20-minute diagnostic
EN
Perspectives / Decisions & capital

A pilot without an exit rule is not a pilot

A founder and an operations leader review a limited pilot proposal in a boardroom

A six-week pilot is approved for one team. Six months later, the supplier is still being paid, two more teams have joined, and no one can say whether the test worked.

The problem is not that the pilot failed. It is that nobody agreed what would happen when it ended.

Decide what the pilot is meant to answer

Before choosing a tool or process, state the decision the pilot is supposed to inform. “Try the new scheduling platform” describes an activity. “Find out whether it can reduce handoff delays without creating more work for dispatch” gives the team something to evaluate.

That distinction matters. A pilot can look busy and still tell you very little. People may attend training, enter data and complete weekly check-ins without answering the question that justified the expense.

Write the question in one sentence. If the group cannot agree on what a useful answer would look like, the test is not ready to start.

Put the limits in writing

A useful pilot has a boundary. Record:

  • one accountable owner who can make the recommendation;
  • the team, location or process included in the test;
  • the time and spending limit;
  • what evidence will be reviewed;
  • the date when a decision will be made.

The point is not to turn a small experiment into a bureaucracy. A short note is often enough. It gives everyone the same understanding of what they have agreed to do and what they have not agreed to do.

Be precise about authority, too. Can the pilot owner adjust the process? Can they add a user or extend the schedule? Which changes require the executive sponsor's approval? If every adjustment comes back to the CEO, the test may recreate the same approval bottleneck it was meant to address.

Two male colleagues review early pilot results together on a workshop floor

Choose evidence that fits the decision

Decide in advance what would support expansion, revision or a stop. Look at more than one convenient number. A new process might be faster but produce more errors. A tool might receive positive feedback from its users while adding work for the team that supports it.

Use a small set of observations tied to the original question. Include the people who will carry the work after a wider rollout. Record important changes in conditions as well; a test run during an unusually quiet month may not tell you how the process performs at year-end.

The evidence does not have to be perfect. It does have to be relevant to the decision you will make next.

Book the decision before the pilot begins

Put the review meeting on the calendar at the start. Ask the owner to bring the original question, the evidence gathered, what changed during the test and a recommendation. At the meeting, choose one of three actions: stop, adjust and test again, or expand within a defined scope.

If more time is needed, name what remains unknown and set a new end date. “We need another month” is not a decision unless the team knows what it expects to learn during that month.

Stopping does not make a well-designed pilot a waste. It can prevent a much larger commitment to a process that does not fit. The test has done its job when it gives the decision owner a sound basis for the next choice.

Notice when the terms keep changing

Sometimes the business case is clear and the owner still avoids the decision. The review date moves, the success criteria are rewritten, or the pilot continues because ending it would require a difficult conversation with the sponsor or supplier.

That is a different problem from incomplete evidence. It may help to ask what the person believes will happen if they stop, and what commitment they think they are making by deciding. The answer can reveal whether the hesitation belongs to the business plan, the relationship or the executive's own role in the decision.

If a decision keeps reopening after the evidence is clear, that pattern may be worth examining directly. In SEAM, the work can begin with one situation like this and focus on a practical next step. Request a free twenty-minute diagnostic to discuss the problem.

Request a free diagnostic