The entrepreneur who cannot leave a failing direction despite clear evidence that it is failing is not making a logical error that logical correction can address. They are experiencing the predictable output of a mechanism that Staw’s research documented with unusual precision: the psychological pressure to justify prior investment is among the most reliably observed patterns in decision research, and it is specifically amplified by the personal responsibility that entrepreneurial decision-making generates.

Staw’s escalation of commitment: personal responsibility as the primary amplifier

Staw’s (1976) foundational research established the key variable in escalation of commitment. People who made the initial decision show substantially greater escalation than those who inherited the same decision in the same state — because the personal responsibility for the prior investment creates the self-justification motivation that sustains commitment regardless of the performance evidence. The escalation is not primarily about the economic logic of sunk costs; it is about the psychological pressure to demonstrate that the prior decision was sound.

For the entrepreneur who made the founding decision, designed the initial product, hired the initial team, and built the culture around the initial direction, each of these prior commitments creates a separate escalation driver. The founding decision created the responsibility. The product design created the creative investment. The hiring created the relational investment. The culture-building created the identity investment. Each additional layer of prior commitment increases the personal responsibility burden that leaving would require admitting was misdirected — which the self-justification mechanism is specifically designed to prevent.

Arkes and Blumer’s (1985) research on the sunk cost effect confirmed the mechanism’s pervasiveness and resistance: participants consistently allocated more resources to losing options when they had already invested in them, even in experimental contexts specifically designed to isolate the sunk cost from any genuine future value. The awareness that sunk costs are logically irrelevant to future decisions does not eliminate the mechanism’s effect — because the mechanism is self-justification, not cost calculation, and self-justification does not respond to logical correction the way calculation errors do.

The Kahneman loss aversion amplification: departure feels like confirming the loss

Kahneman and Tversky’s (1979) prospect theory predicts the specific emotional mechanism that sustains escalation alongside the self-justification mechanism. Departing from the invested direction is not experienced as a neutral reallocation of resources toward a better opportunity. It is framed against the reference point of the total investment already made — and against that reference point, departure means accepting the full confirmed loss of everything that has been invested.

The loss aversion coefficient — losses loom approximately twice as large as equivalent gains — ensures that the anticipated pain of confirming the sunk cost loss significantly outweighs the anticipated benefit of the reallocation, even when the reallocation’s expected value is substantially higher. The entrepreneur who has invested two years and significant capital in a direction does not experience the exit decision as a neutral comparison between two future opportunities. They experience it as: confirm the loss of two years and the capital, versus maintain the possibility that the investment will yet be justified.

The emotional structure of this comparison systematically favours continuation regardless of the future-oriented evidence — which is why the sunk cost fallacy is so resistant to correction by presenting better future opportunities. The future opportunity is competing not against the failing direction’s future expected value but against the full confirmed loss that departing would crystallise.

The confirmation bias companion: how escalation amplifies information distortion

The Boehne and Paese (2000) research on escalation and information processing documented the companion mechanism: committed decision-makers selectively process confirming information about their prior investment while discounting disconfirming evidence. The escalation and the confirmation bias are mutually reinforcing — the escalation creates the self-justification pressure to maintain the commitment, and the confirmation bias then filters the incoming evidence in ways that sustain the commitment’s apparent justification.

The entrepreneur who is escalating into a failing direction is not seeing the market evidence clearly and choosing to continue despite it. They are selectively processing the evidence in ways that confirm the continuation, and the selective processing is driven by the same self-justification pressure that the escalation mechanism generates. The direction appears to have more evidence in its favour than it does, because the evidence processing is contaminated by the commitment that the prior investment created.

The pre-committed kill criterion: the structural correction the research supports

The research on structural corrections for sunk cost escalation converges on a common design principle: establish the exit criteria before the investment creates the justification pressure. Duke’s (2022) Quit research and the Gollwitzer implementation intention framework together predict why this works. The pre-committed kill criterion — “if this specific metric has not been reached by this specific date, we will stop this direction” — is established when the decision-maker has no sunk cost to justify, which means the self-justification mechanism has no existing commitment to protect. The exit condition is set in the absence of the escalation pressure that would later prevent it.

When the kill criterion date arrives, the decision is mechanical rather than evaluative: the metric was met or it was not. The self-justification mechanism has no leverage on a pre-committed criterion because the commitment is to the criterion rather than to the direction. The entrepreneur who established the kill criterion cannot justify continuing past it on self-justification grounds without explicitly overriding a prior commitment they made to themselves — which requires acknowledging the override rather than experiencing continuation as the natural outcome of the evidence.

The Concorde case illustrates the counterfactual: the kill criterion that would have prevented the most documented large-scale sunk cost escalation in business history was not established before the investment began, leaving every subsequent decision to be evaluated against the accumulated prior investment rather than against the future commercial case. Each additional investment was justified by the commitment that the previous investments had created.

Books worth reading on this

The Checklist Manifesto by Atul Gawande is the most practically applicable available account of how pre-committed structural procedures — established before the pressure of the specific situation creates the cognitive distortions this article describes — produce better decisions than in-the-moment judgment under equivalent conditions. Gawande’s account of how pre-commitment to specific decision criteria protects against the self-justification and confirmation bias mechanisms that high-stakes in-the-moment decisions generate maps directly onto the kill criterion structural correction this article identifies as the most research-supported escalation prevention tool.

If the dynamics described here are significantly affecting your wellbeing, speaking with a psychologist is the right next step. UK: Samaritans (116 123, free, 24/7). Mind (0300 123 3393). BACP: bacp.co.uk/search/Therapists. Crisis Text Line — text HOME to 741741 (US, UK, Canada, Ireland). International: internationaltherapistdirectory.com.

This article is for educational and informational purposes only. Sources: Staw, B.M. (1976), Knee-Deep in the Big Muddy: A Study of Escalating Commitment, Organizational Behavior and Human Performance, 16(1), 27–44. Arkes, H.R. & Blumer, C. (1985), The Psychology of Sunk Cost, Organizational Behavior and Human Decision Processes, 35(1), 124–140. Kahneman, D. & Tversky, A. (1979), Prospect Theory: An Analysis of Decision Under Risk, Econometrica, 47(2), 263–291. Boehne, D.M. & Paese, P.W. (2000), Deciding Whether to Complete or Terminate an Unfinished Project, Organizational Behavior and Human Decision Processes, 81(2), 178–194. Gollwitzer, P.M. (1999), Implementation Intentions, American Psychologist, 54(7), 493–503. Gawande, A. (2009), The Checklist Manifesto, Metropolitan Books. Munger, C.T. (2005), Poor Charlie’s Almanack, Donning.