What It Looks Like

If You Are the CEO

The executive mandate without which the structural corrections described in this book are impossible. What the transition requires, what to protect, and four conditions under which you should not attempt it.

The correction in Parts VI and VII requires sustained executive authority. It is subtractive rather than additive: it removes coordination, redistributes decision rights, and makes strategic claims testable. No engineering capability can authorise that change on your behalf.

If an external review recommended better alignment, a new operating model, or a governance framework, recognise that it may have been the most ambitious answer the existing structure could absorb without redistributing authority. It may have been sincerely intended and still structurally incapable of resolving the cause. The transition to autonomous units will eliminate roles, reduce coordination, and redistribute authority. This must be stated plainly, because every previous structural proposal has been presented as additive (a new layer, a new framework, a new governance forum) and this one is subtractive.

Coordination roles become unnecessary: delivery leads who translate between units that cannot talk directly to each other have less work once the translation layer collapses, and programme managers whose authority derives from managing dependencies find fewer to manage once boundaries are explicit and governed by contracts rather than meetings. Narrative authority shifts: unit leads who own a process end to end gain the authority, data, and accountability to describe their own performance, so information that was previously curated arrives unmediated, more precise and less comfortable. The Alignment-Industrial Complex described in Part V, the consulting engagements, framework licences, and governance forums that alignment sustained, loses its market, because an organisation that can read itself does not need a twelve-week diagnostic to understand its own systems. Two things require active protection, and neither will protect itself.

One thing requiring active protection is the transition itself. The coordination layer does not disappear quietly but reassembles through individually reasonable requests: a governance forum that needs to review the new unit's contracts, a compliance requirement that demands an additional approval step, a finance team that insists on monthly reporting in a format that requires the unit to translate its own data into the old structure. Each request is defensible in isolation. Collectively, they reconstitute the coordination overhead the restructuring was designed to eliminate. The governance charter described in Part VI is the structural defence. Protecting it requires sustained executive attention, because the pressure to erode it will be continuous, reasonable, and well-intentioned.

Another thing requiring active protection is the people who can make this work: the engineers, domain experts, and operations specialists who have been carrying the real system in their heads. They are already exhausted, they know which architecture diagrams are fiction and which workarounds have become load-bearing, and they are irreplaceable in the short term and difficult to retain in the medium term. They have been calculating whether the cost of staying is worth it. The transition changes that calculation, but only if they believe the change is real; if they see another operating model that becomes a headcount negotiation, or another governance framework that adds coordination without removing it, they will leave.

The synthesis will occasionally surface a contract or boundary dispute that reaches you instead of being absorbed by programme management. Resolve the boundary rather than reorganising the people around it. Operational visibility is not a request to descend into implementation detail; it is the evidence required to make the small number of structural decisions only you can make. The structural model described in Parts VI and VII is not universally applicable, and applying it in the wrong conditions produces new dysfunction rather than resolving old dysfunction.

The reconstruction fund is capped at twelve months; your commitment cannot be. Plan to protect the resulting structure for at least eighteen months. The coordination layer's immune response begins within weeks; visible results take quarters. An executive who launches the transition and loses interest before the first unit has delivered a full process reconstruction leaves the organisation worse than before: the old structure partially dismantled, the new structure not yet load-bearing, and the people who committed exposed and unsupported. That is a worse outcome than not starting. The question before you authorise the fund is therefore not whether you want the result but whether you will protect the work when protecting it costs you something. If the honest answer is no, the diagnosis holds and the cost of standing still compounds; the timing is yours to judge.

If you are not prepared to redistribute authority, name that before you start, because the redistribution is the mechanism rather than a side effect. The transition moves decision-making from the coordination layer to the units that own processes. Preserve the existing power structure while adopting the vocabulary of autonomy and the result is simulated autonomy, delegated execution under the language of ownership: the same dysfunction reproduced with newer words. The unwillingness does not make the diagnosis wrong; it makes this the wrong instrument until the unwillingness changes.

...

See also: Full contents · Preview chapters · Illusions in the Boardroom