Back

The Tech Debt Trap: When to Stop Building and Start Refactoring

2 MINS

We've all been there: deadlines are tight, the roadmap is packed, and the quick-and-dirty solution seems like the only way forward. This is how technical debt begins. Initially, it's a minor cost of doing business, but left unchecked, it can quickly turn into the Tech Debt Trap, suffocating productivity and freezing innovation.

The core dilemma for any Product Manager or Engineering Leader is: When do we stop building new features and pay the debt?

The Analogy: Building on Quicksand

Imagine your product is a skyscraper. New features are new floors, exciting and visible. Tech debt is the crumbling foundation. For a while, you can keep building taller, but eventually, the foundation starts to sink.

This manifests as:

Slower Delivery: A feature that once took days now takes weeks.
Increased Bugs: Every code change breaks three unrelated things.
Developer Burnout: Engineers spend more time debugging legacy code than writing new, valuable code.

The Three Signals to Stop and Refactor

While you can never eliminate tech debt entirely, there are three clear signals that it’s time for a Refactoring Sprint or a full Debt Paydown initiative.

This manifests as:

The 30% Maintenance Rule: If your engineering team consistently spends 30% or more of its total sprint time on urgent bug fixes, hotfixes, and unexpected stability issues, you’ve hit the danger zone.
Actionable Tip: Track and categorize all maintenance tickets.
Feature Time Triples: If the estimated time to deliver a standard, medium-sized feature has tripled due to complexity, fragility, or dependencies on old code, stop immediately.
Actionable Tip: Calculate the average story point completion rate. A sharp drop means your code base is fighting the team.
The 'We Can’t Scale' Wall: You have a major business opportunity—but your core architecture physically cannot handle the load without a major overhaul.
Actionable Tip: Treat the refactoring needed to scale as a non-negotiable roadmap feature.
Background

Jacob skipped presentations and built real AI products.

Jacob Vos was part of the Spring 2024 cohort at Curious PM, alongside 15 other talented participants.