The executive director broke her wrist on a Saturday, which sounds like nothing. She was back at her desk on Wednesday, typing slowly, mildly embarrassed by the fuss.

What happened in between is the part worth looking at. On Monday morning a contract renewal needed a signature and nobody was sure whether the deputy could sign it. By Monday afternoon somebody had found the bylaws, which addressed the question of officer authority in general terms and did not address this. On Tuesday a county partner called about an incident and was told, twice, that the person who handles that is out. A vendor payment that should have gone Monday went Thursday.

Nothing broke. Four days of a wrist, and the organization did not lose a contract or fail a family. But every single person involved spent that week discovering what they did not know, and the board learned — from an accident, over a weekend — that it had no answer to a question it had been meaning to get to.

Succession planning is a duty most boards hold and almost never discharge. It gets deferred because it feels like a vote of no confidence in a person everybody likes, and because there is never a quarter in which it is the most urgent thing. So it waits for an event, and the event is not usually a resignation. It is a wrist.

Three different questions wearing one name

Most organizations conflate three separate things, produce a document that serves none of them, and conclude that succession planning does not work.

Emergency succession answers: if this person is unreachable tomorrow, who decides what, with what limits, until the board meets? It is short, it costs one board meeting, and it is the only one of the three that is urgent.

Planned departure answers: when someone gives notice, what does the board do first, and in what order? It matters, but it has a runway measured in months.

Bench development answers: are we building people who could do these jobs in three years? It is the most valuable and the slowest, and it is the one organizations start with, which is why the other two never get done.

Do the first one this quarter. It is a single agenda item, and it is the one that protects you from the wrist.

The dependency is wider than the executive director

Here is the part that surprises boards. Ask, for every role and not just the top one: if this person were unreachable for a month, what stops?

The answers cluster in predictable places, and almost none of them are the executive director. The person who knows the payroll exceptions. The one administrator on the system that holds the case records. The single staff member who holds a credential a program is required to have. The bookkeeper who is the only signer besides the director. The person who knows which county contact actually returns calls.

That last category is the hardest to see and the most common. It is not a documented duty, it does not appear on an organization chart, and the person holding it usually has no idea it is unusual — because it has never once occurred to them that it is not obvious to everybody else.

So the exercise has two halves. First, who: name a person for every function, including the functions nobody currently owns. Then, what: ask each person directly what they do that nobody else could pick up, and sort the answers into what is written down, what could be written down in an afternoon, and what is genuinely tacit. The middle category is cheap to fix and almost always the largest.

What an emergency plan actually contains

Not much, and it should fit on two pages that somebody can find when the systems are down.

Who acts, and who acts if that person is also unavailable. What the acting appointee may decide — a spending limit, whether they may hire or dismiss, which contracts they may sign, whether they may speak publicly. What they may not decide. How many days before the board reviews the arrangement. Who must be notified, and when. Where a role carries a license or a regulatory designation, what that specifically requires, cited and confirmed with your own counsel rather than assumed — because an emergency appointment that is not permitted is worse than none.

And then it gets adopted by resolution, which is what makes it a decision the organization has already taken rather than a document somebody drafted.

What to do this month

Put emergency succession on the next board agenda as a single item, with the two-page plan drafted in advance so the meeting is a decision rather than a discussion. Name the acting appointee, write the limits, adopt it.

Then, separately and without the board, ask three people the question: what do you do that nobody else here knows how to do? Write down what they say. You will not fix it this month. But you will know, which is the entire difference between a dependency and a surprise.


The executive transition and key-person risk plan is on the Open Shelf: three kinds of succession kept separate, the key-person map, the single-point-of-knowledge audit, an emergency acting appointment with stated limits, and the ninety-day interim protocol. The licensed-role section must be checked against your own license before adoption. No registration, no email address.