Most corporate disasters are not first mistakes. They're second mistakes. The first mistake is picking wrong. The second mistake is refusing to update. That's where escalation of commitment turns a setback into a catastrophe.
Sunk cost is money, time, or effort already spent that you cannot recover. The mind treats past investment as a reason to keep investing. Logically, it isn't. Decisions should be based on the future, not on what's already gone.
This post connects to Post 3 (the decision operating system) and Post 5 (mental models). Escalation is model rigidity plus identity lock. The framework provides the countermeasure.
The Real Driver: Reputation and Identity
Leaders escalate for emotional reasons dressed in rational clothing. The underlying drivers:
- Looking incompetent: Stopping admits the initial decision was wrong.
- Losing political capital: The leader staked credibility on the project.
- Admitting the model was wrong: The strategy was tied to the leader's identity.
These are legitimate concerns. They're also the mechanism of escalation. The cost of looking wrong feels unbearable, so the leader doubles down instead of updating.
The first mistake is picking wrong. The disaster is refusing to update.
The Repair Fantasy
At organizational scale, the repair fantasy sounds like: "If we ship v3, it'll justify v1 and v2." The belief is that future success will retroactively validate past failures.
This is escalation, not strategy. Past costs are gone. They can't be recovered by future spending. But the emotional logic is powerful: the more you've invested, the more you need the outcome to validate the investment.
The Product That Wouldn't Die: A company keeps shipping features for a shrinking user segment because they've already invested three years in the product. Each quarter, the metrics worsen. Each quarter, the response is "one more sprint." The product is dead. The organization can't admit it because the sunk cost feels too high to write off.
Detection Signals
Signs that escalation is operating:
- "We can't waste what we've built": Past investment as justification for future investment.
- "Just one more quarter": Endless extensions without exit criteria.
- Metrics redefined midstream: When the original goal isn't met, success definitions shift.
- Dissent punished: People who raise concerns are coded as "not committed."
- Updates framed as disloyalty: Changing direction feels like betrayal.
These are not proof of escalation. They're signals to investigate.
Wise Persistence vs. Escalation
Persistence is not inherently virtuous. It depends on the situation.
Wise persistence:
- Evidence of improvement over time
- Learning and adaptation happening
- Costs are tolerable and bounded
- Strategy is changing based on feedback
Escalation:
- Same strategy, higher cost
- Increasing rigidity and defensiveness
- Secrecy and rationalization
- Opportunity costs ignored
The Future-Only Question
The most powerful diagnostic: "If we were starting today with no prior investment, would we allocate resources here?"
This question strips away the sunk cost and forces evaluation based on current evidence and future potential. If the answer is "no," you're likely escalating rather than persisting wisely.
Kill criteria aren't pessimism. They're governance. Exit ramps are professionalism, not failure.
Exit Ramp Charter
For any significant initiative, define in advance:
- Objective: What success looks like (specific, measurable).
- Metrics (3-5): What we'll track to know if it's working.
- Kill criteria: What failure looks like. Specific thresholds that trigger exit.
- Review cadence: Monthly or quarterly check-ins against the metrics.
- Owner + decision rights: Who decides to continue, pivot, or stop.
- Comms plan: How to explain a pivot without shame.
- Post-mortem plan: How to capture learning regardless of outcome.
- Vague metrics that can be reinterpreted to show success
- No kill criteria, so there's no trigger for stopping
- Reviews that get delayed when results are poor
- Political override without evidence (continuing because of status, not data)
The Executive Retention: A company retains a bad executive because firing would admit the hire was wrong. The board championed the hire. Termination would embarrass them. So the executive stays, performance suffers, and the team loses trust. The sunk cost is political capital. The future cost is organizational effectiveness.
Portfolio and Bounded Bets
Design investments so you can stop without ego collapse. If every initiative is framed as existential, every exit feels like failure. Portfolio thinking creates room for learning.
Some bets should be bounded from the start: defined budget, defined timeline, defined kill criteria. If they don't work, you learn and move on. The framing prevents escalation because stopping was always an acceptable outcome.
Tie to Mental Models
Escalation is model rigidity plus identity lock. The model becomes "this strategy will work." The identity becomes "I'm the kind of leader who sees things through." When the model fails, updating feels like identity betrayal.
The solution is to treat strategies as hypotheses rather than commitments. Hypotheses get tested. Commitments get defended. The frame determines the behavior.
Tie to Winner Stories
Survivorship bias legitimizes escalation. The narratives say: "All great wins required perseverance." What they don't show: the many identical perseverance stories that ended in failure. Winner stories are selected on outcomes, not on what caused them.
Weekly Practices
- Add kill criteria to one active project: Define what would make you stop.
- Run the future-only question on one budget line: If starting fresh, would you fund this?
- Schedule a review meeting that includes someone independent: Someone without sunk cost in the decision.
Objections and Clarifications
"Stopping kills morale."
Unclear. Dragging a zombie project often kills morale more. Teams know when something isn't working. Continuing despite evidence damages trust more than honest stopping.
"We'll look bad."
Better to look adaptive than stubborn. Organizations that can pivot based on evidence are more resilient than those that persist based on hope.
When success definitions change midstream, ego has entered the room. Metrics drift is a red flag for escalation.
If your organization struggles to stop initiatives or leaders double down on failing strategies, we can audit your decision processes and build exit ramp governance.
Request AssessmentThis content is educational and does not constitute business, financial, or medical advice.