The Gate That Could Not Say No
You cannot moralise your way out of an incentive structure.
Executive Summary
Somewhere between the discipline the last decade taught us and the assurance this one now demands, the stage gate quietly stopped being a decision and became a performance. It was conceived as a fork in the road — a disciplined moment at which a programme could be continued, redirected, re-resourced, or stopped cheaply before good money followed the first mistake. In a great many organisations it now operates as a milestone dressed as a fork: a scheduled ceremony that consumes weeks of preparation, produces a pack assembled to be approved rather than interrogated, and almost never ends in the one word it exists to make possible — no.
This essay asks why that drift happened and why it has proved so stubborn. The causes are not incompetence or idleness; they are structural. Incentives quietly punish the person who stops their own programme. Compliance regimes have rewritten the gate as an assurance artefact. The calendar, rather than the evidence, now decides when a gate occurs. And accountability spread evenly across a board becomes accountability held by no one. The argument here is not that gates should be abolished — the years since the dot-com collapse have taught us all too plainly what ungoverned spend costs. It is that a gate which cannot say no is not governance at all. It is the theatre of governance, and from the inside the two have become remarkably hard to tell apart.
The Meeting Everyone Has Already Left
Consider a scene familiar to anyone who has sat on a programme board. A major change programme approaches its third gate. For three weeks, a team of four has been assembling the review pack: a revised business case, an updated benefits map, a risk register now several hundred lines long, a plan on a single unreadable page, and a deck that exists mainly to summarise the pack that no one will finish reading. The document runs past a hundred and thirty pages. It is circulated, in the customary courtesy, forty-eight hours before the meeting — which is to say, too late to be absorbed and too early to be an emergency.
The board convenes for ninety minutes. The senior responsible owner speaks well. The programme is behind, but there is a recovery plan; costs have risen, but there is a reforecast; two of the risks have gone from amber to red, but mitigations are “in hand.” A director asks a question about a dependency; it is noted as an action. At the end, the chair looks around the table and asks whether the board is content to proceed to the next stage. It is. The gate is passed. The RAG status remains, as it has for three consecutive reports, a resolute amber.
What has happened here? A great deal of work, and no decision. The word proceed was never in doubt — not because the programme deserved to proceed, but because everyone in the room understood, without it being said, that stopping was not on the menu. The people who prepared the pack could not recommend their own cancellation. The board members lacked the independent grip on the detail to overrule them. And the gate had arrived because the date had arrived, not because a genuine choice had ripened. The ceremony was real. The decision was ornamental.
What the Gate Was Meant to Be
It is worth remembering that the stage gate was a genuinely good idea, and in its origins a radical one. Robert Cooper’s work on Stage-Gate in new product development gave a generation of managers a simple, powerful proposition: that a project is a series of increasingly expensive bets, and that the rational moment to reconsider a bet is before you place the next, larger one. The gate was the place where you stopped, looked honestly at what the last stage had taught you, and decided — on evidence you did not have at the outset — whether to continue.
Public sector practice absorbed the same logic. The Gateway process introduced by the Office of Government Commerce earlier this decade built its reviews around exactly this idea: independent assessment at defined points, with the explicit possibility of a verdict that halts a programme until it is fixed. PRINCE2’s stage boundaries carry the same intent in miniature — the end-stage assessment exists so that continuation is a choice the board actively makes, not a default it passively allows. In each case the design principle is identical, and it is worth stating plainly.
A gate is not a milestone. A milestone marks that something has happened; a gate decides whether the next thing should. The whole value of a gate lies in the reality of the alternative outcome — that it can genuinely stop.
This is the point on which everything turns. A gate earns its cost only if the answer is capable of being no. Remove that possibility — make continuation a foregone conclusion — and you have kept all of the gate’s expense and discarded all of its purpose. You have a milestone with a review pack attached.
The Slow Substitution of Assurance for Insight
The drift from decision to ceremony is rarely a single decision anyone made. It is a slow substitution of one thing for another that looks almost identical from a distance: the replacement of insight with assurance.
The distinction matters more than it first appears. Insight is what a gate produces when it asks a question whose answer could change the plan — when the review genuinely does not know, going in, whether the programme should continue, and comes out knowing more than it did. Assurance is what a gate produces when it sets out to confirm that the programme is on track and, reliably, confirms it. Both generate documents. Both occupy the calendar. Both feel like diligence. Only one of them is worth the price.
Watch how the vocabulary itself drifts. A review that once asked “should we be doing this at all?” comes to ask “are we doing this in accordance with the plan?” The second question is easier, safer, and almost always answerable in the affirmative, because the plan is ours and we wrote it to be followed. The first question is uncomfortable, because its honest answer sometimes implicates the people in the room. So the harder question quietly retires, and the review reorganises itself around the one that can be passed.
“We came to mistake the weight of the pack for the rigour of the thinking, when in truth a heavier pack is more often a sign that no one trusted a single clear judgement to carry the decision.”
There is a particular tell for this substitution, and once you have seen it you cannot stop seeing it: the review pack grows, year on year, while the decisions it informs get no better. Thoroughness becomes a proxy for rigour. The three-hundred-line risk register feels more responsible than the ten risks that actually matter, precisely because it is exhausting — and exhaustion is easily confused with diligence. But a register no one can hold in their head is not a management tool; it is an insurance policy, filed so that someone, later, can demonstrate that the risk was “on the register” all along.
Why It Persists: The Structural Forces
If gate theatre were merely a mistake, it would correct itself; intelligent people would notice the waste and stop. That it does not correct itself tells us the forces sustaining it are structural rather than intellectual. Four are worth naming.
- The asymmetry of incentives. Consider what actually happens to the senior responsible owner who stands up at a gate and recommends that their own programme be stopped. They do not receive a bonus for capital preserved. They lose the programme, the team, the standing, and often the budget line that justified their role. Meanwhile the owner who presses on — even into difficulty — retains all of these and can, for a long time, describe the difficulty as a challenge being managed. We have built a system in which continuing a failing programme is personally rational and stopping it is personally costly, and then we express surprise that so few are stopped.
- The compliance overlay. Over the last few years, governance has acquired a second master. The tightening of financial control and audit expectations — the documentation culture that followed the corporate accounting scandals at the start of the decade — has quietly rewritten what a gate is for. It has become, in part, an evidence-generating exercise: proof, retained for the auditor and the accounting officer, that due process was followed. This is not worthless, but it pulls hard in the wrong direction. A gate optimised to demonstrate that process was followed will always favour the complete pack over the awkward question, because completeness is auditable and judgement is not.
- The tyranny of the calendar. Because gates are scheduled at the outset, they tend to occur when their date arrives rather than when a real decision is due. A gate held because the plan says week thirty-six will find whatever is there in week thirty-six — which is usually a programme mid-stride, neither clearly succeeding nor clearly failing, and therefore easy to wave through. The decision that would have been genuinely live three weeks earlier, or genuinely urgent three weeks later, is not the decision on the table. The date manufactures a non-event and then dignifies it with a meeting.
- The diffusion of accountability. A gate decision taken by a board is a decision taken, in the fullest sense, by no one. Each member defers slightly to the others’ apparent confidence; the specialist defers to the generalist on strategy and the generalist to the specialist on detail; and the collective nods through what no single member, questioned alone in a corridor afterwards, would have signed their name to. A board is a device for sharing comfort, and shared comfort is the enemy of the uncomfortable no.
Notice that none of these forces is a failure of character. Each is an entirely rational response to the situation the individual finds themselves in. That is precisely why exhortation — “boards must be braver,” “owners must be honest about failure” — changes so little. You cannot moralise your way out of an incentive structure.
The Case for the Defence
It would be too easy, and finally dishonest, to leave the argument there. The strongest case for gates is not the ceremonial version this essay has been dissecting; it is a serious one, and it deserves to be met at its strongest.
The defence runs like this. We do not have to imagine the alternative to disciplined gating, because we lived through it. The waste of the late-nineties boom — the projects funded on enthusiasm, scaled before they worked, and abandoned only when the money ran out — was the waste of a world without effective gates. Discipline is not free, and neither is its absence; the review pack that looks like bureaucracy from the inside is often the only thing standing between an organisation and a portfolio of quietly failing commitments that no one is willing to confront. Gates have killed bad programmes. The mere existence of a scheduled point of scrutiny changes how teams behave in the stage before it, because they know they will have to account for themselves. And an imperfect discipline that occasionally catches a disaster may be worth a great deal more than the elegant judgement of individuals who, left ungated, reliably fall in love with their own plans.
All of this is true, and none of it should be surrendered. But look closely at what the defence actually justifies. It justifies scrutiny, discipline, and the reality of stopping. It does not justify the hundred-and-thirty-page pack, the amber that never changes, or the gate whose answer was known before the meeting. The defence is an argument for real gates. It is not an argument for the ones we have. The failure this essay describes is not the failure of gating; it is the failure of gates that have kept the cost of scrutiny while quietly discarding its substance. To defend the ceremony by invoking the value of the principle is to change the subject.
What a Gate That Adds Insight Looks Like
If the disease is the substitution of assurance for insight, the cure is not another layer of process. It is the recovery of the gate’s original nerve. This essay is a reflection rather than a manual, but the shape of a healthier practice is not hard to discern, and it tends to share a few features.
| The gate that adds delay | The gate that adds insight |
|---|---|
| Held because the date arrived | Held because a genuine decision has ripened |
| Asks: are we on plan? | Asks: what would tell us we are wrong, and have we looked? |
| Pack assembled to be approved | Evidence assembled to be interrogated |
| “No” is unthinkable and unspoken | “Stop” is a respectable, survivable outcome |
| Thoroughness stands in for judgement | A few decisive questions carry the decision |
| Owned by a board, and so by no one | Owned by someone who can be named |
The first move is to have fewer gates and mean them. A programme with three real decision points, each capable of stopping it, is governed. A programme with nine ceremonial ones is merely audited. Every gate that cannot realistically say no should be demoted to what it actually is — a progress checkpoint — and stripped of the expensive apparatus of decision it does not deserve.
The second is to change the question. A gate should be built around the inquiry that could genuinely alter the plan: not “is the programme on track?” but “what would we expect to see if this were failing, and have we honestly looked for it?” The most valuable review I have sat through turned on a single such question, asked quietly by someone with no stake in the answer. It did not stop the programme, but it redirected a substantial commitment away from a component the team had privately doubted for months and never found the licence to challenge. That is what insight costs — one honest question, well placed — and it is a fraction of the price of the pack.
The third, and hardest, is to make stop survivable. As long as ending a programme ends a career, no amount of process design will produce honest gates, because the people best placed to recommend stopping are precisely the people it costs most. An organisation that wants real gates has to make the disciplined termination of a programme a mark of judgement rather than of failure — to acknowledge, occasionally and visibly, the owner who handed money back. Until stopping is safe, every gate is a gate that cannot say no, however thick its pack.
The Mirror
In the end, gate theatre is less a governance defect than a mirror. It shows us, with some precision, an organisation’s real relationship with uncertainty. We build gates because we want the feeling of control over things we do not fully control. When the gate works, it gives us something better than the feeling: it gives us a genuine chance to be wrong cheaply, early, and on purpose. When it decays into theatre, it gives us the feeling alone — the comfort of a decision without the discomfort of making one.
The gap between transformation intent and transformation reality is, so often, exactly this gap: the distance between the governance we designed and the governance we can bear to practise. We designed a fork in the road and we practise a milestone, because the fork asks us to contemplate our own reversal and the milestone merely asks us to keep walking. Closing that gap does not require a new methodology. It requires the far rarer thing — the willingness to build a gate that can genuinely stop us, and then to stand at it and, when the evidence says so, actually stop.