A section-by-section board paper skeleton with guidance for each part: the decision sought, background held to one page, options with the rejected ones shown, financials, risks and a recommendation that takes a position.
Directors read papers late, in bulk and under time pressure, and most board papers are written as if the opposite were true. This skeleton gives every section a defined job and a length limit. The whole template sits on this page, with no form in front of it.
Last reviewed 3 July 2026 · Free and ungated
Before the paper reaches the pack, selected senior operators from the Global Board read it the way your directors will and report, confidentially, on where it fails.
Pressure-test this decisionA board paper is not a record of the work done; it is an instrument for a decision. The director reading it has eleven other papers in the same pack, no context on your programme's history, and a well-developed instinct for what is being smoothed over. Papers fail when they are written to demonstrate effort. They succeed when a sceptical, intelligent stranger can find the decision, the cost and the main risk in the first minute.
| Section | What it must contain | Guidance and length |
|---|---|---|
| Decision sought | One sentence a company secretary could minute verbatim: what is being approved, for how much, effective when | First on the page, before any context. If this sentence resists being written, the paper is not ready |
| Executive summary | The recommendation, the cost, the single biggest risk and why now, in that order | Half a page. Written last, by one author, in one sitting |
| Background | Only what changed to make this decision necessary now | One page, hard limit. History that does not change the decision belongs in an appendix or nowhere |
| Options considered | Two to four live options including doing nothing, each with its genuine rejection reason | A rejected option described in one flattering line was never really considered, and directors know it |
| Financial implications | Full-life cost, funding source, benefit profile and the downside case, signed by finance | One page plus a table. A base case with no downside is a forecast, not an analysis |
| Risks and how they are held | The three or four risks that could sink the proposal, each with a named owner and a costed mitigation | Half a page. Ten minor risks listed in a grid signal that the real ones were left out |
| Recommendation | The sponsor's actual position, stated plainly, with the consideration that most nearly changed it | Three sentences. A paper that ends with "the board is invited to note" is asking for attendance, not a decision |
| Conditions and review point | What the board can attach to approval, and the date or event at which the decision returns | A quarter page. This section is what makes approval verifiable rather than ceremonial |
| Appendices | The evidence a challenging director will ask for: model assumptions, diligence summaries, consultation record | Referenced from the body by number. Nothing load-bearing may live only in an appendix |
Write the decision-sought sentence first and the executive summary last; everything between exists to survive the questions those two provoke. The sponsor owns the paper personally. Papers assembled by a working group read like negotiated settlements, and the hedging that creeps in during internal review is precisely what directors are paid to notice. Circulate five working days before the meeting: a paper that arrives with the pack gets skimmed, and a skimmed paper gets deferred.
The author of a board paper cannot read it as a director will, because the author's knowledge fills every gap invisibly. When the decision is large or hard to reverse, the strongest pre-submission test is a cold read by senior operators who have sat on boards and approved, deferred and killed papers like this one. They will find the unanswered question in minutes, and it is far cheaper to hear it before the meeting than during it. A client brief to the Global Board does exactly this, confidentially.
Six to eight pages before appendices, following the length limits above. Beyond ten pages the paper starts defending itself rather than informing a decision, and the probability of deferral rises with every additional page.
They work in opposite directions. This template is for the author constructing the paper; the checklist is for directors and the chair verifying it deserves a vote. A paper written against this skeleton should pass the checklist without surprises.
Treat repeated requests for background as a signal, not an instruction. Boards ask for context when the decision sentence is vague or the options feel pre-cooked. Fix those two sections before adding pages of history that nobody will read.