When Purpose meets Structure
Seeing the hidden architecture of organisations
“Who comes first: your customers, staff or shareholders?” I once put this question to the CEO of a multinational retailer when applying for a job. In my prior career, too often I’d seen short-term financial pressures supersede long-term prudence, and I was wary of the culture I might be stepping into.
He paused for a moment, then said: “If you don’t keep your customers happy, you don’t have a business. If you don’t keep your staff happy, you won’t keep your customers happy. And we have promised a 15% return on capital to the stock market.” I thought it a good answer. And I took the job.
What I found subsequently, however, was that the organisation’s behaviour was not quite aligned with the CEO’s answer. Rather than balancing interests, we seemed more interested in what could be measured. So-called “shareholder value” tended to dominate decision-making.
It is not, I believe, that the CEO was trying to deliberately mislead. Rather what he believed was only part of the picture. He was just one person (albeit a significant one) amongst 90,000. The organisation was an enormous web of relationships; an entire ecology of people, thoughts, habits, information systems and expectations. He wasn’t “in charge” — how could he be? The system was. His role was more political than managerial, steering a path between powerful forces while trying to keep the organisational ship on track.
Systems thinker W. Edwards Deming observed that “a bad system will defeat a good person every time”. I think he was right. I saw how decent, well-intentioned people repeatedly made decisions that compromised their own values, not because they had stopped caring, but because the system made other choices increasingly difficult.
I had joined believing that good leadership was the answer. I left wondering whether something deeper mattered at least as much: the constitution of the organisation itself.
How it starts
Every day, in their thousands around the world, a new enterprise is born. A mother in Pakistan establishes a shop now her children are at school. A carpenter in Nigeria leaves his job and sets out on his own. A student in Argentina realises what she’ll do when she leaves college and pens the outlines of a business plan.
At first, there is little distinction between the founder and the organisation. The founder is owner, board, employee and carrier of values all in one. Their judgement is its governance; their values are its values; their relationships are its relationships.
But if the enterprise survives, something subtly starts to shift. People join. Responsibilities spread. Money comes in. Shares are issued. Gradually, almost imperceptibly, the organisation becomes something separate from the person who started it.
This is a key transition in the life of an organisation. The founder can no longer hold its purpose alone. What was once carried implicitly by a person now has to be carried, somehow, by the institution.
This is where structure begins to matter. How will decisions be made? Who will hold power? What happens when interests diverge? And what, if anything, should remain true when the founder is no longer there? These are the sort of questions that inform the creation of the structure.
At first, it’s purpose that shapes structure. But over time, structure shapes purpose.
But what do we mean by structure?
The hidden architecture
An organisation’s structure is made up of the formal and informal rules, practices, information flows and patterns of relationship that shape how people think, decide and act. Together they create what I think of as the organisation’s hidden architecture.
Like a product or service, this architecture can be designed. Yet surprisingly little of it is. Much develops by default, through templates inherited from previous generations or supplied by advisers. A company adopts standard articles. A charity appoints trustees. A founder issues shares to an investor. A board establishes ways of meeting and making decisions. Each choice seems quite ordinary. Gradually, though, they accumulate into a system that begins to shape the behaviour of everyone within it.
For purpose-led organisations, this matters. But purpose itself is a slippery word. The definition I like best comes from cybernetics pioneer Stafford Beer: “the purpose of a system is what it does.” In other words, an organisation’s purpose isn’t simply what it says about itself, but the function it performs within the larger systems of which it is part.
If purpose is expressed through what an organisation actually does, then the way power, money, information and responsibility move through it cannot be separated from its purpose. Structure doesn’t merely protect purpose. Over time, it helps produce it.
So how might we structure for purpose?
Different answers to the same question
People have adopted many different constitutional arrangements to protect what’s important to them. I have grouped these into how the organisational design answers an implicit question: who, or what, do we trust to keep us honest? Different constitutional traditions answer that question differently.
Trust particular people
Some like to place their trust in people. But not people in general. Certain people. It might be just one person — a founder who only really trusts themself (Elon Musk might be an example). Or it may be members of a particular family. There are many such dynasties, including Bechtel, Hyundai and Rothschild. Mars too has stayed family-controlled for generations, in part as a deliberate choice to retain their independence and values.
A different version of this approach places trust in carefully chosen successors. These may not necessarily be considered the “best” people (i.e. smartest, most experienced, most visionary) but rather those with the deepest understanding of the organisation’s story — who get its DNA implicitly. Some companies appoint their CEO exclusively from within. The chair of retailer John Lewis is always appointed by their predecessor.
Actually, this placing of trust in individuals is a thread running through most companies and charities. A board is appointed, surrounded by rules and expectations, and expected to be responsible. Sometimes it works — but it goes wrong far more frequently than the supporters would care to admit.
There are obvious risks in relying too much on one person or a small handful of people, even specially selected ones. So many prefer to place their faith in markets.
Trust the market
Markets offer a rather different answer. Instead of trusting particular people to make good decisions, we can trust the interactions between many people making decisions for themselves.
At their best, markets do something quite remarkable. They coordinate the activity of millions of people who may know almost nothing about one another, gathering dispersed information through prices, rewarding what people value and withdrawing resources from what they don’t. No central authority has to understand the whole system. Coordination happens via the invisible hand described by Adam Smith.
For businesses, exposure to markets can bring a powerful discipline. Customers can walk away, investors can move their capital and poor decisions have consequences. Public markets have also enabled companies to raise enormous amounts of capital from widely dispersed investors and build enterprises at a scale that would otherwise have been difficult to imagine.
The difficulty comes when we ask markets to judge things they cannot easily see. The invisible hand can also be a blind one. Markets can price a barrel of oil with extraordinary efficiency while struggling to price the consequences of burning it. They can reward attention without distinguishing particularly well between attention that enriches us and attention that harms us. Markets are extraordinarily good at processing some kinds of information, and remarkably poor at valuing others.
Trust the membership
If neither exceptional individuals nor anonymous markets seem sufficient, another answer is to place trust in the people closest to or most affected by the organisation: its employees, customers, suppliers or wider community. The assumption is not that these people are wiser individually, but that collectively they hold knowledge and experience that a small group at the top cannot.
The CEO of Riverford, the organic farmer and online retailer, told me how valuable its employee-ownership model became during Covid. As the business rapidly moved people from sales and marketing into packing and delivery, its co-owner council (the “eyes, ears and voice” of employee owners) provided a trusted network through which information could move quickly in both directions. Leadership could listen and act with a degree of trust that would have been difficult to manufacture in the middle of a crisis.
Get Patrick Andrews’s stories in your inbox
Join Medium for free to get updates from this writer.
Membership can serve another function too: legitimacy. The Forest Stewardship Council brings commercial, environmental and social interests into its governance, deliberately balancing constituencies that might otherwise struggle to trust one another. For a certification organisation whose product is ultimately trust, that matters.
Co-operatives, employee-owned businesses, mutuals and community-owned enterprises all draw on a similar intuition: those with a meaningful stake in an organisation should have some meaningful part in its governance. Knowledge and authority are distributed rather than concentrated in a founder, family or small board.
But membership doesn’t magically make people wise or well-informed. Groups can become inward-looking, participation ritualistic, and dispersed responsibility can sometimes mean nobody feels particularly responsible at all.
For some, this suggests the need for boundaries and rules that cannot simply be changed when they become inconvenient. This points to regulation.
Trust external regulation
Some founders choose to embed their purpose through adopting a regulated legal architecture, placing tight restrictions on the ability of those leading the organisation to deviate from their mission.
This is the thinking behind the Community Interest Company (CIC) in the UK, with its statutory “asset lock” that restricts distributions to members or shareholders. It’s also what drives founders to adopt charitable status, accepting tighter regulation and restrictions on trading activity in exchange for an irrevocable commitment to public benefit.
This can be understood as an act of constitutional self-discipline. Rather than relying on individuals or a small team to keep the organisation in line with its purpose, constraints are imposed by an independent regulator. The legal form provides a guardrail that outlasts any individual.
The disadvantage is rigidity. Legal forms are generic templates, not bespoke solutions. A founder who locks their mission into a particular legal structure may later discover that the structure is also locking out the flexibility they need to adapt, invest, or respond to a changing world.
Trust the design?
Some founders approach the question differently. Rather than looking for one place to put their trust, they become interested in the system itself.
I met Hugo Spowers, the founder of Riversimple, while he was developing a radically different hydrogen-powered car. As an engineer, he understood that changing one part of a complex system changes the behaviour of the whole. His question was whether the organisation behind the car could be approached in the same way. If the product needed to be designed from first principles, why not the company?
I have encountered this instinct repeatedly, often among people whose work already requires them to think in systems. The founders of Library of Things arrived with a different version of the same question. They didn’t feel they were really a conventional for-profit business, but nor did “non-profit” describe what they were creating. Rather than choosing between the available boxes, we began with the organisation itself: what was it trying to do, who was it there to serve, and what relationships would allow it to remain true to that?
This is where constitutional design becomes interesting. It isn’t simply about choosing a legal form or adding protections to a company’s articles. It means looking at the organisation as a living system of relationships, power, information, money and responsibility — and asking how those elements might reinforce one another.
Some of the work resembles failure analysis. What happens if the founder leaves? If an investor wants to sell? If the people governing the organisation lose touch with those it exists to serve? Lawyers are good at asking these questions. But constitutional design asks another kind as well. Who should have a voice here? Where should decisions best be made? What information needs to travel where? What are we trying to protect — and what must remain free to evolve?
The answers rarely point to a single mechanism. They may involve founder judgement, membership, external regulation and constitutional safeguards in different combinations. They may include guardian shares, stakeholder representation or steward ownership. But the details matter: who appoints whom, what happens when people disagree, what can be changed, by whom, and how the organisation learns when its arrangements are no longer serving it.
Perhaps most importantly, good constitutional design doesn’t assume we can foresee everything. It can build in feedback: different voices, flows of information, moments for reflection and opportunities to adapt. The constitution becomes less a set of instructions for controlling the future than a framework within which the organisation can continue to learn.
There is no perfect design. The aim is something more modest: to create conditions in which purpose, people and structure have a better chance of remaining in relationship as the organisation evolves.
And evolve it will. Like raising a child, what is required of us changes as the organisation grows. A structure that works at one stage may not work at another, and we cannot anticipate every question at the beginning.
Fortunately, we don’t have to. Most of the time an organisation simply gets on with its work. But every now and then, an apparently ordinary decision changes what becomes possible later.
These are constitutional moments.
Constitutional moments
“I want my shares back!” I was working with the founder of an ambitious social enterprise who had issued 5% of her new company to a finance director. Within months it became clear he had to go. Firing him seemed straightforward, but getting the shares back wasn’t. He believed he’d earned them, and she discovered she had no power to make him return them. She wasn’t happy!
I’ve seen the opposite too. A founder, keen to “do things properly”, established a charity and appointed independent trustees. Years later, just as the organisation was beginning to thrive, the board removed her. The structure she had adopted to protect the mission became the mechanism through which she lost the ability to shape it.
Neither founder had been particularly foolish. Both had simply crossed a constitutional threshold without realising quite how consequential it was.
These choices are not necessarily permanent. In 2025 the founders of luxury furnishings brand House of Hackney raised new capital and bought out their private-equity investor. They had concluded that their existing ownership structure was constraining the direction in which they wanted the business to evolve. Sometimes an organisation outgrows a structure that once served it.
The point is not to foresee every consequence. We can’t. Good constitutional design is therefore less about getting everything right at the beginning than about recognising the moments that matter, asking better questions, and leaving enough room for the organisation to evolve.
Looking beneath the visible
Modern business pays close attention to what it can see and measure: growth, valuation, market share, profit.
Yet much of what determines the health of an organisation is harder to see. These are things like trust, integrity and relationships. It’s about the distribution of power, the quality of information and the ability to learn. And whether there is a shared sense of what the organisation is there to do.
This invites a different kind of attention. If you are part of an organisation, you might pause occasionally and look beneath its visible structure. Who really has a voice here? Where does information travel freely, and where does it stop? What gets rewarded? What would happen if a founder or other key person disappeared tomorrow? And which parts of the organisation have simply become the way they are because nobody thought to question them?
There may be no immediate need to change anything. Sometimes seeing the structure more clearly is enough to begin. And maybe we are becoming better at seeing this.
Until relatively recently, there was no discipline called sociology. Societies had existed for thousands of years, but it took a particular shift in perspective to begin studying society itself: not simply the individuals within it, but the patterns and relationships that emerged between them.
I wonder whether something similar is happening with organisations. We have become extraordinarily good at creating them. Companies employing hundreds of thousands of people, operating across continents and coordinating supply chains of almost unimaginable complexity have emerged within a few generations. But our understanding of these organisations as whole systems has developed more slowly.
Now we are learning to look differently. Not only at the leaders, the legal form, the strategy or the numbers, but at the relationships between them. At how power moves, information travels, and what gets rewarded. At whose voices are heard, and whose ignored.
Seen this way, some of the founders experimenting with steward ownership, employee ownership, stakeholder governance and other unfamiliar structures may be doing something more interesting than inventing new organisational models. They may be learning to see organisations differently: less as machines to be controlled, and more as living systems to be stewarded.
I don’t know whether this amounts to a new discipline, or even a new way of thinking. Perhaps we are simply recovering something we had forgotten. But I increasingly encounter people asking questions for which the familiar categories — company or charity, for-profit or non-profit, shareholder or stakeholder — no longer seem quite sufficient.
The interesting thing is not that they have found the answer. It is that they are asking different questions.
Purpose and structure go hand in hand. If purpose matters, then so does the structure through which we ask it to endure.
