Over five weeks I posted a short series for the people who carry the can on a major programme – sponsors, senior responsible owners, board members, accounting officers. This is the whole argument, in one place.
The room I keep walking into
In thirty years of being called in to rescue programmes, I have never once found the wreck where the client was looking.
Every recovery I have led has begun in more or less the same room. Serious people, around a serious table, surrounded by reporting that told them everything was broadly fine until, very suddenly, it wasn’t. The schedule was green. The budget was on plan. The risk register was immaculate, forty-one entries all dutifully rated and mitigated. And the programme was already gone – it simply had not been declared yet.
I want to set out, in one place, what those rooms have in common. The pattern is consistent enough to be useful, and almost all of it comes back to a single seat at the table – the one that, by definition, only the owner can fill.
1. Failure does not live in the machinery
Programmes do not fail in the schedules, the cost models or the risk logs. Those are instruments. They report what someone chose to put on them, and they are only ever as honest as the culture feeding them.
Failure lives in the layer above the machinery. It lives in sponsorship that went quiet the moment things got hard. In bad news that took four months to travel three floors. In accountability that blurred the instant two directors disagreed. In a reporting culture where “amber” had quietly come to mean “please don’t ask.”
By the time the numbers turn red, the failure is already history. The red is a death certificate, not a diagnosis – it records something that happened weeks or months earlier, in conversations that never made it onto a slide.
The red is a death certificate, not a diagnosis.
And here is the part that should give any owner pause: every one of those programmes had excellent reporting. Polished, timely, reassuring. That is not a contradiction. On a programme in trouble, the quality of the reporting and the health of the programme are two entirely different things – and mistaking one for the other is how serious people end up surprised by failures that were, in hindsight, visible for a long time.
2. The one role you cannot outsource
When an organisation takes on a programme larger than anything it has attempted before, the instinct is a sensible one: bring in people who have done it. Procure the capability. Buy the expertise. And you can. You can buy design, delivery, programme management, controls, assurance – the lot.
There is exactly one thing you cannot buy: the capability to be an intelligent client of what you have just bought.
The intelligent client is the function that knows what good looks like, asks the question behind the answer, and can tell the difference between a problem being managed and a problem being hidden. It sits on your side of the table. And it cannot be supplied by the people on the other side, because they have an interest in the answer.
This is the most consistently underestimated role in major delivery. Owners spend months evaluating partners and minutes thinking about their own capacity to hold those partners to account. Then the programme drifts, and everyone is surprised – again.
The day you discover you needed an intelligent client is the day it’s already too late to build one.
It does not take a large team. It takes the right one, reporting to the sponsor and equipped to interrogate. But it takes a deliberate decision that the function exists, made early – because the day you discover you needed it is the day it is already too late to build it.
3. When the instruments stopped telling the truth
For most of my career, polish was a fair proxy for thought. A clear, well-argued, well-presented report usually meant someone had done the work to understand the thing they were reporting on. That proxy is now broken.
A competent operator with an AI tool can today produce a confident, articulate, immaculately structured account of a programme they neither understand nor control. The presentation layer has been cut clean away from the thinking layer beneath it.
For the person receiving the report, this is a quiet crisis. The signals you have trusted your whole career to sense trouble – the hesitation, the roughness at the edges, the slightly-too-defensive answer – are being sanded smooth. The report that would once have looked rough enough to question now arrives looking finished.
The answer is not to ban the tools. They are useful and they are here to stay. The answer is to raise your own game as a reader. The skill that now matters is interrogation: not “is this report good?” but “what would have to be true for this to be accurate – and can you show me?” Not the polish, the evidence beneath it. AI has made it cheaper than ever to look in control, which makes it more valuable than ever to be the person in the room who can tell the difference.
4. The decision that sets the trap
Long before any of this, there is a decision that quietly determines whether the owner’s seat is occupied at all: build, buy, or hybrid. Do we build the capability to run this ourselves, buy it from a delivery partner, or do some blend of both?
Watch how the decision is actually made and you will usually see it start from the organisation’s comfort zone rather than the programme’s needs. “We’ve always done things in-house” points to build. “We haven’t got the people, get a partner in” points to buy. And hybrid, too often, is not a decision at all – it is what you end up with when nobody made a clean one.
None of those answers the real question, which is this: what does this specific programme, at this scale and with this risk profile, actually require – and what is this organisation honestly capable of governing? Build without the maturity to deliver gives you an expensive in-house learning exercise on a programme that could not afford one. Buy without the intelligent client capability hands over control and calls it delegation. Hybrid done well is genuinely powerful; hybrid done by default – a bit of everything, owned by no one – is the worst of the three.
The decision deserves a structured, honest readiness assessment before anyone commits, not a reflex dressed up as strategy. Getting it right costs a few weeks of disciplined thinking. Getting it wrong costs the programme – and very often, getting it wrong is precisely the moment the owner’s seat is quietly left empty.
5. The good news: you can know in days, not months
So far this has been an uncomfortable read, and deliberately so. But there is one genuinely encouraging change, and it matters most to owners who already suspect they are in trouble.
There is a particular limbo I have walked into many times. Senior people around a table all privately believe the programme is in difficulty. None of them can say so with authority, because no one has closed the gap between suspicion and evidence. So the programme runs on, money goes out of the door, reputations grow more entangled by the week, and the honest conversation keeps being deferred – because honesty without evidence feels like an accusation.
Here is what has changed. Finding the gap – reading across thousands of pages of contracts, reports, minutes and correspondence to surface where the story stops being consistent – used to take weeks of skilled, expensive forensic effort. It now takes hours.
It now takes hours. Which removes the last excuse for waiting.
That does not mean a machine recovers your programme. It does not. Diagnosing why it went wrong, deciding what to do about it, and executing the turnaround are human work, and they always will be. But the thing that used to keep failing programmes alive for months – the sheer cost and time of proving what everyone already suspected – has largely gone. Which removes the last excuse for waiting.
Taking the seat
Five symptoms, one cause. Programmes fail in the human layer, not the machinery. The one capability you cannot buy is the capability to govern what you have bought. The reports you rely on no longer prove what they appear to prove. The build-buy-hybrid decision is too often made from comfort rather than need. And when it is already slipping, the gap between knowing and proving is now small enough to close in days.
Behind all five sits the same seat. The owner’s real job on a major programme is not to deliver it – that is what the delivery capability is for. The owner’s job is to be the intelligent client of it: to keep sponsorship honest, to read past the polish, to make the resourcing call from need rather than habit, and to close the gap between suspicion and evidence before it closes on them.
None of this requires heroics or a standing army. It requires deciding, deliberately and early, that the seat is occupied by someone equipped to hold it. Almost everything I have ever been called in to recover traces back to that decision being made late, or not at all.