BOLDGROUPTHAILAND Back to overview
All insights
FIELD NOTE 03 / PROCESS TRANSFORMATION

When your stage-gate process becomes the constraint

Stage-gate brought discipline to product development. In faster markets the same structure filters out the work that matters most — and nothing in it measures the delay it introduces.

01IN BRIEF
  • Stage-gate exists to reduce the cost of being wrong. That trade reverses when being slow becomes the more expensive error, and nothing in the structure notices the reversal.
  • Gates require evidence, and novel work cannot supply the kind that incremental work can. One evidence bar for both performs selection rather than risk management.
  • Three checks establish whether this is your constraint, all answerable from records you already hold.
  • The answer is rarely to remove gates. Match the evidence bar to the size of the commitment, and decide when the evidence is ready rather than when the calendar allows.
02THE ARGUMENT

A stage-gate process that is working correctly will still hold your most valuable opportunities at the gate longest, and kill them first. That is not a malfunction. It is what the design was built to do, applied to work it was not built for.

What stage-gate optimises for

A stage-gate process is a sequence of development stages separated by decision points, at which a project must present evidence to receive further funding. The method was developed by Robert G. Cooper, and Stage-Gate is his registered approach rather than a generic term.

Its purpose is to reduce the cost of being wrong. Each gate is an opportunity to stop something before it consumes the next stage's budget. In a market where being wrong is expensive and being slow is survivable, that trade is correct.

The trade reverses when being slow becomes the more expensive error. The structure does not notice the reversal, because nothing in it measures the cost of the delay it introduces.

Three ways it constrains the work that matters

It selects for the predictable. Gates require evidence, and incremental work can supply evidence more readily than novel work — a line extension has comparable products to reference, while a genuinely new proposition does not. Applying one evidence bar to both does not compare them fairly; it advances whichever can produce familiar-looking proof.

It converts decisions into events. When funding is released only at scheduled gates, a project ready in week three waits until week nine. The elapsed time is not spent developing anything. It is spent waiting for the calendar, and it appears in no report as a cost.

It measures compliance with the process. Gate reviews assess whether required documentation is complete. That is a legitimate control, but a team can pass every gate on schedule while the underlying proposition weakens, because the review examines the paperwork rather than the merit.

Together these produce a recognisable condition: innovation theatre. The activity is real, the reviews happen on schedule, the documentation is complete, and the business result does not move. It is not cynicism on anyone's part — it is what a process rewards when it can only assess whether its own steps were followed.

How to tell whether this is your constraint

Three checks, all answerable from records you already hold.

Measure the waiting, not the working. For recent projects, separate elapsed time into time spent doing work and time spent awaiting a decision. Where the second is a large share, the constraint is the decision structure, not the delivery capacity.

Examine what has been stopped. Review the projects killed at gates over two years. If the pattern is that the most novel work stopped earliest, the evidence bar is doing selection rather than risk management.

Ask what the gate would have to see. For a current proposal, ask what evidence would satisfy the next gate. If the answer is evidence that could only exist after the commitment the gate is deciding, the process cannot approve that class of work at all.

Transforming legacy processes without losing the control

The answer is rarely to abolish gates. Organizations that remove them typically rediscover why they existed. Three adjustments retain the control and remove the constraint.

Match the evidence bar to the commitment. A small, bounded test does not need the evidence required for a full development commitment. Set the bar by what is being risked, not by the stage number.

Decide when the work is ready, not when the calendar allows. Where a decision can be made on evidence already available, the waiting adds cost without adding information.

Ask the falsifying question. Replace “prove this will work” with “what would have to be true for this to work, and what is the cheapest way to find out?” The first question can only be answered by work that has already succeeded. The second is answerable in advance, and it makes experimentation explicit and bounded rather than risky and informal.

The sequence is usually the opportunity

The largest gains in process redesign are more often in the order of the work than in the speed of any step.

A turbine maintenance engagement is the clearest instance. The available time turned out to be in the order the stages ran in rather than inside any one of them.

That is the same question a stage-gate review does not ask. Gates assess whether a stage is complete. They do not assess whether the stage needed to wait for the previous one.

The same failure appears in physical inspection, where it is easier to see. A plant checking quality at the end of its line was asking whether the part was finished, and never whether the material could still be recovered — so every reject was found after the one step that made it unrecoverable. The check was competent and it was in the wrong place, which is what a gate cannot detect about itself.

Borrowing beats inventing

The lowest-risk innovations are frequently adaptations of mechanisms already working in another industry. The problem is rarely unprecedented in general; it is unprecedented in your sector.

Working across industries is what makes those patterns visible. A surface-protection method developed for one manufacturing context, a scheduling approach from another — transferred deliberately, these carry evidence that a novel invention cannot, which also makes them easier to fund through exactly the gates described above.

Where this does not apply

Where regulatory approval, safety certification, or contractual obligation sets the sequence, the gates are not inherited and cannot be redesigned away. The distinction worth testing is which of your gates are external requirements and which are internal convention that has hardened into a requirement. In most organizations the second group is larger than assumed.

03SOURCES
  1. The Stage-Gate model: an overview

    Robert G. Cooper, Stage-Gate International · Developed from the 1980s

    Attribution rather than a cited finding. Stage-Gate is Cooper's named approach, not a generic term, and the critique here is of how it is commonly applied rather than of the original design.

04COMMON QUESTIONS
Should we remove our stage-gate process?
Usually not. The control it provides is real. The adjustments that matter are matching the evidence bar to the size of the commitment, and allowing decisions when evidence is ready rather than when the calendar permits.
How do we know the process is filtering out good opportunities?
Review what has been stopped at gates over the last two years. If the most novel proposals consistently stopped earliest, the evidence requirement is performing selection rather than managing risk.
What replaces a gate for genuinely novel work?
A bounded test with a defined question, a cost limit, and an agreed decision rule set before it starts. This is still governance — it is simply governance calibrated to what is actually being risked.
Is this only relevant to product development?
No. The same structure appears in capital approval, IT project governance, and internal service design. Anywhere a decision point requires evidence that only later work can produce, the same constraint operates.
05WHERE THIS LEADS
WHAT SHOULD CHANGE?

Bring the outcome, the constraint, and the accountable decision into the same conversation.

Discuss the outcome that matters