Key person dependency is the condition where a plan's handoffs, knowledge or decisions run through one individual, so that their absence stops work nobody else can pick up. Network research finds that ideas, information and influence concentrate in a few people who bridge groups, and turnover research finds that when one of them leaves, colleagues follow.
What does key person dependency actually look like?
Not a hero on the org chart but a person at a crossing point. The planner who is the only working link between sales forecasts and production. The engineer who understands the legacy system. The manager whom two teams route decisions through. The plan never names them, which is why the dependency survives every review until the person is away.
Burt's research on structural holes explains why these people exist. Opinion and behaviour are more similar within groups than between them, so the person connected across groups sees options that others cannot. In the electronics company he studied, compensation, good performance evaluations, promotions and good ideas were disproportionately in the hands of people whose networks spanned those holes. The same position that makes someone valuable makes the organisation dependent on them, because the bridge they provide has no second span.
Why do new plans create key person dependencies?
Because a new plan creates new handoffs, and the first version of every new handoff runs through whoever understood both sides. A strategy launch, an integration or a system change asks people to connect work that was previously separate, and the few who can do it become the route. The dependency is a by-product of the plan working, early on.
Cross, Rebele and Grant's data, gathered over two decades, show that the time managers and employees spend in collaborative activities has ballooned, and their work describes how that demand lands unevenly on the people others turn to. A plan that adds handoffs adds that demand, and it adds it to the same people, because they are the ones who already know how the pieces fit. Six months in, the plan has a shape nobody designed: a small number of individuals through whom most of the new coordination flows, each of them busier than the plan assumed.
What happens when the key person leaves?
The work they bridged stops, and the people around them start looking. Felps and colleagues, studying 45 bank branches and over a thousand hospitality departments, found that colleagues' job embeddedness and job search behaviour predicted individual quitting beyond what other factors explained. Turnover travels through the same network that made the person central.
Porter and Rigby's review of 42 studies on turnover contagion concludes that employees tend to imitate the turnover-related attitudes and behaviour of their co-workers, so one resignation raises the probability of the next. For a plan this means the loss is rarely one person. It is the person, the handoff they carried, the time it takes a successor to rebuild the relationships on both sides of it, and a raised risk that the colleagues who relied on them leave too. None of that appears in a risk register that lists the role but not the dependency.
How do you find a key person dependency before it bites?
Ask about handoffs, not job titles. For each handoff the plan depends on, name who gives, who receives and who else could do it tomorrow. Where the third answer is nobody, you have found a single point of failure. Then look at the calendars: the people most in demand across teams are usually the ones the plan quietly depends on.
A formal network analysis will show the same thing with more precision, but the handoff question can be asked in a leadership meeting in an hour. The useful output is a short list: five to ten people whose absence would stop a part of the plan, the specific work that would stop, and a view on how exposed each one is, through workload, through an approaching change or through a market that wants them. That list changes what the leadership team does next, which is the point of finding it early.
How do you reduce the dependency without losing the person?
Add a second span to the bridge. Give the handoff a named deputy who does it some of the time, write down what the key person knows that nobody else does, and move some decisions that route through them to the people who could take them. Protect the person's workload while this happens. Overloaded bridges are the ones that leave.
The temptation is to reward the key person and move on. Reward matters, but it does not reduce the dependency, and it can increase the load, because recognition brings more requests. The structural fix is to change what flows through them. That is slower and less visible than a retention bonus, and it is the only one of the two that still works on the day they hand in their notice.
Where this stops
This note is about dependencies on individuals that a plan creates or exposes. It does not cover succession for the top roles, the legal side of key person insurance, or how to design a network analysis. The studies cited come from specific settings, a US electronics company, a bank and a hospitality firm among them, and describe tendencies. They cannot say which person in a given plan is the one it cannot lose.
How Atlas reads it
Atlas treats a new dependency, including a dependency that has quietly come to rest on one person, as one of the six moments that should prompt a reading of the organisation behind the plan. The reading sets what the plan now demands against what the organisation presently supports, and one of the questions it asks of every critical handoff is concentration: whether it is a single point of failure.
Where it is, the recommendation follows the intervention ladder: clarify which handoffs the plan cannot lose, resequence where one person is the bottleneck for several, reassign the decisions that route through them, and reach questions about individual roles last. Each judgement carries the evidence for it and what argues against it, and is bounded to this plan and this horizon, never a prediction about the person.
Sources
- Ronald S. Burt, Structural Holes and Good Ideas, American Journal of Sociology (2004)
- Will Felps, Terence R. Mitchell, David R. Hekman, Thomas W. Lee, Brooks C. Holtom and Wendy S. Harman, Turnover Contagion: How Coworkers' Job Embeddedness and Job Search Behaviors Influence Quitting, Academy of Management Journal (2009)
- Caitlin M. Porter and James R. Rigby, The turnover contagion process: An integrative review of theoretical and empirical research, Journal of Organizational Behavior (2021)
- Rob Cross, Reb Rebele and Adam Grant, Collaborative Overload, Harvard Business Review (2016)
Questions
A situation where a plan's handoffs, knowledge or decisions run through one individual, so that their absence stops work that nobody else can pick up. It usually forms around people who bridge groups, and it rarely appears on an organisation chart, because the chart shows roles and the dependency lives in the handoffs between them.
List the handoffs the plan depends on and ask, for each, who else could do it tomorrow. Where the answer is nobody, there is a single point of failure. Then check who is most in demand across teams, because the people everyone turns to are usually the ones the plan quietly relies on.
Research says it can. Felps and colleagues found that colleagues' job embeddedness and job search behaviour predicted individual quitting, and Porter and Rigby's review of 42 studies concludes that employees tend to imitate the turnover-related attitudes and behaviour of their co-workers. The risk is highest around people others depended on, because their departure changes everyone's work.
It buys time, not resilience. A bonus may keep the person, but it leaves the plan as dependent on them as before and can add to their load as recognition brings more requests. The structural response is to add a deputy to the handoff, document what only they know, and move some of the decisions that route through them.
Because being the bridge between groups attracts work. Burt's research found that ideas, evaluations and promotions concentrate in people whose networks span structural holes, and Cross, Rebele and Grant document how collaborative demand has grown and lands on the people others turn to. The position that makes someone valuable also makes them the default route.