Ask who answers in eighteen months

There is a question we have started asking in transformation meetings, and it changes the temperature of the room more reliably than anything else we know.

Not who owns this. That one has an easy answer, and the easy answer is usually a committee, or a function, or a name that is really a placeholder for a function.

The question is: who will be answering for this in eighteen months.

The difference matters. Ownership in the present is cheap, because in the present the programme is still an intention and everybody supports it. Answering for it later is expensive, because in eighteen months there will be a specific outcome, and somebody will have to sit in front of the board and account for it.

When you ask the second question, one of two things happens. Either a person says their own name, in which case you have something real. Or the room becomes slightly quieter, and someone offers a structure. The steering group will oversee delivery. Governance sits with the transformation office. There is a RACI.

We have learned to treat that quiet as data. It is the most useful signal available, and it is available early, and it costs nothing to obtain.

Shared ownership feels responsible. It looks like maturity and it looks like collaboration. What it actually does, quite reliably, is leave nobody able to force the hard call when the hard call arrives. And the hard call always arrives, usually about fourteen months in, usually on a Thursday, and usually in a form nobody wrote down at the start.

We want to take a serious case, and we want to treat it seriously, because people died.

In October two thousand and eighteen, Lion Air flight six-one-zero crashed shortly after take-off. In March two thousand and nineteen, Ethiopian Airlines flight three-zero-two crashed in similar circumstances. Three hundred and forty-six people were killed across the two accidents. The Boeing seven-three-seven MAX was grounded worldwide.

At the centre of the technical account is a flight control system called MCAS, which could push the aircraft's nose down based on input from a single sensor, and which pilots had not been meaningfully trained on. But the technical account is not the leadership lesson, and we would be wasting your time if we stopped there.

The leadership account, as set out in subsequent congressional investigation and extensive reporting, is about pressure and about who was in a position to stop it. Boeing was responding to competitive pressure and to a schedule. There was a strong commercial reason for the aircraft to require minimal additional pilot training, because that is what made it attractive to airlines. Engineers raised concerns. Some of the internal communications that later became public are, frankly, hard to read.

The thing we keep returning to is this. There were many people who knew something was wrong. There was no single person whose job it was to answer for whether this aircraft was safe to fly, with the authority to stop the programme, and with enough cover to be wrong about a delay.

Responsibility had been distributed across engineering, programme management, commercial, certification and regulatory liaison. Each of those groups was doing its job. Distributed responsibility is not the same thing as collective responsibility. Distributed responsibility means that at the moment a decision needs to be forced, everybody has a partial view, everybody has a reason to believe someone else is handling it, and the person with the strongest signal is the most junior person in the chain.

This is the failure mode we see, in far less consequential form, in ordinary organisations every month. The difference is scale, not mechanism.

We will be brief on the second case, because the shape of it is the same.

In two thousand and fifteen, United States regulators found that Volkswagen had fitted diesel vehicles with software designed to detect when the car was being tested and to alter its behaviour to pass emissions tests. On the road, the vehicles emitted far more than the limits allowed. The chief executive resigned. The financial cost ran into tens of billions of dollars.

What interests us is the demand that produced it. Volkswagen had committed publicly to targets for diesel performance, emissions and cost that engineers reportedly regarded as extremely difficult to meet simultaneously. The target was set at the top. The impossibility was known further down. And there was no route by which someone could say, without career consequence, that the three things could not be done together.

When a target is non-negotiable and the constraint is real, the organisation does not stop. It finds a way to make the reporting comply. That is the fundamental risk of accountability that flows only downward: pressure travels down, and bad news does not travel up, and the two facts together produce exactly this.

Now the case we would nominate as the best single illustration of what accountability actually looks like, and it turns on a moment so small it sounds trivial.

Alan Mulally became chief executive of Ford in two thousand and six. The company was in serious difficulty, heading towards a loss of many billions of dollars for the year.

He introduced a weekly meeting for his senior leaders in which each of them reported the status of their area against plan using a colour. Green, yellow, red. A straightforward management tool. Nothing innovative about it.

For several weeks, every chart from every executive was green.

The company was losing billions of dollars a year, and every single area was reportedly on plan. Mulally is said to have pointed this out with some directness: we are going to lose billions, and everything is green.

Then Mark Fields, running the Americas, put up a red. A vehicle launch had a technical problem and was going to be delayed.

The room, by the accounts of people who were in it, went silent. This was a company where that kind of admission had historically ended careers.

Mulally applauded. He asked who could help Mark with the problem.

That is the whole intervention. And we would argue it did more for Ford's turnaround than any strategic document produced that year, because of what it established. It established that the cost of reporting a red was lower than the cost of concealing one. Everything else follows from that. If it is safe to be the bearer of bad news, then bad news travels, and if bad news travels, leadership is making decisions about the actual company rather than a reported version of it.

Notice also what it required of Mulally personally. He had to absorb the discomfort of a public admission of failure in his own leadership meeting, and respond to it in a way that cost him something, in order to make it cheaper for the next person. That is what setting a culture actually costs. It is not a values statement. It is one leader paying, visibly, at the moment it is expensive.

Ford was the only one of the three large American car manufacturers not to take a government bailout in the financial crisis that followed.

The second success case is about a different kind of accountability, which is accountability for something that has no clean metric.

Satya Nadella became chief executive of Microsoft in February two thousand and fourteen. The company was profitable and large and, by most accounts, internally difficult. It had a reputation for territorial competition between divisions, for defending Windows against every threat including its own products, and for a culture in which being the smartest person in the room was the objective.

The reframe he is known for is the shift from know-it-all to learn-it-all. It sounds like a slogan, and if it had stayed a slogan nothing would have happened.

What made it real was that the strategic decisions were made consistent with it, publicly and at cost. Microsoft put its core productivity software on iPad and on Android, which under the previous logic was arming the enemy. It embraced Linux, which the company had spent years treating as an existential threat. It reorganised around the cloud, which meant subordinating the business that had defined the company's identity for thirty years.

Each of those was a decision that a defensive culture would have refused, and each one told the organisation more about what had actually changed than any internal communication could have.

The results are well known. What we want to point at is the accountability structure underneath. Nadella did not delegate the culture change to human resources with a target for engagement scores. He carried it himself, in the decisions he personally made, over years.

Culture change delegated to a function is a programme. Culture change carried by the chief executive in visible decisions is a change. We have watched several of the first kind. They produce excellent materials.

We should be fair to committees, because we are not arguing that groups are useless. We are arguing something narrower.

A committee is an excellent instrument for a particular set of tasks. Testing an argument from several directions. Bringing information together that sits in different parts of a business. Legitimising a decision so that it holds once people leave the room. Those are real functions and a good committee does them better than any individual.

What a committee cannot do is force a decision that is going to hurt somebody who is sitting in it.

Consider the structure honestly. The transformation requires that one division gives up a revenue line, or that a long-serving team is restructured, or that a system somebody built and is known for is switched off. The people affected are around the table, as they should be, because they have the expertise. Each has a legitimate case for why the timing is wrong, and the cases are good, because these are capable people describing real consequences in their own areas.

There is no mechanism in that room to overrule them. The chair can call it, but the chair is usually chairing precisely because they are neutral between the divisions, and a neutral chair overruling two directors is not a neutral act. So the decision is deferred for more analysis. It comes back in six weeks, better analysed, with the same structure and the same result.

This is not dysfunction. The committee is behaving exactly as designed. It was designed to produce agreement, and agreement is available on everything except the decisions that actually constitute the transformation.

Which is why we say the committee should be the place a named owner comes to be challenged, and not the place the decision lives. Those are two very different designs, and organisations slide from the first into the second without noticing, usually for good reasons, usually after something went wrong once.

There is a related confusion worth separating out, because it costs people a great deal of time.

Consensus and consultation are not the same, and treating them as the same is one of the more expensive mistakes a leader can make in a culture that values politeness. Much of our region does value politeness, and we would not want that traded away. But it has a specific cost here that we should name.

Consultation means everyone affected has been heard, properly, by someone with the authority to change the plan, before the decision is made. It is a genuine obligation and skipping it produces decisions that fail on contact with reality.

Consensus means nobody objects. And in any organisation of a reasonable size, on any decision of a reasonable size, that condition is unachievable without either exhausting people into agreement or diluting the decision until it does nothing. Usually the second. What arrives at the end is a version everyone can live with, which is generally a version that changes nothing anybody cared about.

The distinction we would offer is simple to state and hard to hold. Consult widely, and decide narrowly, and then be extremely clear about which mode the room is in at any given moment. Most of the frustration we have seen in transformation programmes comes from people who believed they were in a consensus process discovering afterwards that they were in a consultation, or the reverse. Say which one it is at the start of the meeting. It takes one sentence and it prevents months of resentment.

So if the argument is that transformation needs a name on it, we should be precise about what that person actually requires, because appointing someone without these things is worse than appointing nobody. It creates the appearance of accountability while guaranteeing failure, and it usually ends the career of a good person.

Three things.

Responsibility for the outcome, not for the activity. Not delivered the workstream on time. Whether the organisation actually changed. If your transformation lead can hit every milestone while the business is unchanged, you have measured the wrong thing, and you will find out late.

Access to the resources. Genuine budget, and genuine ability to move people. If the owner has to negotiate for every engineer with a divisional head whose incentives point elsewhere, they do not have authority. They have a permanent negotiation, and they will spend their eighteen months on it.

And cover to be wrong once. This is the one that gets left out, and it is the one that determines whether anything real happens. A person who cannot survive a single visible failure will not attempt anything that could visibly fail. They will run a safe programme with clean reporting and no risk, and it will deliver nothing, and they will be promoted on the basis of the reporting.

We would go further. If you are the chief executive and you are appointing someone to lead a change, tell them explicitly, in words, that they have one failure available to them. Most people have never been told that, and it changes what they are willing to attempt in the first six months.

We will finish where we started.

Ask who will be answering for this in eighteen months. Then wait, and let the silence run longer than is comfortable, because the temptation to rescue the room with a structure is very strong and it usually comes from the chair.

If a person says their name, ask them the three questions. Do you own the outcome or the activity. Do you control the resources or negotiate for them. Can you be wrong once and still be here.

And if no one says a name, you have learned something important very cheaply. You have learned it now, rather than in eighteen months, at the point where the answer to who is accountable is that everybody was, which has always meant the same thing.

That is a demanding place to stand, and it is the reason this group exists. We bring the decisions we are least certain about to people who have stood there before, and we ask them what it cost.