An asset produces a return whether or not a particular person turns up. Anything else is an arrangement, and it will be priced as one by whoever looks at it hardest.
A chief executive takes three weeks' leave and returns to a queue of decisions nobody was willing to make. The usual reading is that the business needs them. It is evidence of something, but not of value: a queue at one desk describes the operating model, not a compliment.
The reading that matters is a buyer's, a lender's or an insurer's. Diligence looks for exactly this and prices it: which revenue depends on a named individual, which approvals only clear when one person touches them, which customer relationships would reopen the week after a departure. The exercise takes days and the discount is rarely negotiable, because whoever applies it has seen the concentration tested.
Most boards have never run that exercise on themselves. The question has two halves, and the second gets skipped. Which parts of the enterprise would lose material value the day one named individual left — and what is that concentration costing now, while everybody is still here?
The Strategic Context
Key-person concentration is usually filed as a human resources matter or a continuity risk, which places it with the people least able to act on it. It belongs on the capital agenda. It shows up in the multiple a buyer will pay, in credit terms and covenants, in insurability, and in the strategies an enterprise can pursue — a business whose decisions queue behind one calendar cannot take on work requiring many decisions at once.
It is also quiet. Concentration generates no incident until it generates a total one, so it never competes for attention against problems making noise now. Its costs are continuous but unbilled: latency, foregone options, judgement other people never develop.
And it correlates with success rather than failure: the enterprises most dependent on individuals are usually those where the individual is genuinely very good, which is why the dependency is mistaken for strength, defended as culture, and left to compound.
Concentration Reads as Commitment
The first misreading treats this as an insurance problem, solved with a named deputy on a chart and a policy in a drawer. That covers the discontinuous event — the departure, the illness — and leaves the continuous cost untouched. Most of the value concentration destroys goes while everyone is still employed.
The second assumes the dependency is knowledge. Some of it is, and that part is tractable. The larger part is judgement applied to cases that do not repeat, and relationships held in a person's name. Neither transfers through documentation: a document is not a capability, and an enterprise that confuses the two produces a great deal of paper and changes nothing.
The third reads concentration as loyalty, when it is a design property the organisation created and continues to reward. Where indispensability earns standing, capable people will make themselves indispensable. They are responding correctly to the incentive.
The fourth is the leader's own. Personal indispensability feels like evidence of contribution and is better understood as a ceiling on enterprise value. It is also the exposure least likely to be recorded: the risk register is drafted by people who report to the person concerned, and none will write that sentence.
Reframing the Issue
Stop asking what happens if they leave and start asking what it costs while they stay. The first question produces a contingency plan that is filed. The second produces a number, and numbers move capital.
An asset is a claim on future cash flows that does not require a named person to appear. Where output requires one, the enterprise holds an arrangement, and every external party who examines it will say so in the only language that carries: a lower multiple, tighter covenants, a retention clause, an earn-out. The board can run that valuation first, internally, while it still has a choice.
How knowledge is captured and carried between teams is a discipline in its own right, treated elsewhere. The question here is narrower: where the concentration sits, what it costs, and what it does to the enterprise's worth.
Where the Dependency Actually Sits
Five classes behave differently and respond to different remedies, and treating them as one problem is why most attempts stall.
Relationships are held in a person's name — customers, regulators, financiers, suppliers. Judgement covers the unusual case: the pricing exception, the quality call, the acceptance of a residual risk. Method is how the work is genuinely done, as distinct from how it is described. Authority is the approvals that clear only when one person touches them. Identity is hardest: the customer believes they are buying the individual, and in some professional models they are.
Each has its own resolution. Relationships need second and third connections established before a renewal, not after a resignation. Judgement needs the exception decided in company and the reasoning written at the time, because reasoning reconstructed later is a rationalisation. Method needs codification, discussed below. Authority needs delegated limits genuinely used — a governance change, not a training one. Identity may require repositioning what the enterprise sells, or accepting the constraint and pricing it honestly, provided that is decided rather than drifted into.
The last two classes usually belong to the most senior person in the building. They are also the two that never appear on the register.
The Cost Before Anybody Leaves
Latency comes first. Work queues at one desk, and the queue lengthens with growth, so the concentration binds hardest exactly when the enterprise is trying to expand. The scale ceiling follows: an organisation can grow only as fast as its indispensable people can inspect it.
Then the opportunity cost of the leader's own time. Hours spent adjudicating are hours not spent on capital allocation, direction, or the relationships nobody else can hold — rarely counted, because adjudication feels productive.
Capability atrophy is slower and more expensive. A team that never decides does not learn to decide, so the dependency deepens on its own, and each year the successor pool is thinner than the year before.
Finally, governance distorts. Someone indispensable accumulates an informal veto appearing in no delegation schedule, and disagreements with them resolve by avoidance rather than on the merits — a statement about how the organisation handles disagreement [Related article: Your Organisation Has a Default Conflict Style. Who Chose It?].
Build the Standard from the Best Performer, Not from the Ideal
Process documentation fails in two predictable ways. The ideal process, written by someone who does not do the work, is followed in the meeting where it is presented and nowhere else: compliance becomes performance and the real method stays private. The average process, assembled from what most people do now, codifies mediocrity in writing.
The alternative is to build the standard from the best performer's actual workflow. Watch the person who gets the best result and record what they do — the sequence, the timings, the checks, the steps they skip and why, and above all the points at which they stop executing and start deciding.
This does three things. It produces a standard that is demonstrably achievable, because someone in the building achieves it — the difference between a document people follow and one they cite. It converts private judgement into inspectable practice, revealing where the concentration lives: usually three or four decision points nobody had articulated. And it establishes a baseline, since a process never described cannot be improved.
Three cautions belong with it. The best performer's method is credible, not optimal; improvement comes after codification, not instead of it. Essential steps must be separated from personal habit, which is what the observation is for. And expect resistance, because the method is the performer's standing — an incentive the enterprise built and can rebuild, by making whoever codifies their work more valuable.
Codify repetitive, high-frequency execution first. Judgement-heavy work stays with the people accountable for it: standardising a decision taken twice a year removes the thinking and keeps the paperwork. And a responsibility matrix showing two accountable names against one outcome has recorded an unresolved argument, not a control.
Once a standard exists, output can be compared against expectation and performers judged on something other than impression. Whether those measures ever change a decision is a separate and more sceptical discipline [Related article: How Many of Your Workforce Metrics Have Ever Changed a Decision?].
Mastery Is Proven by Succession, Not by Performance
If a senior manager's excellence is judged by personal output, retaining the critical work is the rational response. The concentration is not a character flaw; it is the measurement system operating as designed.
Invert the test. Treat a manager as proficient only when the function survives their absence, and make a tested successor a precondition of promotion rather than an addition to the file. Mastery is demonstrated by producing someone who can do the work — an inversion that costs nothing to state and changes what every ambitious person optimises for.
Transfer happens in unremarkable stages: demonstrate, supervise, let the person work independently and report immediately, move to periodic reporting, then have them develop their own successor. Two rules do the real work. Accountability does not transfer at the rate the task does, so the delegator remains answerable until the handover is formally recognised. And the test is the exception, never the routine — anyone can be handed the ordinary case; the concentration lives in the other one.
One diagnostic is worth keeping. If you find yourself chasing updates, you have assigned a task, not delegated. Delegation transfers the authority to act with the obligation to report; assignment transfers neither, and the follow-up burden returns to the person it was meant to free. The opposite failure is handing work over while dropping accountability — abandonment, which produces the incidents that make leaders reluctant to try again.
All of this assumes the successor wants the role. What an individual seeks from work — technical mastery, autonomy, general management, security — determines whether a succession plan is real or only drawn [Related article: Which Anchor Does This Role Actually Require?].
Decision Framework
Do not attempt to de-risk everything. Score each critical role on two axes: materiality, the share of revenue, margin or delivery capability exposed; and substitutability, how long restoration would take and at what cost. Act where both are high, monitor where one is, leave the rest. Estimates come from the executive, not the incumbent's own team.
| Dependency class | The test to run | What actually resolves it |
|---|---|---|
| Relationship | Which counterparties would reopen terms within a quarter of a departure? | Second and third relationships built before renewal |
| Judgement | Who decides the exceptions, and is the reasoning recorded? | Exceptions decided in company, with reasons written at the time |
| Method | Could a competent newcomer reach standard output from what is written? | Codification drawn from observing the best performer |
| Authority | Which approvals only clear when one named person touches them? | Delegated limits genuinely used, with sampling in place of inspection |
| Identity | Does the customer buy the person or the enterprise? | Repositioning the offer, or accepting the constraint and pricing it honestly |
From Strategy to Execution
The immediate work is an inventory of roles rather than people, with the two scores attached and a check of leave records, since untaken leave is a concentration signal available free. Expect the list to be shorter than feared and more senior than hoped.
Medium-term, two capabilities need building. The first is codification as a routine function — observing an expert working and producing a usable standard — rather than a project run once after somebody resigns. The second is a change to promotion criteria, so a developed successor is evidence of readiness. Delegated authority limits belong here, and count only when the delegating leader stops reviewing what they delegated.
The long-term position is a matter of enterprise worth. Saleability, financing terms and strategic optionality all rest on operating independence from named individuals, and that independence takes years to build. A business worth more when its founder is not in the room has not been diminished. It has been finished.
Signals to Monitor
Untaken leave, particularly at senior level. Approvals that route around the appointed deputy back to the incumbent. Customer contracts naming an individual rather than a capability. Internal candidates declining senior roles, an accurate statement about what those roles cost. Documentation produced after an incident and never opened again.
Watch also for onboarding times lengthening year on year, escalations arriving at one desk rather than through the structure, and the first diligence question from a lender about key people — that question is the market pricing your concentration before you have.
Questions for the Leadership Team
- Which roles, not which people, would cost us material revenue or delivery capability within a quarter of a departure, and who assessed that?
- Which approvals clear only when one named person touches them, and what is the queue behind that desk worth in a year?
- Whose method are our documented processes built from: our best performer's actual workflow, or somebody's account of the right way to work?
- What would a buyer discount in this business today, and would we contest that discount or concede it?
Closing Perspective
Concentration is rarely anyone's doing. It accumulates because it is convenient, is defended because it is effective, and is discovered when something forces the question — an illness, an approach from a competitor, a transaction, or an ambition the enterprise turns out to be unable to pursue.
Leaders who wait discover their concentration at the worst available price, in a conversation they do not control, with a counterparty who has already done the arithmetic. The alternative is to price it now, while the people concerned are still present and can help dismantle what they built.
That last part is the real test. A senior leader who can construct something that makes them steadily less necessary, and who reads that as the achievement rather than as a loss of standing, has understood what they were hired to build.
About the author
Kevin Jogin is Founder & Principal Advisor at EraNorth. Meet the Founder.