All articles

Technical Debt Management: A Practical Plan for Business Owners

Technical debt management for non-technical owners: what it is, how to rank fixes, how much time to set aside, and how to build a simple roadmap.

Technical debt is the future cost of shortcuts taken while building software. Managing it means making that cost visible, ranking it by business impact, and paying it down on a fixed schedule instead of in a panic. If your site, app or internal tools feel slower to change every year, this is why. We build and maintain custom websites, apps and automation at Westcode, and the plan below is the one we would hand a friend who owns the product but does not write the code.

TL;DR

  • Technical debt is any shortcut in your software that makes the next change costlier. It is normal. Unmanaged debt is the problem.
  • Start with a written list of debt items, each with a cost to fix and a cost of leaving it.
  • Rank by business impact: how often the code changes, how risky it is, and how much effort the fix takes.
  • Set aside a steady share of every cycle, commonly 10 to 20 percent, instead of occasional cleanup sprints.
  • Review the list every quarter and keep a short roadmap.

What technical debt looks like in a business

The Software Engineering Institute at Carnegie Mellon defines technical debt as a design or implementation choice that is expedient now but can make a future change costlier or impossible. It also counts delayed upgrades and technology refreshes as debt.

In plain terms, you will notice it like this:

  • A small change takes weeks, and nobody can say why.
  • Only one developer understands how billing, bookings or listings work.
  • Upgrades keep getting postponed.
  • Every release breaks something that used to work.

Common types, as IBM lists them, are architectural debt (tightly coupled parts), code debt (duplicated logic), documentation debt, outdated build and deployment tooling, process debt and security debt, such as skipped patching.

Some debt is a smart choice. If you ship a first version quickly to test the market, you are borrowing on purpose. Martin Fowler warns that the borrowing only pays off briefly: messy code slows new features within weeks, not months. So borrow knowingly, and write down what you borrowed.

Step 1: Make the debt visible

Debt that lives in one developer's head cannot be planned. The SEI recommends building an inventory of debt items and recording the consequences of letting each one grow.

Keep each entry short:

  1. What it is, in one plain sentence.
  2. Which part of the business it touches (checkout, tenant portal, dealer orders).
  3. Cost to fix, as a rough estimate in developer days.
  4. Cost of leaving it: lost hours, outage risk, security exposure.
  5. Who owns the decision.

Put it in the same tracker as your feature work. Atlassian advises against hiding debt in a separate backlog, because it then never competes for attention.

Step 2: Prioritize by business impact

You cannot fix everything, and you should not try. The Software Improvement Group says prioritization should weigh technical severity against business consequences: business criticality, how often the code changes, risk exposure, remediation effort and test coverage. Debt in code that changes often and has weak tests is riskier than the same flaw in well-tested code.

A simple scoring pass works for most small teams:

Question High score means
Does it touch revenue or customers? Fix sooner
Does the code change every month? Fix sooner
Is it a security or data risk? Fix first
Is the fix small compared with the pain? Quick win
Is the code stable and rarely touched? Leave it

A guide on technical debt measurement advises ranking by carrying cost combined with how often the code is touched, not by age or by whoever argues loudest.

Example: a property manager's tenant portal has an old payment form that breaks twice a year, and a messy reporting script nobody opens. Fix the payment form. Leave the script.

Step 3: Decide how much debt is acceptable

There is no single right amount. Steve McConnell's workshop slides point to warning signs: a high interest rate, a high minimum payment, debt nobody tracks, debt taken on by accident and low return. He also asks who in the business owns each debt.

A useful rule for owners: take on debt only when you can name the date and the plan to repay it. Launching a first app with a thin feature set before a trade show is a good reason. "We were busy" is not.

Step 4: Reserve capacity, every cycle

Cleanup sprints sound good and rarely last. SonarSource recommends treating debt like any other backlog work, with a common practice of giving it a consistent 10 to 20 percent of sprint capacity. The same share, decided in advance rather than negotiated each sprint, is what keeps it from being squeezed out.

Fowler's advice is to pay gradually: when you touch a messy area for a new feature, spend a little extra time cleaning it. Budget for this in the plan. Our guides on maintenance budgets for custom software and app maintenance costs show how much to plan for each year.

Step 5: Build a one-page roadmap

A debt roadmap does not need to be fancy. The SEI suggests three stages: first make debt visible by adding a debt category to your tracker, then set goals and measure the share of effort spent on it, then add tooling where debt actually builds up.

For a business, that becomes:

  • Now (this quarter): security and outage risks, plus quick wins.
  • Next (two quarters): debt that slows your most-changed features.
  • Later: large rewrites, such as replacing an old platform. Schedule them against a business event, like a rebrand or a CMS move. Our CMS migration plan shows how to run one with owners and a launch checklist.
  • Never: stable code that works and rarely changes.

A research study of teams doing this found helpful habits: naming a debt champion, using a standard set of fields for each item, and setting a resubmission date so a deferred item comes back for review.

Where AI-built code fits

AI coding tools speed up the first draft. Review matters just as much afterward. SonarSource reports that Carnegie Mellon researchers found early speed gains faded by month three, alongside more static analysis warnings. Treat generated code like any other code: test it, read it, and log what you defer.

Our view

Most owners we meet do not have a code problem. They have a visibility problem. Nobody told them what was borrowed, so the bill arrives as a surprise.

We build custom websites, apps and automation with the client involved from idea to launch. That transparency is how debt stays honest: you see the tradeoffs when we make them, and the list of deferred items is yours to keep.

Our advice is simple. Ship the first version fast if the market needs it. Write down every shortcut. Book a fixed slice of each cycle to pay it back.

 

Loic

Get a debt plan with the team that builds it

If you run a startup, a property management company or a home service business and your software feels harder to change each quarter, we can help you sort the list and plan the repairs alongside new features. We design and build websites, apps and AI and workflow automation from Vancouver and Kelowna. See our services, or if you are in the Okanagan, our Kelowna app development page.

FAQ

What is technical debt management?

It is the routine of finding, recording, ranking and repaying shortcuts in your software. The goal is not zero debt. It is debt you chose, can see and have a plan to repay.

How much technical debt is acceptable?

There is no fixed number. McConnell's guidance is to judge each item by its interest, its monthly cost, whether it is tracked, and the return on fixing it. Many teams reserve 10 to 20 percent of capacity to keep it in check.

What is an example of technical debt?

A billing system that only one person understands. Asana describes a similar case: a team under deadline pressure ships without enough senior review, and bugs surface after launch.

How do I prioritize technical debt?

Rank by business impact, how often the code changes, risk and effort. Fix security and revenue risks first, then quick wins. Leave stable, rarely touched code alone.

Sources

Written by westcode.io, an independent studio in Kelowna and Vancouver.

AI & automation8 min read

Answer Engine Optimization: A Practical Guide for Business Owners

Answer engine optimization explained for business owners: how AEO differs from SEO, the steps that earn AI citations, and how to measure results in Search Console.

AI & automation8 min read

Single Sign-On Setup: A Plain Guide for Business Owners

How to set up single sign-on: what to gather first, SAML or OIDC, step-by-step configuration, testing, lockout protection and when to build SSO into your app.

AI & automation8 min read

IoT Device Management: What It Is and How to Plan It

IoT device management explained for business owners: the five lifecycle stages, what a platform does, how to manage a fleet, and how to plan yours before you build.