A policy brief is a short document that gives an official enough information to make a decision. The way to write one is to lead with the recommendation, keep it to roughly two pages, and cut every sentence that does not change what the reader should do. Most briefs fail not because the research is weak but because the research is all the reader gets.
Officials, staffers and agency managers read in minutes, not hours. A brief that opens with background forces them to hunt for the point. One that opens with the decision, the cost and the deadline gets read to the end. The craft is subtraction: the analysis lives in an appendix or a full report, and the brief carries only what changes the choice.
This guide sets out the structure, the wording rules and the common failure modes. It treats the brief as a working document in a legislative or agency setting, where the reader may be a committee staffer, a program head or a cabinet officer. The broader context of how such decisions get made — how a rule is written, how a bill moves — is covered elsewhere in our policy coverage. We covered a connected angle in The Congressional Review Act: 60 Senate Days to Kill a Rule.
What is a policy brief, and who is it for?
A policy brief is a two-page decision document that states a problem, an option set and a recommendation for a named reader. It is not a summary of a report. The test is simple: if the reader can act on the brief alone, it works. If the reader must open the full study, it is a cover memo. Readers following this should also see Medicaid Section 1115 Waivers: How States Rewrite the Program. We covered a connected angle in Medicaid Section 1115 Waivers: How States Rewrite the Program.
According to Merriam-Webster, the first sense of the verb "write" is forming characters or words by marking on a surface; being the author of, or composing, appears as a later sense (2b(1)). The distinction matters here. A brief is composed for one reader and one decision. Before drafting, write one sentence naming that reader and that decision. Every later choice — length, tone, what gets cut — follows from it.
A brief for a congressional staffer facing a markup deadline looks different from one for an agency program manager planning next year's budget. The staffer needs the vote consequence. The manager needs the funding and staffing implication. Same research, different documents.
What structure do decision readers expect?
Most working briefs follow a five-part structure. Officials scan for these parts in order, and a brief that buries one of them gets set aside.
- Purpose line. One sentence: what decision this brief supports and when it is needed.
- Bottom line up front. The recommendation in the first paragraph, with its single strongest justification. No throat-clearing.
- Background. Three to five sentences of facts the reader must know to judge the recommendation. Anything the reader already knows gets cut.
- Options. Two or three, stated neutrally, each with its main cost and main risk. A brief with one option is an argument, not a brief.
- Recommendation and next step. What to do, who does it, and what happens by when.
Tables work well for the options section: one column per option, rows for cost, risk and legal authority. Readers compare across columns faster than across paragraphs. The same discipline applies to the document's own framing — a brief that reads like advocacy gets discounted, while one that shows the trade-offs gets trusted.
How should the wording be cut and tightened?
Write the first draft long, then cut it by a third. The cuts follow a fixed order: background first, caveats second, adjectives last. Keep the recommendation and the numbers.
Three wording rules carry most of the weight. First, one idea per sentence, and sentences short enough to read aloud without losing breath. Second, concrete nouns over abstractions — "three state agencies must certify by March" beats "interagency coordination challenges remain." Third, plain verbs: "decide," "fund," "delay" instead of "leverage," "address," "facilitate."
Tools can help with the mechanical pass. DeepL's writing assistant offers style and tone adjustment alongside grammar and spelling checks, which can catch the register drift that creeps into drafts written under deadline. But no tool substitutes for the structural cut. A brief that opens with the recommendation survives any editing tool; a brief that opens with context fails regardless of polish.
Define terms of art once, in parentheses, then use them without apology. A reader who knows the term is not slowed by a one-line gloss. A reader who does not is stopped cold without one.
What makes officials stop reading a brief?
Four failure modes account for most discarded briefs, and all four are avoidable in drafting.
- The buried recommendation. The decision appears on page two, after methodology. Readers form a view in the first paragraph; if the point is not there, they move on.
- The single-option brief. Presenting one course of action reads as advocacy. Officials discount advocacy unless they already agree with it.
- The unpriced recommendation. A recommendation with no cost, no risk and no legal authority attached is a wish. Name the trade-off even when it is uncomfortable.
- The hedge stack. Every claim qualified, every number footnoted to a caveat. State the confidence level once, plainly, and let the appendix carry the nuance.
There is a fifth, quieter failure: the brief that ignores process. A recommendation the reader cannot enact — because the committee calendar, the appropriations cycle or the rulemaking clock forecloses it — wastes the reader's time. Timing is part of the analysis. A proposal that needs a legislative vehicle should say which one, in the same way that knowing how committee gatekeepers shape a bill's path changes what a lobbyist asks for.
What does a finished two-page brief look like?
In assembled form, the discipline shows up as proportion. The recommendation and its justification take the top half of page one. Background takes four sentences. The options table takes a third of page two. The closing paragraph names the next step and the date it is needed by. Nothing else appears.
Our analysis of where briefs go wrong is consistent on one point: the failure is almost always structural, not stylistic. A writer who fixes the order — decision, background, options, recommendation — can write in plain prose and be understood. A writer with the order wrong cannot edit enough to save the document.
What remains unknown in any given case is the reader's own constraints: the calendar, the politics, the budget room. A good brief states what it does not know in one line, rather than pretending certainty it cannot support. That line builds more trust than a page of confident prose.
