Systematic Learning: Construction Needs an Open Loop

When an aircraft has a serious incident, a sequence kicks in that construction would not recognise. A protected recorder is recovered. Independent investigators who had nothing to do with the flight examine what happened. The people involved are encouraged to report what they saw, often with protection from being punished for honest mistakes. And then the lesson is written up and published, so that every other operator in the world learns from one crew’s worst day.

That is an open loop. Failure goes in, learning comes out, and the same mistake gets harder to make twice.

Construction runs the opposite system. When a job goes wrong, the instinct is to find out whose fault it is, protect the position, and keep it quiet. The records that exist are built for the dispute, not for the lesson. The settlement comes with a confidentiality clause. The insurer’s policy says admit nothing. So the knowledge of what actually went wrong never leaves the room it happened in, and the next team, on the next job, walks into the same hole.

That is a closed loop. And a closed loop repeats.

We Bury the Very Data That Would Make Us Better

You can see how closed the loop is by trying to find the data. Health and safety is the one area where construction failure is, partly, on the public record. The Health and Safety Executive keeps enforcement and conviction registers, so a serious safety breach can be traced to a firm. Even there it is incomplete. The HSE’s own estimate is that only around half of the injuries that should be reported under RIDDOR actually are.

Move off safety and the shutters come down. Delay and commercial loss is almost entirely confidential. There are well over two thousand adjudication referrals a year in the UK, and they are private by design. Quality and defects is weaker still. There is no public database of building defects. The closest thing the industry has to a defects record is held privately and not shared.

So the domains where the money is actually lost, delay, quality, commercial and cultural failure, are exactly the domains where almost nothing is learned in the open. Avoidable error is estimated to cost the industry somewhere between ten and twenty five billion pounds a year, and most of it is never measured at all, because rework that is quietly absorbed never gets counted. We are flying with the recorder switched off.

This is Not a New Idea, and It Already Works

I want to be clear that I did not invent this thought. The argument that some industries learn from failure and others hide from it has been made powerfully elsewhere, including in Matthew Syed’s writing on how aviation treats a mistake as data and healthcare too often treats it as a stain. The aviation confidential reporting systems have run for half a century and gathered millions of reports. Rail, maritime and healthcare have all built their own versions. The model is proven. Construction has simply not adopted it for the things that go wrong most often.

And here is the part that should give the industry confidence, because construction already has a working example. CROSS-UK is a confidential reporting system for structural and fire safety. It was set up after structural failures decades ago and explicitly modelled on aviation. A reporter sends in what happened. A designated person strips out everything that could identify anyone. Expert panels add analysis. The reporter signs it off. Then a safety report is published with nobody named. Since 2025 it has been the official voluntary building-safety reporting body, and it has gathered more than a thousand reports over its life. As one of its own descriptions puts it, the anonymity is what allows learning to be shared without creating concerns over liability, reputation, client loyalty or insurance.

It works. It just stops at safety.

The White Space is Everything Else

Nobody runs the same open, anonymised, lesson-publishing system for the failures that wreck programmes and margins. Nobody does it for the delay that should have been seen coming, the design clash that surfaced too late, the procurement decision that looked cheaper and cost months, the cultural failure where the truth was known on site for weeks before it reached a director. That is the white space. And it is the white space precisely because that is where the fear of a claim, a premium or a reputation bites hardest.

Closing that loop is what I mean by Systematic Learning. Not a register of blame. The opposite. Analyse the system, not the individual. The aviation principle is sometimes loosely called no blame, but the more honest version, the one I think construction needs, is a just culture. Honest mistakes are treated as data to learn from. Recklessness is still accountable. The point is not to let anyone off. The point is to stop the same avoidable failure happening on the next job because nobody felt safe enough to say what happened on the last one.

It Starts on Your Own Site

You do not need an industry body to start. The same loop runs at the scale of a single project, and it is built entirely from how the people in charge react when someone brings them bad news.

I have written before that the moment you shout at the person who tells you a job is slipping is the moment they stop telling you. The same fear that stops a labourer flagging a hazard stops a project manager telling a director the truth about a programme. Marble Arch Mound is the public version of this, costs known to be climbing and hidden because nobody felt safe saying so. Every business has its own quieter version.

So the first open loop is the one in your own site cabin. Be the person the supply chain comes to first looking for a shared solution, not the person they build a claim against. React well to bad news, and you keep hearing it while it is still cheap to fix. That is Systematic Learning at the smallest scale, and it is where it has to begin.

The Bottom Line

Construction does not lack frameworks, reports or dashboards. It lacks an open loop. The industries that got dramatically safer and more reliable did it by deciding that a failure is too valuable to waste. Construction throws that value away every week, in confidential settlements and quiet rework, and then acts surprised when the same problems come round again.

Aviation made its worst days into everyone’s lessons. Construction can too, if it is brave enough to leave the recorder switched on.