 hours. The 140-hour one is a rewrite risk. Scheduling that one this quarter prevents it from becoming a 300-hour project next year.
  • Post-incident trigger
    Outage traced to known-bad auth module. Payoff hours modeled at 30 six months ago — now 60 plus $X incident impact. Next sprint blocks off the refactor.

When to use which tool

Related

Frequently asked questions

How do I get product to prioritize debt? How-to

Turn debt into dollars or customer impact. Payoff hours × loaded engineer rate + expected incident cost = a number product leadership can weigh against feature work.

Should every hack get this treatment? Trust & accuracy

No — only the hacks that touch actively-developed code or customer-facing paths. One-off scripts and dead code don't compound meaningfully.

How should I use a decision framework in real life? How-to

Use a decision framework to expose the tradeoff, not to outsource the decision. Write down the inputs, compare the output with your constraints, then ask what would change the answer. The strongest use is scenario testing: base case, conservative case, and failure case.

Is this financial, legal, or tax advice? Trust & accuracy

No, this is not legal, financial, tax, medical, or professional advice unless the page explicitly says that use case is supported. It organizes assumptions so you can inspect them. Verify high-stakes choices with qualified people who can review facts, contracts, regulations, and downside risk.

What assumption matters most in a decision model? Edge case

The most important assumption is usually the one you are least certain about and most emotionally attached to. Change that input first. If the recommendation flips after a small change, the decision is fragile and needs more evidence before you treat the model as useful.