How to Read a NOFO: A Compliance Checklist for Nonprofit Grant Applicants

You have a funder in mind and a document in front of you. It might be a federal notice of funding opportunity, or a foundation's guidelines page, or a portal that will not show you its questions until you have made an account. Somewhere in it is everything you have to do.

The instinct is to start writing. The better first move is to stop and sort, because that document is not really prose. It is a list of requirements wearing prose, and the requirements are not all the same kind of thing.

Four kinds of requirement, and why the difference decides your week

Almost everything a funder asks for falls into one of four categories.

Eligibility gates are binary. You meet them or you are out, and no amount of good writing changes the answer. Your organization's age, your state, your budget size, whether you have a fiscal sponsor, whether they accept unsolicited requests at all.

Scored criteria are graded. A reviewer reads what you wrote and assigns it a value. This is where your argument actually lives.

Organizational facts are possessions. A determination letter, a board roster, an audit, a policy document. Either you have it or you need to make it, and neither is a writing problem.

Narrative is the custom argument you build for this funder — the part everyone thinks of as "the grant."

Sorting a requirement into the wrong category is the most expensive mistake available to you, and it is easy to make in both directions. Treat a scored criterion as a checkbox and you write two flat sentences where the reviewer expected your best case. Treat a gate as something you can argue with and you spend a weekend on an application that was decided before anyone read it.

The tell is usually grammar. A gate is written in the imperative and admits no degree: applicants must be located in. A scored criterion is written in graded language: strength of, extent to which, demonstrated. When the verb has a dial on it, someone is going to turn that dial.

The portal tells you more than the guidelines do

Here is the part most guides miss. The same sentence can be any of the four, and the published guidelines often will not tell you which.

"Demonstrate community support" might be a checkbox. It might be a file upload asking for letters. It might be a 500-character text area. Those are three completely different tasks, and the difference is invisible in the PDF.

So when the portal is reachable, open it before you write anything. The input mechanism is the most honest statement of what is actually wanted. A checkbox means the funder wants a fact. A file upload means they want a document you either have or do not. A character-limited text area means they want an argument, and the limit tells you how long an argument they intend to read.

If the portal is not reachable until you start an application, start one and do not submit it. Read the fields. Then close it and go plan.

Building the register

Once you can see the requirements, make a list with one row per requirement and four columns: what it is, which of the four kinds it is, where the answer comes from, and whether you have it yet.

That last pair matters more than it looks. "Where the answer comes from" forces you to name a source — this budget, that board minutes, our program director's head — before you write a sentence that depends on it. Requirements whose source is a person's memory are the ones that turn a fifteen-minute task into a two-hour hunt, and finding them on Tuesday is much better than finding them the night before a deadline.

Work the gates first, all of them, before writing anything. They are fast and they are the only category that can end the exercise. Then the organizational facts, because those have lead times you do not control. Then the narrative, which is the only part that actually needs you at your desk with your best thinking.

Some requirements are conditional, and flat lists get them wrong

A register that reads "independent audit: required" is often wrong, because whether you need one frequently depends on a combination of where the money comes from, where you are incorporated, and how much revenue you had.

State thresholds differ, and — this is the part that catches people — so does what they are measured against. California's Nonprofit Integrity Act requires an independent audit at $2 million or more in gross revenue. New York's threshold is more than $1 million in revenue and support. Pennsylvania sets its audit requirement at $750,000 in annual contributions, with a review or audit required between $250,000 and $750,000.

Read those three again and notice that they are not the same measurement. An organization with $900,000 in total revenue, most of it earned rather than contributed, can be above one state's line and below another's on the same set of books. "Do we need an audit" is not a question with a single answer, and your own number is not one number.

Other states differ again, and some set no statutory threshold at all. Check yours rather than reasoning from a neighbour's.

Federal rules moved recently too. The 2024 revision to the Uniform Guidance (2 CFR 200), effective for fiscal years beginning on or after October 1, 2024, raised the single-audit expenditure trigger from $750,000 to $1 million.

So write conditional requirements as conditions, not as facts. If we receive this award, and we are incorporated here, and our revenue exceeds that, then an audit is mandatory. An organization sitting just under a threshold is exactly the one a flat list will mislead.

The same shape applies to fiscal sponsorship, which is not a status you check off but a pivot that moves the whole compliance burden onto your sponsor. And charitable-solicitation registration is genuinely fragmented — roughly forty states have statutes, with no national standard and no working universal form, so the answer is per-state rather than general.

Checking your draft against the register

When you have a draft, do not reread it looking for typos. Walk the register instead, row by row, and ask a narrower question of each: where, specifically, is this satisfied?

Point at the sentence. If you cannot point at one, the requirement is not met, however good the surrounding paragraph is. A requirement that feels addressed across three paragraphs is usually addressed in none of them, and a reviewer working through a checklist will not assemble it on your behalf.

This is also where you catch the quiet ones — a required attachment nobody named in the narrative, a question inside a compound sentence that got half an answer, a number that appears twice with two different values.

This is not really about federal notices

Federal notices are the clearest example because their structure is standardized, so the four categories are easy to see. But nothing in the method depends on that.

A two-page foundation guidelines page has gates, scored criteria, facts, and narrative in it too, just less visibly labeled. A portal with fifteen fields has them, spread across fifteen inputs. Small funders in particular tend to publish very little and then ask for quite a lot inside the form, which is precisely why opening the form early pays for itself. With foundations there is also a gate that comes before any of this: whether a foundation will even accept your application in the first place.

The habit is the same everywhere: sort before you write, work the gates first, name a source for every claim, and check the draft against the list rather than against your memory of the list. And if the first step the funder asks for is a letter of inquiry, it helps to know what to have ready before your first LOI — most of the facts the register will ask you to source come from that same short list of documents.

One more thing worth saying

If the gates rule you out, that is a good outcome, not a wasted afternoon. The worst version of this work is a strong application to a funder who was never going to consider you — closed to unsolicited requests, wrong geography, a minimum grant size well above what you asked for.

Finding that out on Tuesday costs you twenty minutes. Finding it out after submission costs you the weekend you spent, and the opportunity you did not pursue instead.