Program, admin, fundraising: the overhead question

Every nonprofit eventually has to split its expenses three ways: program, management and general, and fundraising. It shows up on the annual return, in the audited statements, and in nearly every grant application.

It is usually treated as a compliance exercise done once a year under time pressure. It deserves better, because the numbers it produces get read as a judgment on the organization.

What the split actually is

Some costs belong clearly to one function. A program staff member’s salary is program. The annual gala’s venue is fundraising. The audit fee is management and general.

Most costs are not like that. Rent, utilities, insurance, software, the executive director’s time, the finance function itself: these serve everything, and the split is a judgment about how much of each. That judgment is the allocation method, and having one is the whole exercise.

Why the overhead ratio is a bad single measure

The percentage spent on program versus everything else has become a shorthand for whether an organization is well run, and it is a poor one.

The mechanics are the problem. Two organizations doing identical work with identical results can report very different ratios based on reasonable differences in how they allocate shared costs. An organization that pushes hard on the ratio can improve it by allocating more aggressively toward program, by underinvesting in its own finance and technology, or by not counting volunteer-supported functions. None of those made it better at its mission. One of them made it worse.

Meanwhile an organization that invests in the systems that let it report accurately and manage well shows a higher administrative percentage for doing the more responsible thing.

How costs actually get split

A defensible method uses a driver that has some real relationship to how the cost is consumed.

• Time for people costs. Not a guess at year end, but some record of where effort actually went. It does not require timesheets to the minute; it does require something more than a feeling.
• Square footage for occupancy, when different functions genuinely occupy different space.
• Headcount or seats for costs that scale with people, like some software and insurance.
• Direct program costs as a base for spreading genuinely general costs proportionally.

The method does not have to be sophisticated. It has to be reasonable, applied consistently, and written down.

The starvation trap

The pressure to report a low overhead percentage pushes organizations to underfund exactly the things that make them capable: financial systems, competent back office staff, technology, planning capacity.

The cost of that shows up later and somewhere else. Reports that take three weeks. A funder relationship strained by a reporting problem. An audit finding. A key person leaving with the only working knowledge of how anything is tracked. None of that appears in the ratio that the underinvestment was protecting.

This tension is real and it is not resolved by pretending the ratio does not matter to funders. It does. But it is worth deciding deliberately where your organization sits, rather than drifting toward the lowest possible number by default.

Being able to defend your method

The question that comes from an auditor, a sophisticated funder, or an attentive board member is always the same: why this number and not another one.

The answer should be a short written description of the method, the driver it uses, when it was set, and when it was last reviewed. That document is not bureaucracy. It is the difference between an allocation that is a position and one that is a guess.

If you have never done this deliberately

Most organizations arrive at their first real method after a few years of approximation, and the approximations were not wrong so much as undocumented.

Setting it up properly is a finite piece of work: decide the drivers, apply them to the current year going forward, write the method down, and review it when the organization changes shape. It is much easier to do in a quiet month than during an audit, and once it exists it mostly runs itself.