Transformation designed without people is undone by people.

The people who will run the new operating model are the only ones who know what the old one is actually running on. Most transformation programmes never ask them.

The most expensive engagement I have been brought into as a recovery effort had everything a transformation should have.

A rigorous strategy. External expertise. Executive commitment. A programme structure that was genuinely well-designed. Investment that reflected the scale of the ambition.

What it didn't have was the people who actually ran the organisation anywhere near the design process.

They had been consulted. There had been surveys. There had been town halls. A change management workstream had produced communication materials that were, by any measure, excellent.

But the operating model had been designed in a room those people never entered. And when it landed in their hands, they did what any rational person does when handed a new way of working that doesn't match the reality they operate in every day.

They adapted it. Quietly. Without malice. With complete effectiveness.

Within eight months, the organisation was running a hybrid of the new model and the old one. The hybrid looked like the new model in the reporting. It worked like the old one in practice.

The transformation had been delivered. It had not landed.

The structural reason this keeps happening

There is a widely held belief that involving people in transformation is primarily about buy-in. Get them in the room, hear their concerns, address their objections, secure their commitment. The human dimension as a consent problem.

This framing produces consultation exercises. Engagement sessions. Change networks. All of which are useful and none of which solve the actual problem.

The actual problem is not consent. It is intelligence.

The people who run the current operating model are not primarily valuable to a transformation programme because they need to be persuaded. They are valuable because they know things about how the organisation actually works that exist nowhere else.

Every operating model has two versions. The designed version — the org chart, the process maps, the governance documents, the system architecture. And the real version — the informal decision-makers, the workarounds that became the default, the load-bearing relationships that don't appear in any documentation, the knowledge that lives in specific people's heads and disappears when those people leave.

The designed version is what transformation programmes redesign. The real version is what makes the organisation run.

The redesign that ignores the real version will be undone by it. Not deliberately. Not dramatically. Gradually — as the people running the new model encounter the gaps between what was designed and what actually works, and fill those gaps with what they already know how to do.

What I have learned from watching this happen repeatedly

After enough transformation programmes, you start to recognise the pattern early.

The signal is not resistance. Resistance is visible. You can work with resistance.

The signal is something quieter: a gap between what people say in programme meetings and what they do between them. A set of workarounds already forming before the new model has officially launched. A group of people going through the motions of adoption while waiting to see whether the new design will survive contact with the organisation.

That gap almost always traces back to the same root: the design process didn't include the people who knew where the current model was fragile, where the informal structures were load-bearing, and where the proposed redesign would break on contact with operational reality.

Not because the designers were careless. Because the programme was structured as a design exercise followed by a deployment exercise — with involvement as a checkpoint rather than as the primary source of intelligence about what was actually being redesigned.

The difference between consultation and co-design

Co-design is not a more intensive version of consultation. It is a fundamentally different relationship between the design process and the people closest to the work.

Consultation asks: here is what we are planning to do — what do you think?

Co-design asks: what is the current model actually running on — and what would need to be true for the new one to run better?

These produce different conversations. Different data. And a fundamentally different operating model at the end of them.

In every programme where I have seen transformation genuinely land — where the new way of working was still running twelve months after the project team left — the people who had to run it were involved early enough to shape it, not just late enough to receive it.

Not as a consultation exercise. As the people who held the structural intelligence the design couldn't work without.

What genuine involvement requires structurally

Three structural conditions distinguish programmes where people involvement produces something useful from programmes where it produces the appearance of involvement.

The people closest to the work are involved before the solution is defined. The most common failure mode: bringing frontline involvement in to validate a design that is already complete. The intelligence arrives too late to change anything. The session produces minor modifications and a tick in the engagement box. The real model remains unmapped and the new design remains unaware of it.

The questions asked of them are structural, not operational. "What do you think of the new process?" produces surface feedback. "When this process breaks down, what do you actually do?" maps the real operating model. One produces opinions. The other produces the structural intelligence the redesign needs.

What people surface is treated as a design input, not an objection to manage. The moment involvement is framed as resistance to address rather than intelligence to use, the most valuable information stops being shared. People learn quickly whether their knowledge is genuinely shaping the design or politely being noted and set aside.

The cost of getting this wrong

Recovery programmes — the engagements where I am brought in after a transformation has been delivered and has not landed — are significantly more expensive than the original programme would have been if the people dimension had been designed in from the start.

The new model has to be redesigned. The organisation has to move again, from a starting point of reduced trust because the last effort produced movement without change. The goodwill the programme consumed has to be rebuilt before anything new can begin.

There is also a cost that doesn't appear in a budget line. The knowledge that walked out when the programme team left. The people who knew where the current model was fragile — who had been telling informal stories about what didn't work and why, who had decades of institutional understanding about why certain things were done certain ways. In the programmes that involved them structurally, that knowledge became the foundation of the redesign. In the programmes that didn't, it was still there — being used to quietly undo it.

Three questions worth sitting with

The people who run your current operating model — what do they know about how it actually works that doesn't appear in any process document? And when did your transformation design process ask them?

When the last major change in your organisation was delivered, what happened in the first six months? And what does that tell you about where the design process left intelligence gaps?

In your current programme, are the people closest to the work being consulted or co-designing? And how would you tell the difference?

Transformation that sticks is almost never the most elegantly designed. It is the one built with the people who had to run it — not for them.

That distinction sounds small. The difference in outcome is not.

ECC works with organisations in pharma, financial services, and complex regulated environments to design transformation from the inside out — building the operating model around the structural intelligence that only the people running the current one can provide. If this is where your programme is, I would welcome the conversation.