Signs an organization has outgrown its structure

11:47 p.m.

At 11:47 p.m. on a Tuesday, the owner of a company billing twenty million dollars a year approves a 400-dollar purchase over WhatsApp. It is roughly the seventieth message he has answered that day. It doesn't bother him — or so he says: "I'm on top of everything, that's how this works." And he is right about one thing: that is how it worked. That is how he founded the company, how it survived the crises, how it got this far.

The detail is that the company is no longer the one he founded. It bills six times more than eight years ago, employs three times the people, operates in two countries. But it decides exactly as it did when there were twelve people around a table: everything goes through him.

No alarm sounds when an organization outgrows its structure. There is no noise of something breaking. There are scattered symptoms that look like people problems —"I need people who take ownership", "the new manager didn't adapt", "around here you have to stay on top of everything"— and that are, in reality, a single system problem. Structure does not fail loudly: it becomes too small in silence, and the cost of not seeing it is paid every day, in installments.

This article describes the signals for catching it early, the evidence on what ignoring them costs, and the sequence for getting out without bureaucratizing the company along the way.

The math nobody did

Let's start with what almost nobody calculates: team growth is linear; coordination growth is not.

In 1933, the Lithuanian consultant V. A. Graicunas published a calculation that should be taught in every family business: when a manager adds subordinates, the relationships he must attend to do not grow one by one; they explode geometrically, because they include the cross-relationships among subordinates and the combinations of groups. With 4 direct reports, a manager handles 44 possible relationships. With 8, it's 1,080. With 12, it's 24,708.

The exact numbers matter less than the shape of the curve: doubling the team multiplies the complexity of coordinating it twentyfold. That is why the founder who is "on top of everything" is not exaggerating his exhaustion: he is doing a job that arithmetic declared impossible ninety years ago. And that is why "hiring an assistant for the owner" solves nothing: the bottleneck is not the owner's calendar; it is the architecture in which every line passes through a single node.

Anthropology reached the same place by another route. Robin Dunbar estimated that humans can sustain about 150 stable relationships: the maximum size of a community that works "because everyone knows everyone". W. L. Gore, the Gore-Tex company, took the figure seriously decades ago: when a plant approaches 150 employees, it splits it. Below that threshold, informal coordination —the hallway, the trust, the "leave it to me"— is enough. Above it, you need what informality can no longer provide: structure, roles and explicit routines.

Evolution and revolution

The second hard fact comes from a classic that is half a century old and refuses to age. In 1972, Larry Greiner published in Harvard Business Review the model of organizational growth that the magazine itself reissued as a classic in 1998: organizations do not grow along a smooth ramp but through phases of evolution interrupted by predictable crises, and what solves each phase is precisely what causes the next crisis.

The first transition is the one that concerns us. The young company grows through creativity: the founder sells, decides, puts out fires, and the energy of that stage is its competitive advantage. Until volume makes it unviable and the crisis of leadership appears: the organization needs professional direction —structure, priorities, management— and the founder experiences that need as a betrayal of the original spirit. The following phases (direction, delegation, coordination) bring their own crises —of autonomy, of control, of red tape—, but most mid-sized companies in our region stand exactly at the first frontier: the founder's operating system no longer runs on the hardware of the company he himself built.

Two of Greiner's ideas are gold for decision makers.

The crisis is not an anomaly: it is the agenda. If the company grew, the structural crisis is not a sign that something was done wrong; it is a sign that something was done right. The question is not "why is this happening to us?" but "which phase is it time to build now?".

Phases cannot be skipped. The company that jumps from foundational chaos to corporate committees without building the middle —clear roles, real middle management, basic routines— does not professionalize: it puts on a costume.

Seven signs it is already happening

Greiner's crises give notice. These are the seven signals we see most in the field, with the mechanics behind each. None is serious on its own; three or more, sustained over time, make the diagnosis.

1. Everything goes through the owner. Small purchases, permissions, special prices, conflicts between areas: the queue of decisions waits for one person. The test is simple and cruel: if the owner disconnects for two weeks, does the company decide or accumulate? When the answer is "accumulate", the organization does not have a busy leader: it has an institutionalized bottleneck. And waiting decisions carry costs no accounting system records: the urgent freight hired late, the client who didn't wait for the quote, the machine standing idle awaiting a signature.

2. The heroes are the firefighters. In the company that outgrew its structure, internal prestige is earned in the emergency: whoever stayed until whenever, whoever saved the shipment, whoever fixed it with one phone call. Nobody rewards the person who prevented the fire, because the prevented fire is invisible. The result is an invisible incentive system that rewards reaction and punishes prevention — exactly the opposite of what the company needs at the size it has already reached.

3. Hiring does not relieve. People are brought in "to decompress" and, months later, the overload is intact — only now with more salaries. The reason lies in Graicunas's math: without clear roles and defined interfaces, each hire adds more coordination load than it removes in work. New people do not find a position: they find a nebula, and learn the trade by oral tradition.

4. Knowledge lives in people, not in processes. There is one employee without whom "invoicing is impossible" and another who is the only one who knows how the big client's order is put together. Each of them is, at once, an enormous asset and an uninsured risk. The audit question is direct: how many critical processes depend on a single head? Each of those single points of failure is a business interruption waiting for a date.

5. Meetings without decisions, decisions without owners. Much is discussed and little resolved; the same topics return week after week; everyone opines on everything and nobody is accountable for anything specific. It is the symptom of a structure where roles are defined by history ("Jorge has always handled that") and not by design. The acute version: decisions that are made and unmade, because there is no record of what was decided or who stands behind it.

6. The numbers arrive late, or not at all. The income statement is known on the 25th of the following month; the real cost per product line is a heroic estimate; the management dashboard is, in practice, the bank balance and the owner's gut feel. Without timely numbers there is no possible delegation —how will someone let go of decisions when they have no way of finding out quickly if something goes wrong?—. Informality of information is the mother of centralization.

7. The company grows and the margin does not. The most silent signal and the most expensive: revenue rises and profitability flattens or falls. These are the internal diseconomies of informality: emergencies paid at a premium, rework, purchases without process, discounts without control, structural overtime. Disorderly growth has a rising unit cost that appears on no accounting line under that name — it appears dissolved in all of them.

What ignoring them costs

So far, the symptoms. What does it cost not to treat them? For a long time the answer was rhetorical ("a lot"). Today there are numbers, and they are among the best data management research has produced in the last two decades.

Nicholas Bloom (Stanford) and John Van Reenen (LSE) built, with Raffaella Sadun (Harvard), the World Management Survey: thousands of mid-sized companies assessed across more than 30 countries with a protocol that measures something very concrete — whether the company has structured management: targets, indicator tracking, review routines, performance management. The findings, published in the leading academic economics journals since 2007: structured management practices are strongly and consistently associated with higher productivity, profitability and survival, across every country and industry measured; in their "management as a technology" research line, about 30% of the productivity gap between countries is attributed to management quality; and family firms run by hereditary succession tend to score systematically lower — not for lack of talent, but because informality does not fix itself with a generational handover.

The obvious objection is causality: does good management make better companies, or can better companies afford good management? Bloom and his team answered with a now-classic experiment (Does Management Matter?, QJE 2013): they took mid-sized textile plants in India and randomly assigned them consulting in basic management practices — the same routines, indicators and controls we are discussing here. The treated plants increased productivity 17% within the first year, with gains in the order of 300,000 dollars per plant per year. Same owner, same people, same machines: the management system changed. Causality was demonstrated in the uncomfortable direction: informality had been costing, every year, a figure nobody saw because it appeared on no financial statement.

McKinsey reaches consistent magnitudes by its own route: its Organizational Health Index — built on millions of responses across thousands of organizations — shows that organizations in the top quartile of health deliver, over the long run, about three times the total shareholder returns of those in the bottom quartile, and that those improving their health record EBITDA gains well above their peers.

Structured management is not an aesthetic taste of large corporations: it is a technology — probably the cheapest one your company has not yet installed.

Professionalizing without bureaucratizing

The founder's resistance holds a legitimate argument: he has seen companies drown in forms, committees and slide decks, and wants none of it. He is right in the fear and wrong in the conclusion. Bureaucracy is process without purpose; professionalization is the opposite: the minimum structure required so that decisions do not depend on a single calendar. The acid test for every new element is one question: does this speed up or slow down a decision?

The sequence that works in the field — and the order matters:

1. Diagnosis with data, not opinions. Two or three weeks watching how decisions are made, where they wait, which numbers exist and which get invented. The symptom map above, measured in the real company.

2. Structure before people. Define the roles the operation needs — with responsibilities, authority and metrics — and only then ask who fills them. The reverse order ("we have Marta, let's invent her a position") perpetuates the problem under a new org chart. It is Alfred Chandler's old lesson: structure follows strategy, not the other way around.

3. Management routines. A short weekly cadence per area with indicators and commitments; a monthly results review; one-page minutes where decisions have an owner and a date. Boring on purpose: professional management is the systematic elimination of surprise.

4. Few indicators, on time. Before an encyclopedic dashboard, five numbers per area that arrive fast and are taken seriously. The perfect indicator that arrives on the 25th loses to the reasonable indicator that arrives on Monday.

5. Delegate through pilots, with real authority. Delegation is not a speech; it is a verifiable transfer: this decision, up to this amount, with this indicator and this reporting routine. Start with one area, learn, extend. The owner does not "let go of everything": he lets go in stages of something that now, at last, has a net underneath.

Six to twelve months of that work is usually enough to cross Greiner's first frontier. What you get on the other side is not a colder company: it is a founder who goes back to working on what only he can do — the future of the business — because the present, finally, runs on a system.

The good news is the same as at the start: this is not a rare pathology; it is the scheduled crisis of companies that grew well. It has a diagnosis, a sequence and a way out — and the way out is not working longer hours. It is building, once and for all, the company you already have.