When Sunk Cost Matters

As Peter Drucker stated long ago, “What gets measured, gets managed.” Metrics matter: they drive behavior.

Consider a company trying to stay alive in a declining market with obsolete technology: it can be done, but they probably should not make share gains one of their measures of success. Conversely, the leader in a new, rapidly growing, rapidly changing market might not want to lean too heavily on cost control metrics, which could undermine their ability to adapt nimbly.

With this in mind, Sopheon’s Michael Tulaney recently presented a webinar  in which he challenged the orthodox view that financial decisions should always exclude sunk cost. A sunk cost is one that has already been incurred and thus cannot be recovered. What has already been spent, management theory argues, should not be considered when considering future expenses. The enterprise should focus on maximizing returns on expenses/investments that can be controlled – e.g., future ones – rather than engaging in the “sunk cost fallacy” in which good money is thrown after bad to justify past decisions.

While this is certainly correct in theory, the webinar argued, consider the real world behaviors it might drive in the area of innovation project management.

I once had a colleague who was frustrated because she had never launched a new product. While reviewing some discouraging data that had just come in on her latest project, she exclaimed, “We’ve got to get this out the door or my career is over!” Her solution: accelerate spending so that by the next management portfolio review, the project would look good when excluding sunk cost.

One of the philosophical underpinnings of Stage-Gate® process design is that project risk should be managed by deferring commitment of expensive resources until some of the project uncertainties have been reduced. At each successive stage, there should be less uncertainty as various questions get answered, and thus there is less risk in committing more resources. But if a manager knows that sunk costs will be excluded from future management decisions, he or she may feel incented to commit resources at an earlier, riskier stage.

Another real-world consideration is how to evaluate project success during post-launch reviews. Should project NPV be reset each time sunk costs are excluded from a decision? How do we evaluate a project that fails against going-in assumptions but succeeds against later assumptions that excluded sunk costs? This is not an academic issue, as future policies and strategic choices are built on the organization’s learnings from past projects.

Mike and I are certainly not arguing for always including sunk costs in project analyses and decisions; the sunk cost fallacy is real. But likewise, we do not believe they should always be excluded. What is needed is a balanced approach that may help to avoid incenting spending acceleration. Portfolio decisions should acknowledge failure to achieve going-in assumptions, but at the same time should be based on rational, forward-looking criteria. This is a case for both, not either.

This post has discussed one example of how a key metric can drive managerial behavior in an undesirable direction. Sopheon Innovation Management Consultants are experts in partnering with their clients to identify an optimal set of decision-supporting metrics. We would love to talk with you about your critical issues.

Key consideration: Has your company debated which financial metrics are critical for analyzing NPD efforts? What have you decided regarding sunk costs and what have been the repercussions?

2018-05-02T16:06:00-05:00September 14th, 2015|

About the Author:

Mark Friedman
Mark has over 30 years of experience managing product innovation, project portfolios, business planning, and team development. After nearly a decade with P&G in brand management, he helped lead marketing and innovation at several leading CPG companies, and built a successful consulting practice.


  1. Mark Friedman
    Mark Friedman February 21, 2019 at 5:03 PM - Reply

    Hi Kenneth – thank you for your comments! I completely agree with your second scenario – sunk costs should not lead to rejecting a project when the future looks profitable. In the first scenario, I would argue that sunk costs and value recovery might be be considered because the future outlook of both projects is identical; in that case, however, a coin toss may be as reasonable a solution as considering sunk costs. Past expenditures generally should not play a role in developing a financial model of the future. However, in the real world, there are behavioral implications of ignoring sunk costs, as discussed in the article.

  2. Avatar
    Kenneth Siler February 21, 2019 at 2:54 PM - Reply

    Thanks for this article. Yes, the orthodox view is for sunk costs to be excluded. This orthodox view sometimes fails the test of practicality in my current role, but in a different way than your article examines.

    Scenario one: In my role, I must assign project costs and will need to choose which project to undertake when we are close to max capacity. For simplicity in this example, I can only choose between one of two projects, all project costs are equal, and both projects generate similar returns.

    However, staff hours for each project are not equal and no new staff will be hired or laid off regardless of the project decisions. Unless I’m mistaken, the stability in the paid staff salaries are now sunk costs because I will have paid them regardless of which project that I choose or even if I reject both. The question becomes one of value recovery (or proper resource utilization), a major source of the sunk cost fallacy. To choose the project that offers best value to the company, I do need to consider these sunk costs of staff salaries as 1) actual projects costs and 2) as evaluation points when choosing between the two.

    Scenario two: Alternatively, when we are not at max capacity and staff is underutilized, if I include sunk costs for salary in the project evaluation for less valuable projects, this can make projects look unprofitable… while in reality having some projects with meager profit margins are better than no projects because a lack of projects means that highly trained staff will be forced to spend hours in areas that do not recover the value of their salaries. In this case, I would not include the sunk costs of staff salaries.

    There is real danger of falling into the sunk cost fallacy, but value recovery in staff salaries doesn’t seem to be something that I can discount in either scenario. There is also a question of comparing past project costs when cost evaluations have variability in sunk cost inclusion. Care to comment or correct my understanding?

Leave A Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.