Strategy
4 min read

Escaping the Sunk Cost Trap: Killing Features with Conviction

The hardest part of product leadership is not choosing what to build, but having the courage to abandon what you already built.

Sarah Chen·Product Strategy Director·Aug 10, 2026
Core Takeaways
  • Treat past engineering hours as zero-cost in current decision matrices.
  • Apply the "Clean Slate" test to contentious roadmap items.
  • Celebrate feature deprecations as victories for codebase velocity.

Every software team has experienced the sunk cost fallacy: spending engineering weeks fixing a legacy feature that users do not use simply because "we already invested so much effort into it."

Resources spent yesterday are gone forever. Today’s decision must be evaluated solely on future expected value divided by future opportunity cost.

Ask yourself: "If this codebase and feature did not exist today, would we choose to invest our next sprint building it from scratch?" If the honest answer is no, the feature must be sunset.

Apply this model to your current problem

Run your decision through ThinkRoom's 4-perspective engine now.

Launch workspace