Every reorganisation begins with a chart and a hope: redraw who reports to whom, and performance will follow. It rarely does. The chart is the easy part to change and the least likely to fix anything, because what slows an organisation down usually lives in the spaces between the boxes, not in the boxes themselves.

The quick version

  • Your operating model is the whole system of how your organisation turns strategy into delivered value, structure, but also decision rights, processes, metrics, incentives and skills. The org chart is one slice of it.
  • A reorganisation changes the structure. An operating-model redesign changes the whole system on purpose. Reorganising without touching the rest is why so many "restructures" disappoint.
  • Most redesigns underperform: McKinsey's survey found fewer than a quarter of organisational redesigns succeed at both meeting objectives and improving performance.
  • The fix isn't a better chart, it's coherence. Jay Galbraith's Star Model says structure, processes, rewards and people must all move together to follow the strategy, or behaviour won't change.

The idea in depth: a model is a system, not a chart

Start with what an operating model actually is, because the word gets used loosely. It is the architecture that connects strategic intent to frontline execution: how decisions get made, how work flows, how performance is measured and rewarded, and what people are capable of. McKinsey frames it as the set of choices that turn strategy into the day-to-day, identifying roughly a dozen elements, from purpose and governance to processes, behaviours and talent, that a serious redesign has to address together (see McKinsey's "A new operating model for a new world"). The structure, the reporting lines, is exactly one of those elements. It is the one everyone can see, which is precisely why it gets changed first and asked to do too much.

The practical consequence: stop reaching for the chart as the instrument of change. Before anyone redraws a line, write down what is actually broken. Which decisions are slow? Which hand-offs drop the ball? Which outcomes does nobody own? If the answer is "marketing and product never agree on priorities," a new reporting line might help, or a shared metric and a weekly forum might fix it without moving a single person. Diagnose the friction first. Structure is one possible remedy, not the default one.

Why most reorganisations fail to lift performance

Here is the uncomfortable evidence, stated plainly. In McKinsey's research on organisational redesign, summarised in "Getting organizational redesign right" (Aronowitz, De Smet & McGinty, McKinsey Quarterly, 2015), fewer than a quarter of redesigns succeeded at meeting their objectives and improving performance, and a large share simply bogged down in implementation and were never finished. The same body of work found that companies which followed more of McKinsey's nine practical "rules" of redesign were far more likely to call the effort a success.

Why so much failure? Because reorganising the structure while leaving everything else untouched changes the picture without changing the behaviour. People get new managers but the same incentives, the same broken processes, the same unclear decision rights, so they keep doing roughly what they did before, now with added confusion and a few months of distraction. The reorganisation looks like action; it rarely is.

Moving the boxes changes the picture. It only changes the behaviour if the incentives, processes and decision rights move with them.

The lesson is to treat structure as the last lever, not the first, and never the only one. When you do change it, change the things around it in the same breath: who decides what, how teams hand work to each other, what gets measured, and how people are rewarded. A new line on a chart with old incentives underneath it is theatre.

The discipline behind it: Galbraith's five levers

The most durable framework for getting this right is the Star Model, developed by organisational theorist Jay Galbraith and laid out in his book Designing Organizations. Galbraith argued that performance comes not from structure alone but from the coherence of five design choices: strategy at the centre, which then drives structure (where decision power sits), processes (how information and work flow across the lines), rewards (what behaviour gets reinforced), and people (the skills and mindsets you select and build for). His central insight was diagnostic: leaders kept changing structure and leaving the other four points of the star alone, and then wondered why behaviour didn't shift.

flowchart TD
  S(["Strategy
direction + capabilities needed"]) --> ST(["Structure
where decision power sits"])
  S --> P(["Processes
how work flows across lines"])
  S --> R(["Rewards
what behaviour is reinforced"])
  S --> PE(["People
skills + mindsets you build for"])
  ST -.-> B(["Behaviour
+ performance"])
  P -.-> B
  R -.-> B
  PE -.-> B
Galbraith's Star Model, strategy sets the requirements; the other four levers must move together to change behaviour. Leaders Loop

In practice, use the star as a checklist the moment a reorganisation is on the table. For every structural change, ask the other four questions out loud. Does this strategy actually require a new structure? Which cross-cutting processes need rebuilding so the new boxes can work together? What in the reward system would quietly pull against the new design? What skills will people need that they don't have yet? A redesign that answers only the structure question is, by Galbraith's logic, almost designed to disappoint.

An honest limitation. The Star Model is a thinking tool, not a measurement, and it can flatter the planner. It tells you which levers to consider; it does not tell you the right setting for any of them, or how much disruption a given change will cost the people living through it. Coherence on paper is not the same as a workable transition, and the model is largely silent on the human cost of getting from the old design to the new one. Treat it as a way to avoid the obvious omission (changing structure alone), not as a guarantee that a coherent-looking design will land. The transition is where redesigns actually fail, and no framework substitutes for managing that well.

A worked example

Take a mid-sized insurer, call it Meridian. (Illustrative throughout; this is a teaching example, not a real company.) Customer complaints are rising and product launches keep slipping. The executive team's instinct is a classic one: merge the three regional product teams into one central product function, so there's "one throat to choke." A reorganisation. The chart gets redrawn over a weekend.

Run it through the diagnosis-first lens instead, and a different picture appears. The real friction isn't that product is split three ways, it's that product, underwriting and IT each own a piece of every launch and none of them owns the launch end to end. The slowness lives in the hand-offs, not the structure. Centralising product would reshuffle the boxes and leave the hand-offs exactly where they are.

flowchart TD
  A(["Symptom: launches slip,
complaints rise"]) --> B{"Is the problem
structure, or the flow
between boxes?"}
  B -->|"Reflex: merge the
product teams"| C(["Reorg only:
new chart, same hand-offs
little changes"])
  B -->|"Diagnose the friction"| D(["No single launch owner;
broken hand-offs;
misaligned incentives"])
  D --> E(["Redesign the system:
name a launch owner,
rebuild the cross-team process,
align the metric"])
  E --> F(["Behaviour shifts
launches speed up"])
The reflex reorganisation versus the operating-model redesign, same symptom, very different intervention. Leaders Loop

The operating-model answer touches four of Galbraith's five levers at once. Structure changes modestly, appoint an accountable launch owner who can pull the cross-functional team together. Process changes more, a single, mapped launch flow with clear hand-off points replaces three informal ones. Rewards change, the launch owner and the contributing teams share a delivery metric, so nobody is rewarded for protecting their own box at the launch's expense. People change, the new owner needs influence-without-authority skills the old role never required. The chart barely moves; the behaviour does. That is the difference between a reorganisation and a redesign.

Frequently asked questions

What's the difference between an operating model and an org structure?

The structure is the reporting lines, who sits where and who manages whom. The operating model is the whole system that turns strategy into delivered value: structure plus decision rights, processes, metrics, incentives and capabilities. Structure is one component of the operating model, and usually not the one causing your problem. Changing the structure alone is the most common and least effective intervention.

When is a reorganisation actually the right answer?

When the friction genuinely traces to the structure, decision power sits in the wrong place, accountabilities overlap so nothing gets owned, or the structure contradicts the strategy (a structure built for regional autonomy when the strategy now demands one global product). Even then, change the surrounding levers, processes, rewards, skills, in the same effort, or the new structure inherits the old behaviour.

Why do so many reorganisations fail?

Two reasons dominate. First, leaders change structure and leave processes, rewards and people untouched, so behaviour doesn't move. Second, the redesign bogs down in a messy transition that was never planned for, McKinsey found a large share of efforts simply stall during implementation and are never finished. The design is the easy part; the transition is where it lives or dies.

How often should we reorganise?

Less often than most organisations do. Reorganisations are expensive in attention and morale, they pause real work, unsettle people, and reset relationships that took years to build. Frequent restructuring is often a sign that the underlying operating-model problem was never diagnosed, so each new chart fails and triggers the next. Fix the system once, properly, rather than redrawing the boxes annually.

How do we keep people productive during the redesign?

Decide and communicate the new design quickly, then over-invest in the transition: name owners early, be explicit about who decides what from day one, and don't leave roles ambiguous for months. The damage in most redesigns is done in the limbo period, when nobody knows who owns what and the best people start looking elsewhere. Speed and clarity on the transition matter more than perfecting the chart.

Related in the Toolkit

An operating-model redesign is a change programme, so the disciplines of leading change apply directly, pick a change model to structure it, and plan from the start for the resistance any restructure provokes (managing resistance & driving adoption).

Where to go next