The transition logic establishes the order in which the structures must be built. It does not establish how to begin. An organisation that has understood that the authority design comes first, that the instruments depend on it, and that the structures must be absorbed before they are extended still faces the question of what to actually do in the period before any of the structures exist — the period in which the organisation is still operating in its current condition and must begin the change without the foundations that the change will eventually rest on. That period has a natural horizon. It is the first ninety days, and what happens in it determines whether the transition is founded correctly or founded on the same defects it is meant to correct.
The first thing to be clear about is what the first ninety days are not. They are not the period in which the governance architecture is built. The governance architecture cannot be built in ninety days, and any plan that promises to build it in that time is promising the comprehensive transformation that the preceding chapter warned against. The authority design alone — the foundational structure on which everything else depends — is the work of considerably longer than ninety days, because it requires confronting and redistributing decision rights across an organisation that has reasons to resist the redistribution. The first ninety days are not the construction of the architecture. They are the establishment of the conditions that make the construction possible. The distinction is the entire content of the chapter, because an organisation that spends its first ninety days trying to build the architecture will fail, and an organisation that spends them establishing the conditions for building it will have founded the transition correctly even though, at the end of the ninety days, almost nothing of the architecture itself will yet be visible.
The condition that must be established first is diagnostic, not constructive. Before an organisation can reform its authority design, it must know where its authority design is actually defective — where accountability has been assigned without the decision rights that discharge it, where ownership is symbolic rather than operational, where the questions enter and do not exit. This is the ownership velocity reading applied diagnostically, before any of the instruments that normally produce it are in place. It can be done without the instruments, through direct examination: taking the questions the organisation is currently failing to close and tracing each one to the position that should have closed it, then asking why that position did not. The answer, in case after case, locates a specific defect in the authority design — a holder without decision rights, an accountability without authority, a question that belongs to no one. The first weeks of the transition are spent producing this diagnosis, because the diagnosis is what tells the organisation where its authority reform must be directed. Without it, the reform is general, and a general reform of authority is a reorganisation, which produces disruption without precision.
The diagnosis produces a map of where genuine holding is absent, and the map directs the first constructive act of the transition, which is narrow and specific rather than broad. The organisation does not reform its entire authority design in the first ninety days. It selects a small number of the defects the diagnosis has located — the ones where the cost of the absent holding is highest and the difficulty of supplying it is lowest — and it repairs those. Repairing a defect in the authority design means giving a specific holder the specific decision rights their position requires, establishing the cost-absorption position that lets them own the consequences of their decisions, and ensuring they have access to the signal their decisions depend on. This is done for a few positions, not for all of them, because the purpose of the first ninety days is not to complete the authority reform but to demonstrate that it works — to produce a small number of positions that genuinely hold, so that the organisation can see what genuine holding produces and can build the rest of the reform on the evidence of the first repairs rather than on the promise of a transformation it has not yet seen.
The second constructive act follows from the first, and it is the introduction of the decision infrastructure for the domains where the repaired holders now operate. Once a few holders genuinely hold, the decisions they make are binding, and binding decisions can be held. The decision infrastructure introduced in the first ninety days is not the comprehensive decision layer for the whole organisation. It is the decision layer for the repaired domains — a place where the genuine closures the repaired holders are now producing can be recorded, held current, and made accessible. This is deliberately small. It covers only the domains where the authority reform has reached. But it produces, within the first ninety days, the first instance of the property the whole framework exists to produce: a domain in which questions are closed by holders with genuine authority and the closures are held in infrastructure that delivery can act on. That property, present in even a few domains, is the proof of concept on which the rest of the transition is built. The organisation can point to those domains and say: this is what decisive operation looks like, and it is producing the velocity the framework promises, and it is doing so here, now, in our organisation, not in a description of someone else’s.
What the organisation must resist in the first ninety days is the pull toward breadth. The instinct, once the first repairs produce results, is to extend immediately — to reform more positions, to expand the decision infrastructure, to begin installing the pulse and the measurement across the whole organisation. This instinct is the enemy of the transition, because breadth in the first ninety days comes at the cost of the absorption that the structures require. The few repaired positions need time to operate before more are added, because the organisation is learning, through them, how genuine holding works, and the learning cannot be rushed. The small decision infrastructure needs to prove itself in the repaired domains before it is extended, because its extension to domains whose authority has not been reformed would fill it with the symbolic agreements that the unreformed authority produces, and the infrastructure would lose, in the extension, the property that made it valuable in the repaired domains. The first ninety days are an exercise in deliberate narrowness, and the discipline of staying narrow against the pull toward breadth is the hardest discipline the period requires.
At the end of the first ninety days, the organisation’s governance architecture is not complete, and it is not meant to be. What it is, if the ninety days have been spent correctly, is correctly founded. There exists a diagnosis that locates the defects in the authority design precisely. There exist a small number of positions that genuinely hold, repaired against the diagnosis, demonstrating that the reform works. There exists a small decision infrastructure that holds the genuine closures those positions produce, demonstrating that the closures can be captured and held. And there exists, in the domains the reform has reached, the first real instance of decisive operation — questions closed by genuine holders, closures held in infrastructure, delivery acting on current decisions. None of this is the full architecture. All of it is the correct foundation for the full architecture, and the difference between an organisation that has this foundation at the end of ninety days and an organisation that has spent the same ninety days building instruments on an unreformed authority design is the difference between a transition that will succeed and one that has already failed without yet knowing it.
The ninety-day horizon matters for a reason beyond convenience. A transition that shows nothing for longer than ninety days loses the organisation’s belief, and belief is a resource the transition cannot do without. The organisation undertaking the change needs to see, within a period it can hold in view, that the change produces something real. The few repaired positions and the small decision infrastructure provide exactly that — visible, real results within the horizon the organisation can sustain its attention across. This is not a concession to organisational impatience. It is a structural requirement of the transition, because the transition depends on the organisation continuing to invest in it, and the organisation will only continue to invest in what it can see is working. The first ninety days produce the visible proof that sustains the investment through the much longer period in which the rest of the architecture is built. An organisation that understands this designs its first ninety days to produce that proof, and an organisation that does not understand it designs its first ninety days to build as much of the architecture as possible, produces nothing visible because the architecture is not yet complete enough to function, and loses the belief that the longer transition requires.
The ninety days have a shape, and describing the shape is useful provided it is understood as a structural sequence rather than a project schedule. The opening weeks are diagnostic, and they produce the map of where genuine holding is absent. This work is not glamorous and it is tempting to compress, but compressing it is the most common way the ninety days are wasted, because a transition directed by an incomplete diagnosis repairs the wrong positions and produces results that do not demonstrate what the transition needs to demonstrate. The diagnosis must be thorough enough to locate the defects precisely, and the weeks it takes are not a delay before the real work but the work on which everything that follows depends. An organisation impatient to begin building should understand that the diagnosis is the beginning of building, and that the weeks spent producing it are the weeks that make the subsequent repairs land where they will produce the most proof for the least difficulty.
The middle weeks are the first authority repair, and they are the weeks in which the transition becomes visible. The few positions selected from the diagnosis are repaired — given the decision rights they lacked, the cost-absorption position they needed, the signal access their decisions required — and the holders in those positions begin, for the first time, to genuinely hold. This is the moment the transition produces its first real closures, and it is also the moment the transition produces its first resistance, because the authority given to the repaired positions was taken from positions that held it without exercising it, and those positions feel the loss. The middle weeks are therefore the weeks in which the transition’s sponsor matters most, because the resistance that the first repairs produce must be held against, and holding against it requires a sponsor with the authority to ensure that the repairs are not reversed the moment they are resisted. An organisation attempting the transition without a sponsor who can hold against the resistance the first repairs produce will find the repairs undone, and the ninety days will end with the authority design exactly as it was, because every repair was reversed by the resistance it provoked.
The closing weeks are the establishment of the first decision infrastructure and the accumulation of the first held closures. The repaired holders are now producing genuine closures, and the closing weeks build the infrastructure that holds them — small, covering only the repaired domains, but real, holding decisions that are binding because the holders who made them have genuine authority. By the end of the closing weeks the organisation has, in a few domains, the complete chain that the framework produces: questions closed by genuine holders, closures held in infrastructure, delivery acting on current decisions. The metrics to watch through these weeks are not compliance metrics but the early form of the outcome metrics — whether the repaired domains are closing questions faster than they did before the repair, whether the closures are being held and accessed, whether the delivery teams in those domains are encountering current constraints. These metrics, even in a few domains, are the proof, and watching them through the closing weeks is how the organisation confirms that the foundation it has laid is producing what the foundation is supposed to produce.
What “correctly founded” means at the end of the ninety days can now be stated concretely, against the shape the ninety days have taken. It means the organisation possesses a thorough diagnosis of its authority defects, a small number of genuinely repaired positions demonstrating that the reform works, a small decision infrastructure holding the closures those positions produce, and the early outcome metrics confirming that the repaired domains are operating with more velocity than they had. It does not mean the architecture is built, the authority comprehensively reformed, or the instruments fully installed. It means the organisation has proven, in its own context, with its own people, that the framework produces results — and it has proven this in a form small enough to have been achievable in ninety days and real enough to sustain the investment that the much longer remainder of the transition requires. The organisation that reaches this state has not finished the transition. It has founded it correctly, and the difference between a correctly founded transition and an incorrectly founded one is the difference between a transition that will succeed over the years it takes and one that has already failed in its first ninety days without knowing it.
The first ninety days, founded correctly, produce the proof. They do not produce agreement. The reform of even a few positions in the authority design takes decision rights from some positions and gives them to others, and the positions that lose authority do not experience the loss as the correction of a defect. They experience it as a diminishment, and they resist it. The resistance is not incidental to the transition. It is a structural feature of it, because the transition redistributes authority, makes deferral costly, and removes the ambiguity that some positions have learned to use as a resource — and the positions affected by these changes are often the most powerful in the organisation. The pattern of that resistance, and the structural rather than political responses that address it, is the subject the next chapter takes up.