(And why large, complex programs still run this way)

The moment everything changed

“How confident are we in the revised delivery date?” the client asked.

“60%,” I said.

The room went quiet. Not polite quiet – more shock than anything else.

The client team looked back at me, trying to work out whether this was candour or career suicide. Under the table, my CEO’s leg had started bouncing – he knew what this could trigger. No supplier had ever put a confidence rating that low on the table.

After the initial shock – and a predictable burst of outrage – the conversation shifted. Once the posturing subsided, I was able to show when and how system reliability – and confidence – would climb, and over what time period.

That reaction almost always comes from the commercial side. The engineers already know how the system behaves.

We were controlling it.

It wasn’t stuck at 60%.

With the right changes, it would move into the nineties – but never perfect. A complex system. We would be confident in the main paths through it.

My opposite number wasn’t completely surprised – I’d kept him close to where we were and what it meant for delivery, even if the percentage itself landed hard. He didn’t know how his stakeholders would react.

We agreed the program needed a cold shower – after too many self-inflicted false dawns on both sides.

I wasn’t the first to sit in that seat. Three predecessors had come and gone before me – each losing the confidence of the client.

They knew how to operate in that environment. And pushing back wasn’t part of it.

But the number wasn’t the point.

It exposed something most programs never test properly.

Five years of activity – and nowhere near enough progress to meet the targets that had already been committed

The program had all the hallmarks of control. Multiple suppliers. Regular reviews in central London. Detailed reporting. Plenty of activity.

And yet not single number that reflected reality.

Each forecast was built to survive until the next review – and by then something would have shifted.

A delay elsewhere. A dependency slipping. Occasionally even the client’s own issues surfacing just in time to blur the picture again.

It wasn’t unique to that program. You will find it in most large, complex programs.

This wasn’t a visibility problem

People could see what was going on. The engineers knew where things were breaking. Suppliers could see the collisions at the boundaries. Parts of the client organisation had stopped believing the reports altogether.

The problem was a lack of permission to make binding decisions.

Everything revolved around a single question:

“Are we on track?”

It sounds reasonable. It isn’t.

It invites reassurance, encourages interpretation, and produces answers that are technically defensible and practically useless.

It keeps the program moving, but it doesn’t move it forward.

The question that actually matters

The shift came when that question stopped mattering and was replaced by something more useful:

“What would it take to make this work?”

That question changes the rules.

The moment you ask it properly, the program stops being a narrative and becomes a problem to be solved.

You are forced into the arithmetic – how much work there is, how long it takes, what depends on what, and whether any of it holds together under real conditions.

At that point, the story falls away.

You are left with facts.

Why this almost never happens

Because the system doesn’t reward frankness and honesty.

Suppliers learn quickly that optimism is commercially safer than openness.

Prime contractors know that admitting loss of control invites commercial scrutiny they may not survive.

Program boards say they want transparency – but tend to reward continuity over disruption.

But that is how it behaves.

When it becomes impossible to judge honestly

Each decision made sense in isolation.

Together, they created something else entirely – a version of reality just credible enough to continue, but not robust enough to deliver.

That is how you get years of progress reports that no one believes.

The program doesn’t fail outright.

It becomes impossible to judge honestly.

And the longer that persists, the harder it becomes to stop, because too many decisions, reputations, and commercial positions are tied to the assumption that it will eventually come good.

By the time I arrived, that had become normal behaviour.

The client no longer believed the reports, but the process continued. The suppliers no longer believed each other, but they kept presenting. My own organisation – the prime contractor – had become adept at saying “almost there” with a straight face.

The reviews had taken on a familiar rhythm. A stage show. Plenty of challenge, plenty of theatre, but very little that changed the outcome.

Nothing changed because nothing forced it to.

What actually changed things

So we stopped – stepped outside the program for long enough to test whether it actually worked.

We asked for six weeks with no new commitments, no new dates, and no attempt to defend the existing position – just enough time to establish what was actually true.

When you strip a program back like that, the noise falls away quickly.

What remains is arithmetic.

Once the arithmetic is visible, it’s very hard to ignore.

·  How much work is there?

·  How long does it take?

·  Where are the dependencies?

·  Do they hold?

·  What happens when the whole system runs under load?

Those questions have answers – and once you have them, the range of possible outcomes narrows very quickly.

In this case, it quickly narrowed to one.

Which is how I ended up back in that room, saying “60%.”

What happened next

What followed was not the reaction people expected.

There was no backlash, no performative outrage, and no attempt to negotiate the number. What replaced the silence was recognition.

For the first time in years, the client felt they were being told the truth – not a polished version, not a defensible version, but something grounded in facts they could test and understand.

That changed the dynamic immediately.

Once the conversation moved to “What would it take to make this work?”, the program had somewhere to go.

The six-week pause made sense. The plan that followed was not optimistic.

It was workable. It was realistic.

Both sides believed in it – everyone understood how it had been constructed.

We put an end to guessing – and the inter-company political games that came with it.

We did what we said we would do – with a few weeks to spare.

What this really was

The number didn’t change the program. It exposed it.

It forced a simple question that had never really been asked:

Will this actually work – when everything comes together?

We didn’t have a name for it at the time.

But that was, in effect, an Independent Program Review – a deliberate step outside the program to test whether it actually works.

The only question that matters

Most programs never get close to answering that question.

The system is set up so that it’s never asked.

The program becomes impossible to call out – even when people can see it isn’t going to work.

By then, the dates are committed, the positions are taken, and too much depends on it continuing.

So it continues.

Until reality forces the answer.

And by then, it’s usually too late – to avoid major delays, eye-watering overspends, and settling for less than you expected.

About the author

David Hilliard is founder of Mentor, specialists in strategic program execution.

You can call him on 0118 359 2444 or email david.hilliard@mentoreurope.com.