Companies restructure constantly and change almost nothing. The review is how the structure survives.
getty
“We need to relook at our operating model.”
Every executive has been in the room where that sentence lands. It arrives late in a difficult conversation, after the numbers have been walked through twice and nobody has anything useful left to say about them.
It is the most serious thing an organization can say about itself: that the way it is built may be the problem. It is also, reliably, the sentence that ends the meeting. The tension drops. The problem has been handed a container. Someone will be assigned to it, a workstream will be stood up, and everyone can go back to work. No sentence in corporate life sounds more like the beginning of an examination, and none more dependably marks the end of one.
The Examination Is Never Voluntary
An organization does not look at its operating model because it became curious about itself. It looks because something outside forced it to: a technology shift it cannot absorb, a competitor doing something it cannot match, a quarter that will not survive the earnings call, an investor with a slide deck.
Nothing internal generates the examination. A head of operations who cannot get a price approved in under nine weeks will raise it for two years without producing a review. A competitor repricing its portfolio produces one in a single quarter.
That is because friction gets absorbed. People route around it. They build workarounds, and private knowledge of who to call to make something move, and the reputation for being someone who can get things done in a system where getting things done requires knowing people. The organization becomes genuinely skilled at operating despite its structure, and that competence is what hides the condition. Only an external force large enough to overwhelm the workarounds gets the question asked at all.
The Ritual Has a Vocabulary
What follows is a script so consistent it can be predicted before the first meeting is scheduled. A diagnostic is commissioned, it runs six to ten weeks, and it comes back with bloat, uneven spans of control, duplicated capability, decayed processes, and a talent gap in two or three critical roles.
Every one of those findings is true. Every one of them is also legible: countable, presentable, and addressable by remedies that already exist. So the remedies follow. Mend the broken processes, remove a layer, consolidate the duplicated functions, rightsize.
Rightsizing is the operating-model word for people losing their jobs. The euphemism is doing structural work. It converts a decision into an adjustment, and an adjustment has no author.
Motion is generated. The board is satisfied that action was taken. And it was.
The Diagnostic Stops Where the Remedies Can Reach
Bain studied 57 major reorganizations and found that fewer than one in three produced any meaningful improvement in performance. Some destroyed value. Their own case is Chrysler, which restructured three times in the three years before its bankruptcy, and none of it had much effect.
Two findings from a 2025 McKinsey survey of 757 senior executives sit oddly together. Two-thirds of their organizations had redesigned the operating model within the previous two years, and half expected to do it again inside two more. And 89% were still running a traditional hierarchical structure: a business unit or holding company model, a matrix, or a functional organization. Near-continuous redesign, and almost nothing structurally different at the end of it. The researchers read the gap as a problem of scope, that leaders concentrate on structure and neglect the rest of the system. That is one reading. It is not the only one available.
The more revealing number is newer. In a January 2026 study of roughly a thousand executives and employees, Bain found that 88% of leaders believed their new structure would achieve its goals. Among the employees working inside that same structure, 36% agreed. Only 22% said they had received adequate support to operate in it.
The instinct is to treat that gap as a communication failure. It is not. Both groups are reporting accurately about different things. The leaders are evaluating a design, and the design is usually sound. The employees are reporting on a system, and the system did not change. They understand perfectly well what moved on the chart. What they cannot locate is how decisions now get made, who can approve what, and which of the old routes still work. The distance between 88 and 36 is the distance between a structure that was redrawn and a structure that was rebuilt.
Because the diagnostic never examines the deeper structure that produced the failure. It stops at the level the remedies can reach. Headcount can be cut, so excess headcount gets found. Processes can be mended, so decayed processes get found. Decision rights, incentive architecture, the difference between where authority sits and where the chart says it sits: none of these has a corresponding remedy sitting on the shelf, and so none of them is what the instrument is built to detect.
The Structure Defending Itself
There is another reading available. The operating-model review may not be the organization trying to change. It may be the structure protecting itself from change.
The ritual works because it looks like change. It absorbs external pressure, the analyst note, the board’s impatience, the disruption everyone can see coming, and converts it into visible motion: a new chart, a headcount number, an announcement with a date on it. Enough motion to satisfy everyone watching, while the structure it was meant to change stays exactly where it was. The reorganization is not the cure that fails. It is the antibody that protects the structure from the cure.
Gary Hamel has spent a decade arguing that bureaucracy is innovation-phobic, that its rule-bound power structures frustrate renewal and hobble ingenuity. What the operating-model review adds is subtler than obstruction. It does not resist anything. It absorbs. A defense that rejected the pressure outright would be visible and could be overruled. This one metabolizes the pressure into activity, returns the organism to its previous state, and everyone involved reports the procedure as a success.
Nobody decides this. That part matters. No executive walks into a room intending to stage a change rather than make one. It emerges from rational local choices. The chief executive needs visible action within the quarter. The finance chief needs cost out of the run rate. The advisory firm is scoped to what can be delivered in twelve weeks. Nobody’s compensation depends on redistributing decision authority, and everyone’s depends on the number. Each choice is defensible on its own terms. What they sum to is a defense mechanism. It is nobody’s intention and everybody’s behavior.
Which is why the mechanism needs no defender. It does not require anyone to protect the structure. It only requires that the remedies be chosen before the diagnosis is finished, and they always are, because the remedies are what the organization already knows how to buy. The scope of the review is set by what can be executed, and what can be executed was decided by the last structure. Everything the structure needs to survive is supplied by people acting reasonably.
Seen, Named, and Declared Unchangeable
The most efficient version is the one where the structure is examined properly, named correctly, and then ruled out of bounds, because in that version the organization does the defending itself, voluntarily, and calls it maturity.
The verdict arrives as realism. That is a three-year program and we do not have three years. That is how we are built. It is in the DNA. Denial can be corrected with evidence. This cannot, because it does not dispute the evidence. It agrees with it and files it under things that are true and cannot be acted on. An organization that cannot see its structural problem might yet be shown it. An organization that has seen its problem and defined it as unchangeable has pre-agreed that the one thing that would work is the one thing that cannot be done.
Years ago I watched a chief operating officer, eleven months into a role she had taken because she was told the mandate was to fix the model, walk her executive team through what she had found. The regional P&L structure that made cross-regional collaboration a personal cost to whoever attempted it. Approval thresholds set when the company was a third of its current size. Three people who could unblock any major decision, all reporting to the same executive, none of them able to act alone. It was a good diagnosis. Nobody in the room disputed a line of it.
Then someone said three years, and someone else said DNA, and the chief executive called the analysis excellent and asked what could be done inside the current cycle. Within forty minutes the conversation was about spans and layers. By the end of the session there was a workstream with her name on it that would produce a headcount number by the third quarter. She delivered the number. She left fourteen months later.
The more consequential thing happened to the eleven people who watched her present that diagnosis, and who learned in a single session what naming the structure correctly is worth in that organization. None of them will make that presentation. They are not cynical. They are calibrated.
Structural Sight
The question that would change the outcome is not how do we restructure. It is whether the thing in front of us is a people problem or a structural condition, asked seriously, before the diagnostic is scoped, in the room where the answer can still alter what gets examined. It is a harder question than it sounds, because the honest answer usually implicates the people asking it.
Three things make that question real rather than rhetorical. Before the diagnostic is commissioned, write down which finding would be unacceptable to receive. If nothing appears on that list, the review has already been decided and everything that follows is procurement.
Then name who in the room is authorized to act if the answer comes back as decision rights rather than headcount. If nobody is, the instrument will not produce that finding, because instruments do not produce findings nobody can use.
Finally, require the diagnostic to name one thing that cannot be fixed inside the current cycle, and put it in the board pack anyway, unresolved. An organization that keeps its unchangeable problem written down somewhere visible has not solved it. It has only refused to let it disappear, which is the part of this that is actually within reach.
The cost is not the reorganization that did not work. It is what the organization learned: that it looked straight at the thing producing the failure and was told the thing cannot be changed. The next time someone says the structure is the problem, they will be right, and nobody will move. What gets spent is not the money. It is the organization’s capacity to be told the truth about itself.
