Breaking the Escalation Trap: When Good Decisions Go Bad

The worst decisions often follow the best ones.

This is not a paradox. It is a structural problem in how we process commitment. When we make an initial choice—invest in a project, hire someone, back a strategy—we create a psychological anchor that distorts every subsequent decision. We do not evaluate what comes next on its merits. We evaluate it through the lens of protecting what we have already chosen. The decision that felt rational at the time becomes a cage.

This is escalation of commitment, and it is everywhere. A company launches a product line. Early sales disappoint. Rather than cut losses, they increase marketing spend, reasoning that the initial investment must be justified. A researcher pursues a hypothesis that data no longer supports, but abandoning it means admitting the previous months were wasted. A leader doubles down on a hire that is clearly underperforming because acknowledging the mistake feels like a referendum on their judgment.

The mechanism is not irrational in the moment. It is rational given a false premise—that the past investment is still in play, still recoverable, still relevant to the choice at hand. But sunk costs are not in play. They are gone. The only rational question is whether the next dollar, the next month, the next resource should go forward. Yet we consistently treat the initial commitment as a variable we can still influence by throwing more at it.

What makes this particularly insidious is that escalation often wears the mask of consistency. We tell ourselves we are being decisive, committed, unwilling to quit. We frame retreat as weakness. In some contexts—long-term relationship building, skill development, organizational culture—consistency is genuinely valuable. But escalation of commitment is not consistency. It is the confusion of past choices with present ones. Consistency would mean applying the same decision criteria to new information. Escalation means applying emotional protection to new information.

The research is clear on the conditions that make escalation worse. Public commitment amplifies it—when others know about your choice, backing away feels like public failure. Ego involvement deepens it—when the decision reflects on your identity or competence, the stakes feel personal rather than financial. Ambiguity extends it—when success metrics are unclear, you can always interpret new data as "not yet proven" rather than "proven wrong."

But here is what matters for decision-makers: escalation is not a character flaw. It is a decision architecture problem. It emerges from how we frame choices, how we measure progress, and what we allow ourselves to see.

The antidote is not willpower. It is structural. First, separate the evaluation of past decisions from the evaluation of future ones. Ask explicitly: "If we had not already spent this money, would we spend it now?" If the answer is no, the decision is made. The past choice is not on trial. The future choice is.

Second, build in predetermined exit criteria before you commit. Not pessimistically, but realistically. What would need to be true for this to be working? What would need to be false? When would we reassess? This removes the moment of escalation from the heat of commitment and places it in the clarity of planning.

Third, create distance between the person who made the initial choice and the person who evaluates it. Not as punishment, but as protection. Fresh eyes are not disloyal. They are honest.

The hardest part is accepting that good decisions can produce bad outcomes, and that recognizing this is not failure—it is learning. The trap closes when we treat every setback as a reason to prove the original choice was right. The way out is to treat every setback as new information about a new choice.

The best decision-makers are not those who never escalate. They are those who notice when they are about to, and stop.