Skip to content
    Outline · Influence

    Updated: 12 August 2026 at 00:00

    The Sunk Cost Fallacy

    mid

    Why teams continue unprofitable initiatives because of past investment, how to recognize the trap, and how to decide based on future value.

    Source

    Sunk cost fallacy

    book_cube note about sunk-cost decisions and adjacent cognitive effects in leadership contexts.

    Open source

    Thinking / Critical Thinking / Leadership

    Sunk cost fallacy appears when a team keeps a strategy alive because of already spent time, budget, or reputation, even when new evidence suggests a better path is to stop or reframe.

    1. What matters for tech leads

    • Past cost is unrecoverable and should not drive the next decision.
    • Evaluate options by expected forward value, risk, and opportunity cost.
    • In engineering work this bias often hides behind phrases like "we are already too far in to stop now".

    2. Adjacent cognitive effects

    IKEA effect

    People overvalue outcomes they assembled or customized themselves, even when objective value is lower than available alternatives.

    NIH syndrome (Not Invented Here)

    Teams avoid external solutions just because they were not built in-house, then keep spending on a weaker internal replacement.

    3. Decision-debug questions

    Question 1

    If we had not already invested so much effort, would we still push to continue this initiative today?

    Question 2

    If X had not already been done, would we choose to do it again under current constraints?

    Question 3

    What alternatives do we have, including active changes and an explicit status quo option?

    A practical pattern: force 2-3 competing alternatives before discussing whether to continue the current plan. This reduces tunnel vision in high-commitment initiatives.

    4. Common anti-patterns

    The decision is defended by past spend, not by expected future value.

    The team avoids stop-loss discussions because stopping is treated as reputational failure.

    The argument "we already did too much" replaces a comparative option analysis.

    The choice is framed as "continue or fail" while viable reset/scoping options exist.

    5. Recommendations

    Separate sunk cost from forward-looking value: sunk cost is not a valid prioritization argument.

    Run pre-mortem before go/no-go decisions and define conditions that must trigger a stop.

    Treat status quo as a real alternative, not as invisible default behavior.

    Make the three decision-debug questions mandatory in initiative reviews.

    6. Sources and follow-up

    Progress tracking is off. Turn it on in settings.

    Learning evidence

    Reading is only the start. Move the idea into a real workplace experiment and reflection.