Every governance framework I have ever inherited was originally held together by a person.
Sometimes it was the founder — the one who remembered why the review cadence existed, why the escalation criteria were written that specific way, why Rule 47 had an exception for Rule 23. Sometimes it was a senior architect who had been with the organization long enough to have lived through the failures the rules were designed to prevent. Sometimes it was an operations lead whose judgment had shaped the delivery framework over years of calibration. In every case, the governance existed on paper. But the governance that actually functioned — the judgment that made the rules work — lived in one person.
When that person left, the rules persisted. The judgment did not. The organization continued to follow the rules literally, without the context that made the rules coherent, until the rules started producing outcomes that were worse than no rules at all. Or the organization quietly abandoned the rules, because no one left could defend them, and reverted to whatever ad-hoc governance the remaining leaders improvised.
Nineteen years of program delivery across 10-plus countries — founding the PMO at OpenText, founding the PMO at Full Potential Solutions, directing multi-million-dollar programs at HPE — have put me inside enough organizational transitions to see this pattern as structural rather than anecdotal. Governance that depends on a person does not survive that person. The only governance that survives leadership changes is governance that exists independently of any individual.
This article explains what makes governance portable, what makes it fragile, and why the knowledge-silo failure is the default outcome when governance is not designed for portability from the start.
Why Person-Dependent Governance Fails
Three structural patterns account for nearly every governance collapse I have observed during leadership transitions. They appear in startups, enterprises, government programs, and academic institutions with nearly identical mechanics.
The Institutional Knowledge Evaporation. Governance rules are only useful when the judgment behind them is available. Why does the architectural review panel require three approvers for cross-domain changes? Because two approvers missed a cascading failure in 2019. Why does the deployment window exclude Friday afternoons? Because a Friday deployment in 2017 consumed an entire weekend in recovery. Why does the data migration process require a dry-run against a production-sized dataset? Because a migration that worked in staging corrupted half a production database in 2015.
These are not arbitrary rules. They are compressed lessons from specific failures. But the compression works only if the person reading the rule can decompress it — can understand why the rule exists, when it applies, and when the context has changed enough that the rule needs to evolve. When the person who holds the compressed context leaves, the rule persists in the document. The decompression key does not. New arrivals follow the rule literally without understanding it, or challenge the rule without understanding what it is protecting against.
I observed this at a Fortune-1000 enterprise content vendor where I founded the PMO. The delivery governance framework had been built over three years by an operations lead whose judgment calibrated every threshold, every exception, every escalation criterion. When that person's availability was interrupted, the framework continued to exist. Teams continued to follow it. But within weeks, projects began missing gates that should have been caught, because the rules no longer contained the judgment that had made them precise. The escalation criteria were still in the document. The reading of the criteria — which situations qualified, which warranted exception — had left with the person who had calibrated them. The rules were intact. The governance was not.
The Undocumented Exception Network. Every real governance framework accumulates exceptions. A particular client gets a faster review cycle because of a contractual arrangement from three years ago. A particular codebase skips the integration test requirement because of a legacy constraint that has since been resolved but not formally documented. A particular type of change bypasses the security review because historically those changes have been low-risk. Each exception was reasonable when it was created. Each exception is part of the actual governance framework, even though the exceptions are not written into the formal documents.
The exception network is what makes the formal governance workable. Without exceptions, the framework would be uniform and the friction tax described in the proportional governance model would destroy the system. With exceptions, the framework is proportional — but the proportionality is held in institutional memory rather than documented structure.
When leadership changes, the exception network becomes invisible. The new leader sees the formal governance, not the exceptions that made it function. Either the new leader enforces the formal governance strictly — destroying the operational workability the exceptions provided — or the new leader abandons the governance — because the formal framework without exceptions is unworkable. Either outcome is a structural failure caused by the governance being only partially documented.
At Full Potential Solutions, where I founded the PMO, the first draft of the project governance framework had roughly forty documented rules and, by informal estimate, another twenty-five undocumented exceptions that had accumulated during the PMO's first year of operation. When I transitioned oversight of specific program domains to new program managers, the documented rules transferred cleanly. The exceptions required deliberate transfer sessions — walking through each one, explaining the reasoning, documenting the context that had been missing. The exceptions that were not transferred explicitly reappeared as problems within months, either because they were enforced when they should not have been or abandoned when they should not have been.
The Authority Shadow. Some governance rules function only because a specific person has the authority to enforce them. The rule says "all architectural decisions require approval from the chief architect." This is operable when the chief architect has enough standing in the organization to actually refuse unapproved work. When that standing is attached to a specific individual — earned through years of relationships, political capital, and track record — the authority transfers to the role only partially. A new chief architect has the title but not the earned authority. The rule still exists. The enforcement does not, because enforcement depends on authority that has not yet accumulated to the new incumbent.
When I directed multi-million-dollar programs at HPE, this pattern appeared in subtle ways during program lead transitions. A program lead who had been in the role for three years had accumulated relationships with delivery teams, executive sponsors, and external partners that gave them the latent authority to enforce governance decisions without escalation. When that program lead moved and a new lead took the role, the governance rules were identical. The authority behind them was not. Decisions that had been final under the previous lead became subject to negotiation. The formal structure had not changed. The operational reality had.
These patterns compound when they interact. Institutional knowledge evaporates, taking the exception network with it, and the new authority figures do not have the standing to rebuild the exception network or enforce the remaining formal rules. Within a year of a significant leadership transition, an organization that had functioning governance can find itself with a framework that exists on paper and is followed nowhere in practice.
The problem is not the leadership transition itself. Transitions are normal. The problem is that governance built around people does not contain the portability mechanisms that would allow it to survive the transition.
The Structural Portability Properties
Governance that survives leadership changes has specific structural properties. These properties are not about the content of the governance. They are about how the governance is encoded and enforced. Governance with any content can be portable if the encoding is structural. Governance with the best content can be fragile if the encoding is person-dependent.
Enforcement Embedded in Tooling, Not Authority
The first property is that enforcement lives in the system, not in a person. The previous article on phase gates explained how structural gates work — the system produces the verification evidence, the human reviews the evidence, and the system cannot be talked past. The critical portability implication is that when the human changes, the enforcement does not degrade. The new reviewer sees the same evidence the previous reviewer saw. The new reviewer may decide differently, but they decide based on the same structural input. The enforcement mechanism does not depend on accumulated authority or institutional knowledge — it produces the evidence, and the evidence is what the review is about.
In the DIOSH framework I built across the 18-venture portfolio, this shows up as hook-enforced blocking. A pre-tool-use hook intercepts file writes that would violate gate requirements. The hook does not know who is doing the writing. It does not have accumulated authority. It checks the gate conditions and either permits or refuses the write. When the operator changes, the hook behaves identically. The governance is portable across operators because the enforcement is not operator-dependent.
Lessons Captured as Loadable Artifacts
The second property is that the judgment behind the rules is captured as artifacts that can be loaded by any operator, not as tacit knowledge held by specific individuals. The governance memory layer in DIOSH contains lesson entries that record why a rule exists, what failure it prevents, and when it applies. A new operator working on a task loads the relevant lessons automatically. The loading is not optional and not dependent on remembering to check. The structural design ensures that the judgment that shaped a rule is available to anyone who encounters the rule, regardless of whether they were present when the rule was created.
The academic frame is useful here. Institutional knowledge in most organizations has the properties of tacit knowledge — it lives in practice and relationships, cannot be easily articulated, and transfers only through apprenticeship. Portable governance requires converting that tacit knowledge into explicit knowledge — documented, structured, and loadable by anyone with access. The conversion is work. Organizations that fail to do the conversion end up with governance that works only as long as the apprenticeship chain remains intact. One skipped generation of apprenticeship, and the tacit knowledge is gone.
Rules That Explain Themselves
The third property is that each rule contains the context required to apply it. A rule that says "require three approvers for cross-domain changes" is fragile. A rule that says "require three approvers for cross-domain changes because cascading failures in cross-domain changes historically go undetected with fewer approvers, as demonstrated by the 2019 incident in which a two-approver change broke three downstream services over a weekend" is portable. The new operator does not need to have been present for the 2019 incident. The rule contains its own context. The judgment required to apply it — when it applies, when an exception might be warranted, when the rule needs to evolve — is accessible from the rule itself.
This sounds verbose in the abstract. In practice, it is the difference between documents that can be handed to a new operator and documents that require a three-month apprenticeship to decode. The verbosity is the portability mechanism.
Exception Networks Made Visible
The fourth property is that the exception network is documented, not held in institutional memory. Every exception to a formal rule is recorded with the conditions under which it applies, the reasoning behind it, and a review date after which the exception needs to be re-validated. The new operator inherits both the rules and the exceptions. The proportional governance the exceptions enable continues to function across the transition.
This is structurally uncomfortable because exceptions feel like admissions of inconsistency. Formal governance tends to pretend exceptions do not exist. But exceptions always exist — the question is only whether they are documented or hidden. Hidden exceptions produce fragile governance. Documented exceptions produce portable governance. The discomfort of admitting that the framework has exceptions is a small price for the portability the documentation provides.
Operational Evidence
Scale. The DIOSH governance framework operates across 18 ventures under HavenWizards 88 Ventures OPC with a single operator — a setup that forces every portability property to be structural rather than personal. The framework cannot depend on my continuous presence because I am not continuously present on any single venture. The session for a venture I have not touched in weeks must produce the same governance outcomes as the session for a venture I worked on yesterday. The only way this works is if the governance is in the system, not in my working memory. The 18-venture operation is, in effect, a continuous stress test of the portability properties.
Recovery. When I joined the Australian digital agency network during the turnaround engagement, the prior delivery governance had been held together by individuals whose judgment had calibrated the processes over years. Those individuals had transitioned out. The rules persisted without the judgment. The agency was running at -20% to -60% losses in part because the governance framework had become person-dependent and the dependent person was no longer there. The recovery rebuilt the governance with the structural portability properties — enforcement in tooling, lessons as loadable artifacts, rules with embedded context, documented exceptions. Profitability reversed to +40% to +60%. The recovery was not just a turnaround of operations. It was a reconstruction of governance as portable rather than personal, which is the structural precondition for operations to be recoverable at all.
Prevention. In the PMO I founded at a Fortune-1000 enterprise content vendor, I had the unusual experience of building governance that I knew would outlast my involvement. The question "what happens when I am not here" was a design input from day one. Rules were written with their context embedded. Lessons were captured as artifacts that could be loaded by any future PMO lead. Exceptions were documented with their reasoning and review dates. The governance was not pretty. It was longer and more explicit than frameworks that could assume continuous expert presence. But when my direct involvement ended, the framework continued to function — not because of heroic effort by the remaining team, but because the framework had been designed to function without depending on any specific person.
Compounding. Across the 18-venture portfolio, every governance failure during a portability test — a session where governance produced a wrong outcome because a specific lesson had not been loaded, or a rule had been applied without its context — becomes input to the portability infrastructure. The lesson gets captured. The rule gets expanded with its context. The exception gets documented. Over hundreds of sessions, the portability properties have strengthened with use. The strengthening happens automatically because each failure becomes structural input rather than an operational incident.
Where This Does Not Apply
Structural portability is not universally the right answer. Several contexts warrant different governance design.
Highly specialized domains with long apprenticeships. Some fields — certain legal specialties, certain medical specialties, certain engineering disciplines — have governance that genuinely requires tacit knowledge accumulated through long training. The knowledge cannot be fully converted to explicit artifacts, because the judgment requires pattern recognition developed through experience. In these domains, person-dependent governance is not a failure mode but a structural reality. The portability mechanism is the apprenticeship chain, not documentation. Attempting to encode everything structurally would flatten the judgment in ways that damage the governance.
Small teams with stable personnel. A small team whose members have worked together for years and expect to continue working together benefits less from portability infrastructure. The overhead of documenting every rule with its context, every exception with its reasoning, every lesson with its applicability, is real. If the team composition is stable, the tacit governance works well enough that the portability overhead is not justified. The calculation changes if composition stability can no longer be assumed — but in genuinely stable configurations, person-dependent governance can be locally optimal.
Organizations in early formation. Before an organization has enough governance to require formal portability, investing in portability infrastructure is premature. Early-stage governance is correctly improvised, discovered through trial, and held loosely. The portability properties matter when the governance is worth preserving across transitions. Governance that has not yet been validated by operational use is not yet worth preserving, and imposing portability infrastructure on pre-validation governance can ossify rules that should still be evolving.
Contexts where institutional continuity is not a goal. Some work is structurally temporary — a defined-end project, a specific-crisis response, an initiative that will conclude regardless of outcome. Governance for temporary work does not need to survive leadership changes because there is no institutional continuity to protect.
These are boundaries. Portable governance is the right design for governance that is meant to survive. It is not the right design for every governance situation.
The Principle
Governance is a system that must function when the people who built it are no longer there.
If your governance depends on institutional memory, it is fragile. If it depends on accumulated authority, it is fragile. If it depends on exception networks held in the heads of senior operators, it is fragile. Each of these forms of fragility is invisible while the dependencies are intact, and visible only when the transition happens and the governance fails to survive it.
The diagnostic question is simple. If everyone currently operating the governance framework left tomorrow, and new operators arrived with access to the documents but no access to the people, how much of the governance would still function? If the answer is "most of it," you have portable governance. If the answer is "very little, because the rules only make sense if you were here when they were written," you have personal governance dressed up as institutional governance.
Structural portability is not glamorous. It requires rules that are longer because they contain their context. It requires lessons captured as artifacts when no one feels like capturing them. It requires exceptions documented when documenting them feels like admitting inconsistency. The overhead is real and, in the moment, often feels unnecessary.
It feels unnecessary right up until the transition, when the invisible dependencies become visible by failing. Build governance that can be inherited. The inheritance test is the only one that matters, because leadership changes are not hypothetical — they are the one thing every organization encounters eventually.
