Give stage gates teeth by shrinking what they decide

What the paper develops

Ask a portfolio board when its last review changed how much money, how many people, or how much scope a project received. If no one can name a case, the review may be tracking progress rather than governing the investment. A board with well-chosen projects may rarely stop anything, so the test is not a count of stops. It is whether the review could have changed what the project received, and whether it ever did. A gate has teeth when it can change what a project receives: its money, its people, or its scope.

Reviews tend to fail in one of two ways. Some lack the authority to act: leaders ask for another slide, and the project keeps its funding, staff, and scope. Others vote yes or no on the whole project, once. A yes given early is hard to take back, because contracts are signed, people are hired, and other programs plan around the result. Research on escalation of commitment, the tendency to keep backing a failing course, also links personal responsibility with greater escalation, although the average effect of money already spent varied widely across studies.

A gate has teeth only over a choice that is still open, so the fix starts with where the gate sits. A stage drawn around a phase of work, such as design or build, can end after the point where reversing a step has become expensive. A gate should sit before the first step that is hard to reverse and decide that step: the next commitment, meaning a step that is hard to reverse because it spends a sum the board would not want to lose, pulls scarce people from other work, signs a contract, or lets other programs plan around the result.

Decide the next commitment, then give the board what it needs

To find the next commitment, list the steps the project will take, in order, and mark each as cheap or costly to reverse. A short test with no contract is cheap. A multi-year contract or a shared platform is costly. Approve the cheap work up to the next costly step, and review that step on its own with stronger evidence in front of it.

Three controls make the decision real. The board has written authority to continue, redirect, narrow, pause, or stop, with the effect of each outcome defined before the meeting. The evidence fits the size of the commitment and sits on a one-page evidence map. Someone other than the sponsor owns the key facts that could change the decision. Each approval then carries an end date, the three or four facts it rests on, and a trigger that sends the decision back to the board.

Test one gate before changing the whole process

Choose one mid-sized decision with real stakes. Write the next commitment in one sentence, and list what each of the five outcomes would do to money, people, scope, and dependencies. Record the decision, the assumptions, the triggers, the end date, and any dissent in the meeting, then check that money and people moved as directed.

After two or three reviews, compare what the board committed at each step with what was requested. If nothing ever changes and the board could not have changed it, fix the authority and the size of the commitment before adding templates. A gate earns its place when evidence can change the next commitment while the organization still has options.