Key-Person Risk Is Not One Risk: Five Dependencies Boards Should Separate

A board can agree that a founder, CEO, technical leader, or commercial operator is unusually important and still have no shared account of what would fail if that person became unavailable. The concern tends to arrive as one large shape: we depend on her; he knows everything; the relationships are his. That language conveys seriousness while making the work harder to assign.

The first useful move is separation. What information would become inaccessible? Which calls depend on experienced interpretation? Where does approval stop because authority has not moved? Which stakeholders respond to the individual rather than the office? Which meetings, handoffs, and informal reconciliations happen because one person quietly keeps the system together?

Each answer points toward a different continuity mechanism. The dependencies can overlap inside one decision, but they do not disappear through the same intervention. A full interview may leave authority untouched. A perfect org chart may say nothing about coordination. A contact list cannot transfer trust.

Information dependency exists when material facts, records, commitments, definitions, or access paths are known to one person and are not reliably available to others. Some of this is mundane: a private spreadsheet, an undocumented renewal date, a supplier’s actual commercial terms. Mundane does not mean harmless. The simplest omission can stop work at an inconvenient moment.

The response is usually institutional: recover the record, establish ownership, document provenance, make access durable, and create a maintenance routine. Search and retrieval matter as much as storage. A fact that technically exists but cannot be found by the person making the call is weak continuity.

Walsh and Ungson’s organizational-memory framework is useful here because it treats memory as acquisition, retention, and retrieval. The board should therefore ask more than whether something was written down. Who knows where it is, when it applies, whether it is current, and what retrieves it at the moment of need?

Two executives can read the same customer history and reach different conclusions about whether a concession is ordinary, dangerous, or overdue. The difference may sit in cue recognition, comparison with prior cases, an expectation about what happens next, or a threshold learned through costly experience.

Judgment can be represented, challenged, and developed, though never captured as a perfect internal algorithm. Consequential incidents, decision timelines, contrast cases, and scenarios reveal more than a list of principles. The aim is to make distinctions visible enough for another person to practice and for the organization to identify where escalation remains necessary.

Research showing persistent manager effects across firm policies gives boards reason to take individual influence seriously. It does not establish that tacit judgment caused every difference, nor that preserving an incumbent’s style is always desirable. Some inherited judgment should be questioned as strategy and conditions change.

A named successor does not necessarily possess the right to act. Material decisions may still be routed to the departing leader because approval matrices, bank authorities, customer expectations, board habits, or the incumbent’s informal veto have not changed. The vacancy is filled on paper while operational authority remains elsewhere.

Authority requires explicit transfer: decision rights, limits, escalation routes, signing power, access, and visible backing from the board or owners. The outgoing executive also has a role. If every difficult question continues to return to them, the organization teaches stakeholders that the new authority is conditional.

Some personal authority cannot be reassigned by resolution. A successor earns it by making calls, accepting consequences, and behaving consistently. The structural part should still be moved cleanly so that personal credibility can form around a real mandate rather than a borrowed title.

Relationship dependency concerns history, obligations, influence, credibility, and the way a particular person changes what others are willing to say or do. Context can be transferred through stakeholder maps, joint meetings, and careful introductions. Personal trust must be earned again. Treating those as the same thing invites the predecessor to remain permanently between the successor and the relationship.

Coordination dependency appears when a person connects pieces of work that the formal design leaves apart. They reconcile the sales promise with operating capacity, know which disagreement needs a private call before the meeting, or translate between a technical team and a board. Hutchins’ account of distributed cognition is a warning against locating all competence inside the individual: performance may arise from a system of people and artifacts, with the critical person acting as one essential junction.

The response may be a recurring forum, a new integrator role, clearer interfaces, shared artifacts, or a redesign of the workflow. Documenting the coordinator’s habits helps, but the organization still needs someone—or some dependable routine—to perform the coordination after departure.

Once the dependencies are separated, the continuity plan becomes inspectable. Information has an owner and retrieval test. Judgment has cases, practice, and escalation boundaries. Authority has formal rights and a visible transfer date. Relationships have context, introductions, and space for the successor to earn confidence. Coordination has an operating mechanism that does not rely on the former leader’s private intervention.

The categories are a practitioner diagnostic, not a validated risk score. Their value is in forcing different questions and preventing a single artifact from being mistaken for complete preparation. A repository can be excellent and the transition still exposed. A well-liked successor can have strong relationships and no authority to resolve the underlying issue.

The Map is appropriate when the organization needs to locate these dependencies around a particular critical role. The Atlas becomes relevant when the resulting material must remain owned, retrievable, and maintained by the institution. Legal, tax, compensation, search, and fiduciary questions remain with the specialists responsible for them.