Escalation of Commitment: Why Teams Keep Funding Failing Projects

Escalation of commitment occurs when people keep investing resources in a course of action despite negative feedback, with earlier commitments helping drive the decision to continue. In project management, it can look like another budget increase, another postponed benefit, or another rollout phase approved because stopping would be difficult to explain.

“We have already spent too much to stop” is a familiar warning. But an overrun does not automatically make continuing a mistake. The useful question is: does the next commitment make sense given the remaining benefits, costs, risks, and alternatives?

This guide explains the concept, gives workplace examples, and provides a practical review process. Jump to the worked example or copy the project commitment review template.

In this guide

1. What is escalation of commitment?
2. Escalation vs sunk-cost fallacy
3. Why teams keep committing
4. Seven warning signs
5. Workplace examples
6. When continuing is justified
7. Continue, change, pause, or stop
8. Worked rollout example
9. Copyable review template
10. How to reduce escalation
11. Frequently asked questions
12. Research and scope

What is escalation of commitment?

Escalation of commitment describes persistence in a questionable course of action when unfavorable information should prompt a fresh assessment. Earlier investments, decisions, or promises can make further commitment feel necessary even when the current case has weakened.

For example, a company funds a product launch. Trials show weak demand, but the sponsor requests a larger launch budget primarily to recover the original investment and protect a public promise. The concern is not the additional budget alone; it is the reasoning behind it.

Barry Staw’s 1976 study examined a simulated business investment decision. Participants committed the most resources when they were personally responsible for negative consequences. This experimental finding helps explain why reviewing one’s own decision can be difficult; it does not establish that every struggling project should stop.

Escalation of commitment vs sunk-cost fallacy

The concepts overlap but emphasize different things. Sunk-cost fallacy concerns how unrecoverable past costs influence a present choice. Escalation of commitment describes a pattern of further commitment despite negative feedback. Sunk-cost reasoning can contribute to that pattern alongside reputation, group expectations, incentives, and organizational constraints.

Why teams keep committing to struggling projects

Escalation is not adequately explained by calling a sponsor stubborn. A multilevel review of escalation research examines influences at group and organizational levels and beyond the organization. The setting matters as well as the individual.

• Personal responsibility: changing direction can feel like admitting that the original decision was wrong.
• Public promises: an announced launch or strategic commitment can make revision costly to someone’s reputation.
• Incentives: people may be rewarded for launching or completing work while receiving little credit for preventing further waste.
• Filtered information: summaries can preserve reassuring status while excluding inconvenient evidence.
• Perceived proximity: “almost finished” can become persuasive even when essential work remains uncertain.

Groupthink may suppress challenge. The Abilene paradox describes a different problem: people support a decision they privately do not want because they misread others’ preferences. Either can complicate a review, but neither is another name for escalation of commitment.

Seven warning signs in project reviews

1. Past spending leads the argument. The request explains what has been invested more clearly than what the next investment will achieve.
2. Success criteria move after disappointing results. Revised targets appear without a transparent explanation or renewed approval.
3. “Almost finished” replaces evidence. Remaining work is repeatedly described as minor without a credible completion assessment.
4. The review hears only from defenders. Users, operational teams, or independent reviewers cannot challenge the proposal.
5. Alternatives receive unequal scrutiny. Stopping must meet an impossible standard while continuing is treated as the default.
6. Benefits stay vague while commitments become specific. Money and people are requested now; value remains an aspiration.
7. No reconsideration trigger exists. Nobody can say what evidence would justify changing direction.

These are prompts for investigation, not a diagnostic score. Changed targets, for example, may reflect legitimate learning. Ask who approved the change, which evidence supports it, and whether the updated case still stands.

Escalation of commitment examples at work

The following scenarios are hypothetical illustrations, not reported client cases.

A software rollout keeps expanding

A pilot reveals low adoption and unresolved integration failures. The steering committee approves wider deployment because implementation has already consumed a year. A better review would test whether expansion has a credible operational case before adding users and support costs.

A product launch receives repeated rescue funding

Customer tests weaken the demand assumption, but each marketing request promises that more exposure will solve the problem. The review needs to distinguish inadequate reach from an offer customers do not value, using evidence that could support either conclusion.

A transformation programme protects its original scope

Several initiatives compete for the same specialists. Benefits slip, yet every workstream remains “strategic.” Comparing alternative uses of capacity through transformation portfolio management can reveal what continuing each commitment prevents the organization from doing.

When continuing is the right decision

A troubled project can still have a strong remaining case. The original forecast may have been wrong while the next step remains better than stopping. Compare options from the current position rather than pretending the organization can return to the beginning.

• Unrecoverable spending cannot be recovered merely by spending more.
• Usable assets and capabilities can still create future value.
• Exit, transition, and contractual consequences may change the cost of stopping.
• Remaining risks and opportunity costs affect every option, including doing nothing.

A bounded experiment may be worthwhile if it can resolve an important uncertainty at an acceptable cost. Continuing becomes more defensible when the rationale identifies what will be learned, what resources are exposed, and what happens if the evidence disappoints.

How to decide whether to continue, change, pause, or stop

The following six-step process is a practical synthesis, not a validated diagnostic model. Use it at a funding gate, rollout decision, major scope change, or material breach of an assumption.

1. Define the next commitment

State the decision in concrete terms: approve another quarter, fund a limited pilot, sign a contract, or expand to additional teams. Name the amount, people, timing, and accountable decision-maker. “Do we believe in the programme?” is too vague.

2. Separate past spending from future consequences

Record past expenditure for accountability, then compare the incremental consequences of each option. Include reusable assets, remaining obligations, switching costs, operational disruption, and opportunity cost. Do not count already-spent money as a benefit of continuing.

3. Reassess the evidence

List the assumptions that supported approval and the evidence now available. Distinguish observations from explanations: “only three teams use the workflow” is an observation; “they need more training” is a hypothesis. A delivery assurance review can help examine the evidence behind reported confidence.

4. Compare credible alternatives

Assess continuation, changed scope, a time-bounded pause, and stopping where feasible. Apply the same criteria to each. A pause needs an owner, a question to resolve, and a deadline; otherwise it can preserve cost without producing useful evidence.

5. Challenge the recommendation independently

Include someone who can question the original assumptions without having to defend the original approval. Ask for the strongest contrary evidence. Preserve the sponsor’s knowledge while making decision authority explicit through appropriate TMO governance.

6. Record review conditions

Document the chosen option, assumptions, resource limit, action owners, and review date. Specify evidence that would reopen the decision. Exceptions to agreed conditions should require a recorded rationale rather than an automatic extension.

Worked example: the rollout that keeps asking for another quarter

This is a hypothetical software rollout. All figures and thresholds below are illustrative, not industry benchmarks.

An organization has spent $600,000 on a new workflow platform. The sponsor requests another $250,000 for wider deployment. Pilot adoption is weak and integration issues remain unresolved. The claim that stopping would “waste the $600,000” does not establish what the next $250,000 will deliver.

In this example, the decision-maker authorizes a six-week pilot capped at $40,000, rather than approving the full expansion request. This is justified only if the pilot can produce evidence that materially changes the next decision.

Before work starts, the group agrees three local conditions:

• At least 80% of eligible pilot transactions use the new workflow for two consecutive weeks.
• All integration checks designated critical before the pilot pass, with no unresolved critical incident at review.
• The revised case explains why remaining benefits justify the remaining costs and risks compared with retaining the existing solution.

The product owner supplies usage evidence, the integration lead supplies test results, and an operational representative challenges readiness. The sponsor decides at the end of week six. Failure to meet a condition triggers reassessment; meeting the conditions does not automatically authorize full rollout.

The $40,000 must be included in subsequent funding calculations. Do not assume that spending it reduces the original $250,000 request dollar for dollar: remaining work and costs need re-estimation. The existing $600,000 remains part of the history, not the reason to approve either amount.

Copyable project commitment review template

Copy this worksheet into a project document or funding proposal. Keep the reasoning short enough to review, and link to the supporting evidence.

PROJECT COMMITMENT REVIEW

Decision required:
Resources requested and commitment date:
Accountable decision-maker:

Original assumptions:
New evidence and its limitations:
Past expenditure (context, not a reason to continue):
Usable assets and remaining obligations:

OPTIONS
Continue:
Change scope or run a bounded test:
Pause for specific evidence:
Stop:

For each option:
- Remaining costs, capacity and time
- Expected benefits and uncertainty
- Operational risks and exit consequences
- Opportunity cost and dependencies

Strongest objection to the recommendation:
Independent reviewer and response:

Decision and rationale:
Approved limit and action owners:
Evidence thresholds and review date:
What happens if conditions are not met:

For the example above, the decision record would state: “Approve a six-week pilot capped at $40,000; expansion remains unapproved. Review usage, critical integration checks, and the revised future case at week six.” That is more informative than “continue with close monitoring.”

How to reduce escalation of commitment before it starts

Set review conditions while the team is still willing to discuss alternatives. A premortem analysis can surface plausible failure causes before approval. Convert important concerns into evidence checks, owners, and triggers; an imagined failure is not itself proof of a risk’s probability.

Use staged commitments where practical, preserve the assumptions behind each approval, and review benefits as well as completed activities. Give people a credible way to recommend stopping or changing work without treating every revision as misconduct. Accountability should examine decision quality and evidence, not reward persistence regardless of results.

Independent challenge also needs authority and access to information. Adding a reviewer will accomplish little if unfavorable findings are excluded from the funding discussion.

Frequently asked questions about escalation of commitment

What is a simple example of escalation of commitment?

A team keeps funding a product despite weak customer evidence mainly because it has already invested heavily and promised a launch. Further investment might still be justified, but that requires a current case rather than an appeal to past effort.

Is escalation of commitment the same as sunk-cost fallacy?

No. Sunk-cost fallacy concerns the influence of unrecoverable past costs. Escalation describes continued commitment despite negative feedback and can involve additional pressures, including responsibility, reputation, group dynamics, and organizational incentives.

Is continuing an over-budget project always a mistake?

No. Compare the remaining benefits, costs, risks, and alternatives from the current position. A project can exceed its original budget and still have a stronger future case than stopping or switching.

How can a project manager challenge escalation?

Define the next approval, present the changed evidence, and compare realistic options using consistent criteria. Ask for an independent review and a documented decision. Challenging a rationale is more useful than accusing a sponsor of bias.

Can a premortem prevent escalation of commitment?

A premortem can help establish concerns and review conditions before commitment. It cannot guarantee that later decisions will use the evidence. Follow-through, decision authority, and the treatment of dissent still matter.

Research and scope

• Staw, B. M. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action. Foundational experimental research on responsibility, negative consequences, and further investment.
• Sleesman, D. J., Conlon, D. E., McNamara, G., and Miles, J. E. (2012). Cleaning Up the Big Muddy: A Meta-Analytic Review of the Determinants of Escalation of Commitment. A synthesis of research on escalation’s determinants.
• Putting Escalation of Commitment in Context: A Multilevel Review and Analysis. A review examining group, organizational, and external context.

The review process, worksheet, and warning signs are practical editorial syntheses, not validated diagnostic instruments. The project examples, budgets, and pilot thresholds are hypothetical. Research findings do not establish a universal threshold for continuing or stopping a project.