Sunk Cost Fallacy

Sunk Cost Fallacy

Overview

The Sunk Cost Fallacy describes the systematic tendency to continue investing in a project, relationship, or course of action because of resources already committed, even when discontinuing would be more rational. Formalized by psychologists Hal Arkes and Catherine Blumer in their seminal 1985 paper "The Psychology of Sunk Cost," this bias represents a fundamental violation of rational decision-making: past costs should be irrelevant to future choices, yet they dominate our behavior.

The classic example from Arkes and Blumer: You've bought a $100 ski trip to Michigan and a $50 ski trip to Wisconsin for the same weekend. You expect to enjoy Wisconsin more. Which do you choose? Rational analysis: The money is gone either way (sunk cost). Choose based on which trip you'll enjoy more (Wisconsin). Human behavior: Most people choose Michigan because "I spent more on it" or "I don't want to waste the $100."

This manifests across domains: companies throwing good money after bad into failing projects, individuals staying in unfulfilling careers because of years invested, governments escalating military commitments to justify past casualties, product teams continuing to build features no one wants because they've already spent months on development.

The mechanism is psychological, not logical: sunk costs create emotional investment (loss aversion, desire not to "waste" resources, need to justify past decisions, ego protection). Admitting a sunk cost was a mistake feels like personal failure. Continuing feels like honoring the investment. But this confuses accounting (tracking what's been spent) with decision-making (choosing optimal future actions).

Related to escalation of commitment, status quo bias, and loss aversion, the Sunk Cost Fallacy is particularly insidious because it masquerades as virtues: persistence, follow-through, not being a quitter. Culture often reinforces it: "Winners never quit" becomes "Keep investing in losing strategies."

Key insight: Sunk costs are sunk. The only decision that matters is: given where we are now, what action maximizes future value?

When to Use

Apply Sunk Cost Fallacy awareness in these situations:

• Project decisions: Continuing development on products, features, or initiatives with poor prospects
• Strategic pivots: Evaluating whether to persist with current strategy or change direction
• Hiring/firing: Keeping underperforming employees because of time invested in hiring/training
• Relationships: Staying in romantic or business partnerships due to years invested rather than current value
• Financial decisions: Holding losing investments, continuing failing businesses, finishing expensive but unrewarding degrees
• Resource allocation: Allocating budget to existing initiatives over new opportunities
• Product development: Shipping features because they're 80% done, not because they add value

Trigger question: "If I were making this decision from scratch today, with no history, would I choose this path?"

Process

1. Use the "Fresh Start" Test

Reframe decisions as if starting from zero, ignoring past investment:

• "If I didn't already own this stock, would I buy it today at this price?" (If no, sell)
• "If I were hiring for this role today, would I hire this employee?" (If no, start transition)
• "If I were starting this project today, would I commit these resources?" (If no, cancel)
• "If I just met this person, would I enter this relationship?" (If no, reconsider)

Action: For any continuation decision, write out the "fresh start" question and answer it without referencing past investment.

2. Separate Accounting from Decision-Making

Explicitly acknowledge sunk costs, then exclude them from analysis:

• Accounting: "We've spent $2M and 18 months on this product."
• Decision-making: "Going forward, will this product generate positive ROI? If yes, continue. If no, cut."
• Create two columns: "Sunk Costs (Irrelevant)" and "Future Costs/Benefits (Relevant)"
• Make decisions based only on the right column

Action: Build a decision template with sunk costs listed and visibly crossed out, then analyze only forward-looking costs and benefits.

3. Implement Kill Criteria Upfront

Before starting initiatives, define objective conditions for cancellation:

• Set "trip wires": measurable metrics that trigger mandatory review or shutdown
• Example: "If we don't have 1,000 users by Month 6, we cancel the product"
• Make criteria specific, measurable, and committal—no room for reinterpretation
• Revisit kill criteria when hit, not when emotionally attached months later

Action: For any new project, write down 3-5 kill criteria before starting. Commit to honoring them regardless of sunk costs.

4. Conduct Quarterly Portfolio Reviews

Regularly evaluate all ongoing commitments as if allocating resources fresh:

• List all projects, initiatives, relationships, investments currently consuming resources
• For each, ask: "If this didn't exist, would we start it today with what we know now?"
• Rank by future expected value, ignoring past investment
• Reallocate resources from low-value (high sunk cost) to high-value (fresh) opportunities

Action: Schedule quarterly "zero-based" reviews where every initiative must re-justify itself as if newly proposed.

5. Reframe "Waste" as Learning

Eliminate the emotional driver by redefining sunk costs not as waste but as tuition for knowledge:

• "We spent $500K learning that this market isn't viable" is progress, not waste
• "I spent 3 years learning what I don't want in a career" is valuable information
• Shift from "I wasted X" to "I invested X in learning Y"
• This reduces ego protection motivation: cutting losses isn't admitting failure, it's capitalizing on learning

Action: When cutting a project with sunk costs, write a "what we learned" document. Frame the investment as education expense.

6. Use Precommitment and External Oversight

Create accountability structures that force rational evaluation:

• Board oversight: Require independent board approval for continued investment beyond milestones
• Peer review: Have uninvested third parties evaluate whether to continue
• Sunset clauses: Projects automatically terminate unless explicitly renewed with fresh justification
• External advisors: Consultants with no emotional attachment evaluate continuation decisions

Action: For major initiatives, appoint an advisor with no sunk cost attachment who must approve continued investment quarterly.

7. Institutionalize "Permission to Quit"

Create organizational culture that celebrates cutting losses, not just persistence:

• Celebrate kills: Publicly recognize teams that shut down failing projects early
• Failure resume: Share lessons from discontinued initiatives, normalizing pivots
• Anti-awards: "Best Project Killed" award for identifying sunk costs and cutting losses
• Language shift: Replace "quitting" with "reallocating to higher-value opportunities"

Action: In retrospectives, explicitly discuss: "What should we have killed sooner?" Reward honest answers.

Example

Scenario: Your startup has spent 18 months and $1.5M building a consumer app. User growth is flat, engagement is low, but you're "almost done" with a major feature.

Sunk Cost Fallacy in action:

• Reasoning: "We've invested so much time and money. We can't give up now. We're 80% done with this feature—just a few more months and it'll turn around. Shutting down now would mean all that work was wasted."
• Emotional drivers: Fear of admitting failure, guilt about money spent, attachment to original vision, team morale concerns
• Decision: Continue development for another 6 months, burn another $500K
• Reality: Feature ships, still no traction. Company runs out of money. Total loss: $2M and 24 months
• Result: Opportunity cost was massive—could have pivoted to adjacent market, returned capital to investors, or started new venture

Better approach using this framework:

1. Fresh start test:
   • Question: "If I were starting a company today with $1.5M and an 18-month runway, would I build this exact product for this exact market?"
   • Honest answer: "No. I'd target a different customer segment or a different problem entirely."
   • Conclusion: Sunk costs are biasing the decision. The product isn't good enough to start today, so don't continue it.
2. Separate accounting from decision-making:
   • Sunk costs (irrelevant): $1.5M spent, 18 months invested, feature 80% complete
   • Future costs/benefits (relevant): $500K+ to finish, 6+ months to ship, low probability of traction based on data
   • Decision: ROI on future investment is negative. Cut the project.
3. Kill criteria:
   • Revisit criteria set at launch: "If we don't have 10K MAU by Month 12, pivot or shut down"
   • Reality: Month 18, only 2K MAU, declining engagement
   • Criteria clearly met. Honor the commitment.
4. Quarterly portfolio review:
   • Ask: "What's the best use of our remaining $500K and 6 months?"
   • Options: (A) Finish this feature, (B) Pivot to enterprise SaaS, (C) Return capital to investors, (D) Explore adjacent market
   • Analysis: Option A has lowest expected value. Sunk costs make it emotionally appealing, but logically weak.
5. Reframe as learning:
   • "We spent $1.5M learning that consumer social apps require virality we don't have, and that enterprise SaaS is a better fit for our team's strengths. That's not waste—that's tuition."
   • Document learnings for future ventures. Framed this way, cutting the project feels like capitalizing on knowledge, not admitting failure.
6. External oversight:
   • Present situation to board: "Should we continue?" Board members, uninvested in sunk costs, recommend pivot or wind-down
   • Their advice isn't clouded by 18 months of emotional attachment
7. Permission to quit:
   • CEO tells team: "Shutting this down isn't failure. We learned invaluable lessons. Now we're reallocating to higher-probability opportunities. That's smart strategy."
   • Celebrate the team's honesty in recognizing the pivot point rather than doubling down

Result: Pivot to enterprise SaaS using lessons learned. Remaining $500K goes to new product. Team ships v1 in 4 months, gets first paying customers in Month 6, raises Series A in Month 12. Cutting losses early created opportunity for success.

Anti-Patterns

"Just a little more": Repeatedly extending deadlines or budgets with "just one more sprint" or "just $50K more," indefinitely deferring the kill decision.

Honoring the team's effort: Continuing projects to avoid "wasting" team members' hard work, conflating respect for effort with rational resource allocation.

Public commitment trap: Refusing to cancel because of public announcements, press releases, or stakeholder expectations, prioritizing reputation over outcomes.

Incremental entrenchment: Making small continued investments that individually seem reasonable but cumulatively represent massive sunk costs.

Success theater: Redefining metrics to make failing projects appear successful ("engagement is up!" when it's still abysmal), justifying continued investment.

Comparison to worse alternatives: "At least we're not as bad as [failed competitor]" as justification to continue rather than evaluating on absolute merits.

Waiting for breakeven: "We've spent so much, we just need to get to breakeven to justify it" when breakeven itself isn't strategically valuable.

Ego protection over outcomes: Treating project cancellation as personal failure rather than organizational learning, preventing honest evaluation.

Related Frameworks

• Loss Aversion: Losses loom larger than gains, making sunk costs feel more painful than opportunity costs
• Status Quo Bias: Preferring current state over change, even when change is superior
• Escalation of Commitment: Intensifying investment in failing courses of action, closely related to sunk cost fallacy
• Endowment Effect: Overvaluing things we already own/created, related to attachment to sunk costs
• Opportunity Cost: The value of alternatives foregone; sunk cost fallacy ignores opportunity costs
• Prospect Theory: Reference-dependent preferences explain why sunk costs loom large
• Confirmation Bias: Seeking evidence that sunk costs will pay off, ignoring disconfirming data
• Zero-Based Budgeting: Countermeasure requiring fresh justification for all expenditures
• Pre-Mortem: Imagining failure upfront to set kill criteria before sunk costs accumulate
• Pivot vs. Persevere: Lean Startup framework for deciding when to change direction despite sunk costs