What Scale Reveals

Early in my time on one of the largest developments in the world, I took two engineers up to see the excavation. I'd worked with both of them in Australia years before, on projects that felt substantial at the time. We stood at the edge and looked down.

None of us said much at first. The space was enormous, the machines in it looked like toys, and you could see just how challenging a project this would be. Both had spent their careers building difficult projects. Neither had seen anything like it.

Then one of them asked when it was meant to be finished. I told him the date, which was less than three years away.

All three of us laughed.

We laughed at the same moment, without conferring, and that is the part I keep coming back to. Nobody did any arithmetic. Nobody asked about sequencing, procurement, or how many shifts you'd need to run. Three experienced people looked into a hole in the ground, heard a date, and answered with their bodies before their heads caught up.

What we were laughing at was the distance between two things. The ambition was extraordinary. Unfortunately, the plan attached to it was not grounded in reality.

It took me considerably longer to work out why it hadn't actually been funny.

Being fair about this matters, because the vision itself was not the problem. It was bold, it was coherent, and it rested on drivers that made real sense. Most organisations never plant a flag like that, and the ones that don't rarely build anything worth talking about. The people were capable too — I'd put many of them against anyone I've worked with in thirty years. Nothing here is a failure of individual competence.

The problem was that the ambition was serious and the plan was not, and at that size you cannot hide the difference.

I have laughed at an impossible date on a suburban subdivision. Everybody in this industry has. But that is a different laugh — the laugh of people who expect the date to be theatre and have already priced it in. Nobody grieves a subdivision. What made this one different is that the vision deserved better. It was the kind of ambition that comes along a few times in a career, and it warranted a plan as serious as the thinking behind it. A small ambition failed on a poor plan is a nuisance. A large one is a waste.

That is what scale reveals. Not new problems. The same problems, at a size where you can no longer look away from them.

Here's the part that should trouble us more than it does. I used to hold up aeronautics as the industry that had solved what we hadn't. Then I looked properly at the A380.

The French plants designed in CATIA software. The German plants used Computervision. Design files didn't transfer cleanly, and when the wiring went into the first aircraft — around 100,000 wires, 530 kilometres of cable — the cables were too short. They had been routed through a fuselage that existed in two incompatible versions of itself. The programme ran more than two years late and cost roughly €5.5 billion.

That is not a manufacturing failure. It is an interface failure between parties using incompatible systems, each protecting its own way of working, inside a single company. It is our disease, in their industry.

But I was wrong about the example, not about the industry, and the distinction is the whole point.

Two years and €5.5 billion was the worst programme failure in modern commercial aviation. It was treated as one. It took executives with it, was picked apart in public, and the business was restructured around it. Now hold that against our own standards. A two-year overrun on a major development barely registers as news. We'd call it a challenging program and move on. Their catastrophe is our ordinary Tuesday.

And they recovered. Airbus went on to build 251 of them, at rate, each cheaper and faster to build than the last. The failure was a phase in the programme, not a condition of the industry. Aeronautics plans, sequences and delivers better than we do. The proof isn't that they never fail. It's that when they do, the failure is bounded, examined, and closed out.

Ours isn't a phase, because we don't keep the organisation that could learn from it. Airbus was still Airbus the following year, and the year after that. The entity that delivers a major development is assembled for the job and taken apart at the end of it, and whatever it learned leaves in the heads of people already moving to the next one.

The numbers are what you'd expect. Between 1939 and 2020, American car manufacturing went from about 4.8 vehicles per employee per year to roughly twenty-five. Over the same period, residential construction productivity has gone approximately nowhere — close to where it sat in the 1930s. The economists who examined that blame land-use regulation and the small firms it produces, and they may be right. I'd only add that an industry starting every job from zero was never going to build a compounding curve.

And there's a cost beyond the numbers. When something genuinely bold fails on an unserious plan, the lesson the world takes away is don't be that ambitious, never plan it properly next time. The boldness gets discredited rather than the execution, which makes the next bold thing harder to fund. Meanwhile the people who believed in it, who moved countries and gave it years, absorb the disappointment personally. I was one of them. I'd do it again. But I'd want the plan to be worthy of the vision.

So, as an industry: how do we start to hold ourselves accountable for delivering on a vision, rather than merely announcing one? And how do we genuinely carry the learning from one project to the next?

Because at the moment, we don't.