Reducing Dependency on the One Person Everyone Asks

Every team has one and every team describes them as a great asset. They are, and they are also a structural risk that gets treated as a cultural strength, which is why nothing is done about it until they hand in their notice.

"One resignation and we're blind."

How one person becomes the route

It is a feedback loop with no natural brake. They answer well, so they get asked. Being asked deepens their context, so they answer better. Better answers attract more questions. Within a year they are the route to most of what the team knows, and this was produced entirely by helpfulness.

Two things accelerate it. Speed: people route to whoever answers fastest rather than whoever is most appropriate. And uncertainty about ownership: when nobody knows who owns an area, the default is the person who always knows, which is covered in ownership clarity for distributed teams.

What it costs

Four costs, and only the last one is usually discussed.

Their output. A day spent being an index is a day not spent on the work only they can do. Their own contributions move to early mornings and evenings, which is the pattern to watch for: the senior engineer who was never truly offline.

The team's speed. Every question routed to them queues behind every other question routed to them. On a distributed team, add a night's delay per queue position.

Everyone else's growth. If the fast answer is always available, nobody develops the context to answer it themselves. The dependency deepens by preventing its own cure.

The resignation risk. The one that gets attention, and it is the least urgent of the four, because it is hypothetical while the other three are happening this week.

Why they cannot fix it themselves

This matters, because the usual response is to suggest the person set better boundaries.

Consider their position. Declining to answer a colleague's question is visibly unhelpful, and helpfulness is a large part of how they are valued. Nobody's performance review credits them for the interruptions they refused. Meanwhile the cost they are paying is invisible: their own output looks fine because they are making it up in the evenings.

So the incentives point entirely at continuing, and asking them to change is asking them to accept a reputational cost to fix a structural problem they did not create. The intervention has to come from their manager.

Separate retrieval from judgement

Sort a week of their incoming questions and the split is stark.

Type	Share	Needs them?
Retrieval: where is it, what was decided, was it tried	The large majority	No
Judgement: should we, is this wise, what would you do	A minority	Yes, and this is the value

Retrieval is the volume and judgement is the value, and the current arrangement spends their attention overwhelmingly on the first. That is the inefficiency to attack, and it is attackable without asking anybody to be less helpful.

With StandIn, they spend ninety seconds at the end of each day confirming a brief, mostly pre-drafted from the work they already did. While they are off, including during focus blocks and meetings, their StandIn answers questions from that brief in their words, with a source under every answer, clearly labelled, and it never guesses: when the answer is not there it says so and names who to ask.

They keep the expertise and lose the interruptions, and their name stays on every answer because it came from what they wrote. That last point matters more than it might seem, since the status of being the person who knows is real and worth preserving. See StandIn for leadership teams.

What to do in the first month

Four moves, none of which require the person to change their behaviour.

Week one, measure: ask them to tally questions for five days, marking each as retrieval or judgement. Week two, publish an ownership list so routing failures stop landing on them by default. Week three, route the next three pieces of work in their area to someone else with them reviewing rather than implementing. Week four, look at what still arrives and ask whether the retrieval half needed a person.

And say out loud, to the team, that this is the goal. The person in this position often assumes the load is a sign they are valued and that reducing it is a demotion. It is worth being explicit that the opposite is intended.

Common Questions

How do you reduce dependency on one key person?

Separate retrieval from judgement. Make the retrieval answerable without them, publish ownership so questions route correctly, and deliberately give some of their work to others with them reviewing. Asking them to be less available does not work.

Is it a problem if one person knows the most?

Expertise is not the problem; being the only route to it is. The test is what happens when they take two weeks off. If work stops, you have a dependency rather than a specialism.

Will this person feel sidelined?

Only if it is done silently. Say what the aim is, keep their name on the answers that come from their work, and give them back the time for the work only they can do. Most people in this position are relieved.

How quickly can this change?

The retrieval share can drop within a month, because it depends on a habit rather than on knowledge transfer. Judgement takes quarters, because it moves only by other people making calls and being reviewed.