Nonprofit staff retention: why staff leave

The program manager who resigns in March has usually been leaving since November. By the time she gives notice she has a new job, a rehearsed reason, and no interest in a difficult conversation on her way out the door, so she says it was about the money. Sometimes it was. More often the money is just the answer that requires the least explanation. Nonprofit staff retention is mostly decided months earlier, in a hundred small moments nobody logged, and a board that only hears the exit-interview version will keep drawing the wrong conclusion.

Boards tend to respond to turnover with a salary conversation, because a salary conversation is one a board knows how to have. It is worth having. It is rarely the whole story.

What actually drives nonprofit staff retention

Ask people who left small nonprofits why, once they are far enough away to be candid, and four things come up more than pay.

Unclear expectations. The job was described in one paragraph and grew by accretion. Nobody can say what success looks like this quarter, so the employee is never finished and never quite adequate. This is the most fixable item on the list and the most commonly ignored.

Unsustainable workload. Not a hard week; a hard year, structurally. Two people covering three roles because the third was never filled, or a program that grew 40% while staffing stayed flat. People will absorb this for a surprisingly long time, and then they leave suddenly, which is why it reads as a shock to everyone above them.

No path forward. At a nine-person organization there may genuinely be nowhere to be promoted to. That is a real constraint, but it is not the same as no growth, and organizations that lose people to it usually offered nothing at all: no training budget, no new responsibility, no conversation about what someone wants to be doing in three years.

Hearing about decisions after they land. A funding change, a program shift, a schedule change, learned about in the same email that announced it. This one is corrosive out of proportion to its apparent size, because it tells people they are staff to be informed rather than people to be consulted.

Sector workforce research from the National Council of Nonprofits has tracked vacancy and turnover pressure across small organizations for years, and the pattern holds: compensation matters, and it is not the only thing, and organizations that treat it as the only thing keep losing people after they fix it.

What a board can influence, and what it can't

The line matters, because a board that reaches past it does damage even with good intentions.

A board genuinely controls the budget, which means salary bands, benefits, paid time off, whether a vacant position gets funded or quietly absorbed, and whether there is a training line at all. It controls how much churn it creates from above: a board that changes strategic direction every 18 months, or asks for a new report three days before every meeting, is a workload driver whether or not it thinks of itself that way. And it hires, supports, and evaluates one person, the chief executive, whose management practice is the single largest factor in whether anyone stays.

A board does not supervise anyone else. It does not resolve individual grievances, decide who gets promoted, or address a specific manager's behavior. Those belong to the chief executive, and a board that takes one of them on has stopped governing and started managing, which is a much harder mistake to walk back than it is to make.

The practical version: the board's lever is the chief executive, plus the budget. Almost everything a board can usefully do about retention runs through one of those two.

How to hear from staff without undermining the chief executive

Here is the delicate part. A board that hears directly from staff is routing around the person it hired to run the organization. Done casually, it is destabilizing: it tells staff there is an appeal route above their boss, it tells the chief executive the board has doubts it has not voiced directly, and it puts the board in possession of information it cannot act on without managing.

Done well, it is one of the more useful things a board can do. The difference is entirely in the conditions.

Agree it with the chief executive in advance, and ideally have them ask for it. This is not a courtesy. A staff survey the executive learns about secondhand is a vote of no confidence regardless of what the results say.

Ask about the organization, not about a person. The questions should cover conditions of work, not the performance of named individuals, and everyone should know that before they answer.

Report in aggregate, always. The board sees averages and themes across the team. It never sees who said what, because there is no mechanism to find out, and nobody should ever be asked to follow up with whoever raised something. On a team of seven, that promise is the only thing standing between an honest survey and a polite one.

Say plainly that this is not a grievance channel. If someone has a complaint about a specific person, it goes to their supervisor, the chief executive, or whatever the personnel policy says, and the survey does not replace any of that.

Give the chief executive the results at the same time as the board, not after. They should never be surprised in a room by something the board read first.

This is a different exercise from a board looking at its own performance, and it is different again from evaluating the executive, though the results inform that. Boards that blur the three end up doing all of them badly, and the distinction between assessing the board and evaluating the executive is the one worth being clearest about.

What gets asked, and what it takes to see results

The staff questions cover five areas: role clarity, leadership communication, growth and recognition, culture and belonging, and retention confidence. It is deliberately narrow. There are no items naming a manager, and no open-ended catch-all inviting someone to say anything at all, because that is the shape of a complaint box rather than a survey.

Results are withheld until two conditions are both met: at least 4 responses, and at least 60% of the roster. Both, not either. A team of six needs 4. A team of 12 needs 8. Below that the section is not shown at all, not shown partially, and not summarized loosely, and the reason is arithmetic rather than policy: with three answers out of six, averages start pointing at people. The same logic sets the thresholds for board and community results, and it is not adjustable by anyone, including the board.

Tell your staff the threshold before they answer. It is the credible part of the confidentiality promise, and the one people actually check.

Asking and then doing nothing is worse than not asking

This is the honest limit, and it is worth being blunt about it.

A staff survey is a request. It asks people to spend social capital telling their employer something uncomfortable, on the implied promise that it goes somewhere. If the board reads the results, says "this is really helpful," and takes no visible action, the staff have learned something specific: that the organization gathers information about them and does nothing with it. The next survey gets fewer responses and softer ones, and the team is now worse off than if nobody had asked. Restoring that takes years.

So do not run one until you can answer three questions. Who reads the results, on what date, and in what meeting. What gets communicated back to staff, by whom, and how soon. And what the board is prepared to actually change if the answers are unwelcome, given that the two things it controls are the budget and the chief executive.

If the honest answer to the third is nothing, that is worth knowing before you send anything, and it does not mean the situation is hopeless. It might mean the board's real work this year is a budget conversation, or a candid one with the chief executive about workload, and that both can happen without a survey.

If you do run one, the assessment collects staff responses anonymously, holds them until the threshold is met, and reports them alongside what the board and the chief executive said. It's free, it takes a few minutes per person, and there's nothing to buy.

Turnover at a small nonprofit is usually read as a personnel problem, one departure at a time. It is more often a design problem: roles that grew without being redrawn, decisions that arrive finished, and a workload nobody sized. Those are all things a board can see, if it is willing to ask in a way people can safely answer.