The unasked question that sinks good businesses

On 1 March 2017 Carillion published accounts that showed a healthy company. The board paid a record dividend of £79 million and awarded large bonuses to senior executives. Four months later the company warned that its contracts were worth £845 million less than it had said. Come January 2018 the whole thing was gone: liquidation, nearly £7 billion owed, £29 million left in the bank.

The Commons committees reported in May 2018. Their phrase was "a story of recklessness, hubris and greed". Fair enough. But the line I keep coming back to is quieter. The non-executive directors, the MPs wrote, "failed to scrutinise or challenge reckless executives". Nineteen years of unqualified audits. A full board. Advisers on retainer.

I am a civil engineer, not an accountant. I spent twelve years on major projects in the UK and the Middle East, and the last five years researching and writing a book about failures like this one. The pattern that finished Carillion is the one I keep finding in rooms where everyone is clever and the data is fine.

On the evening of 27 January 1986, the engineers at Morton Thiokol told NASA over a teleconference not to fly if the seals would be colder than on any previous launch. Their floor was 53 degrees Fahrenheit, and the morning forecast sat far below it. They had watched erosion on the seals after earlier cold flights, and they knew they were outside their own data. NASA's managers found the recommendation unacceptable and asked for a reassessment. Thiokol's people went off line to caucus, and when they came back management had overruled its own engineers. The recommendation was withdrawn. Seven people died the next day. The warning had been on the record the night before, from Thiokol's seal engineers.

Twenty-four years later, on 20 April 2010, the crew of the Deepwater Horizon accepted conflicting pressure test readings and continued work. Eleven men died. On board that same day, BP and Transocean executives were congratulating the crew on seven years without a lost-time incident.

Five years of reading the investigation reports for my book taught me they all finish in the same place. The data existed. Someone in the room had the doubt, usually more than one someone. The question that would have stopped the day never got said aloud, or never got said by a person the room would listen to.

Why do smart rooms go quiet? A plan with a date and a budget behind it has a pull of its own. Raising a doubt starts to feel like sabotage, so people sit on it and wait for somebody more senior to go first. Senior people are busy being confident. Meanwhile the whole room keeps score of what happens to the awkward one. When the last person to challenge a plan gets moved off it, the lesson travels fast, and it is never the lesson the company meant to teach.

What would have changed those rooms costs very little. Take the Carillion board before those accounts were published. One person around that table could have asked: which of our big contracts is losing money, and how do we know? What has to be true for these numbers to hold? When did someone last stand up in this meeting and tell us something we did not want to hear, and what happened to them afterwards? None of those questions needs an accountant. They need somebody whose job that day is to be the awkward one, so give the role a name and rotate it. And let the most junior director answer first, before the senior voices set the weather.

Now the Thiokol caucus. The engineers had just said the seals were outside their own data. Before any reversal, the question is: what has changed? New data, a new analysis? Nothing had changed except the pressure. A decision that size, reversed in one short caucus, deserved the question asked early: if this goes wrong tomorrow, what will we wish we had asked tonight? Then the disagreement goes on paper before the vote, so that "we all agreed" means something.

Make the question a standing line at every large commitment: what are we missing? And watch the people who used to argue. When they go quiet, that is information, and it is rarely good news.

I use a five-part check for this, put together over five years of researching the failure record: frame the problem before solving it, do the research, interpret honestly, pull the threads together, and keep an eye on your own thinking while you do all of it. The last part is the hardest, because the mind that made the plan is the mind checking it. It is the core of my book, What Are We Missing? How to Think More Clearly Before You Decide, which is planned for February 2027.

Carillion's board minutes will be full of answers. The record I would want to see is the other one: the question that was asked, who asked it, and what happened next.

‍

Zafar Alam is a civil engineer and Chartered Construction Manager. More about him and the book at zafaralam.co.uk

‍