Back to Blog & Insights

Leadership

Before the Next Spend, Decide What Would Make You Stop

A strong project review does not defend yesterday’s investment. It defines the outcome still worth pursuing, tests the next commitment against current evidence and establishes clear conditions for stopping or restarting.

By WWA Editorial Team 3 min read
Business leader reviewing project evidence and stop criteria with a colleague before approving the next investment

The difficult decision is the next one

The prototype missed its performance target. The launch date is slipping, and the team has already spent months refining the concept. Now the project leader faces a familiar problem: approve another development cycle or stop.

The weakest review question is, “How can we walk away after investing so much?” Past spending matters as context. It can reveal what the team learned, which capabilities it built and whether execution problems are repairable. But spent money and elapsed hours cannot be recovered. The decision in front of a leader is whether the next dollar, hour and choice are justified by current evidence.

That distinction becomes harder when the decision maker supported the original direction. Three incentivized studies reported in Frontiers in Psychology found that reinvestment was more likely when people were responsible for prior choices, preferred the initial investment or viewed outcomes as losses. A project review therefore needs more than confidence and persistence. It needs conditions established before the next commitment.

Build a stop-or-continue review

Name the outcome that remains valuable

Start by separating the desired business result from the current route. “Launch this prototype” is a route. “Give midmarket customers a faster way to complete a recurring task at a viable cost” is an outcome. If the outcome remains commercially useful, the team may test another route. If customer need, strategic fit or economics have changed, continuing the existing plan will not restore its value.

List evidence that could change the decision

Write down what would support continuation and what would require stopping. Relevant evidence might include customer adoption, technical performance, delivery capacity, unit economics or a partner commitment. Make each item observable. “Customers seem interested” invites interpretation; “five pilot customers use the feature weekly for four weeks” creates a test.

A 2021 Frontiers in Psychology delay-and-choice task found that participants could base quitting partly on time already waited; the observed sunk-cost sensitivity disappeared with an attention check. The setting was experimental, not a corporate launch, but the finding supports a practical discipline: direct attention to decision-relevant evidence instead of allowing elapsed time to dominate the review.

Set the date and threshold now

Before authorizing more work, specify the smallest next investment, the review date and the threshold it must meet. Agree on what happens if results fall below, meet or exceed that line. Precommitting reduces the temptation to reinterpret disappointing results after another deadline passes.

Then ask an uninvolved colleague to challenge the case. That person should test assumptions, identify missing alternatives and ask whether the same proposal would look attractive without the project’s history. The role is not to kill the work; it is to improve the decision.

Preserve the learning, end the weak route

Consider an illustrative example. Maya leads a product launch for a workflow tool. Its prototype was supposed to cut task-completion time by 30 percent, but testing shows only a 10 percent improvement. Her team proposes another broad redesign because it has already invested six months.

Maya reframes the review. The worthwhile outcome is reducing a costly customer bottleneck, not rescuing this interface. She authorizes a two-week test of one narrower workflow, requiring a 20 percent improvement with no increase in support requests. A colleague from another product group challenges the measurement plan and asks whether integration, rather than interface design, is the stronger route.

The test misses the threshold. Maya stops the redesign, documents customer observations and retains the reusable technical components. She then presents a restart case for an integration-led concept with a new scope, budget and evidence standard. The original route ends; the learning survives.

This is not failure disguised as process, nor is it impulsive abandonment. A disciplined stop follows declared evidence and protects resources for a better route. Persistence remains valuable when evidence supports the next commitment—not merely when the organization remembers the last one.

Sources