A Board Paper Is the Record of the Decision, Not the Case for It
Most board papers are written as advocacy.
An executive has reached a view, the paper explains why the view is correct, and the recommendation sits at the end as a conclusion the reader has been led to.
That's a natural way to write, and it misunderstands what the document is for.
A board paper isn't primarily read at the meeting. It's read afterwards, by people who weren't there: auditors, a regulator, a successor board, a court, an acquirer's diligence team.
What they're looking for isn't whether the decision was right. It's whether the board understood what it was deciding.
That distinction has practical force, because the two documents look different.
A persuasive paper presents the recommended course and the reasons for it.
A paper that functions as a record presents the decision, the alternatives considered and why they weren't preferred, the assumptions the recommendation depends on, what happens if those assumptions are wrong, and what the board is being asked to accept.
The first is shorter and reads better. The second is what a governance review looks for, and its absence is what a governance review finds.
This is frequently misread as a call for caution or for longer papers, and it's neither.
A board that approves an ambitious course after seeing a genuine downside has made a stronger decision, on the record, than one that approved a modest course on a paper showing only upside.
Governance isn't a bias toward inaction. It's a requirement that the risk actually taken was the risk the board knew it was taking.
One necessary caution. Directors' duties, the status of minutes and the remit of committees are governed by legislation and listing rules that differ between jurisdictions. Nothing here is legal advice, and the applicable law and the company's own constitution always govern.
A governance review is not asking whether the decision was right. It is asking whether the board knew what it was deciding — and only the paper can answer that.
Write the Minute First, Then Write the Paper That Makes It True
The most useful discipline in drafting a board paper is to write the minute first. Not the secretary's wording, but the substance: what must this minute record, and does the paper make each element of it accurate?
A minute saying the board considered the proposal and approved it records almost nothing and protects nobody.
A minute recording what was considered, what alternatives were examined, what the principal risks were understood to be, and what conditions attached is a record. Every element has to be traceable to something actually in front of the board.
This inverts the drafting order productively. Working back from the minute exposes immediately what a paper is missing, because each clause has to point at something.
If the minute is to record that the board considered alternatives, the paper must set out alternatives. Not a token pair of obviously inferior options, but the courses genuinely available, including doing nothing.
The option of doing nothing deserves emphasis, because it's the alternative most often omitted and most often asked about later.
Almost every proposal has a counterfactual where the company doesn't proceed, and that counterfactual has consequences which are frequently not neutral. A position deteriorates, an opportunity closes, a competitor moves.
Conditions are regularly under-served. Boards frequently approve subject to something: a valuation confirming, a diligence item closing, a consent obtained, a price not exceeding a limit.
Where those are set out in the paper, the minute records them precisely and the delegation is clear. Where they're agreed in discussion, the scope of the delegation depends on a sentence written afterwards.
What a well-formed minute commonly needs to be able to record, and what the paper must therefore contain. Working backwards from this list exposes most of what a draft is missing.
-
What was proposed, in terms specific enough to bind
The transaction, the amount, the counterparty, the term. A minute recording approval of a proposal that the paper described only in general terms leaves the scope of the authority genuinely uncertain, and the uncertainty surfaces when someone acts on it.
-
What alternatives were considered, including doing nothing
The courses genuinely available and why each was or was not preferred. Token alternatives are recognisable and counterproductive: a paper offering two obviously inferior options reads as having constructed a choice rather than presented one.
-
What the principal risks were understood to be
Stated in the paper, in the body, in terms a director can restate. A risk disclosed only in an annexure was available to the board; whether it was understood by the board is a different question and a harder one to answer afterwards.
-
What the recommendation depends on
The load-bearing assumptions, identified as such. This is the element most often absent, and it is the first thing a review looks for when a decision has turned out badly for a reason that was foreseeable.
-
What was approved, subject to what, within what limits
Conditions, delegations and financial limits, drafted in the paper so the minute can adopt them. Conditions agreed only in discussion depend on a sentence written afterwards by someone reconstructing the sense of the room.
-
Who declared an interest, and what followed
Disclosure, and whether the director participated or withdrew. Procedurally decisive, easily recorded at the time, and effectively impossible to reconstruct later if it was not.
The Downside Belongs in the Paper, Written by the Person Recommending
Every proposal has a case in which it doesn't work. The question isn't whether that case exists but whether it appears in the paper, in what form, and written by whom.
The weakest version is a risk register at the back listing generic exposures with generic mitigations, none quantified and none capable of changing the recommendation. That satisfies a formal requirement and informs nobody.
The strongest is a downside in the body, quantified on the same basis as the recommended case, stating what happens if it materialises and whether the recommendation would still stand. That last element distinguishes a real downside from a decorative one.
A paper presenting a downside and then observing that the proposal remains attractive within it hasn't presented a downside. It has presented a mild variation of the base case.
A genuine downside is one where the recommendation would change, and identifying it requires the author to consider what would change their own mind. It's the most valuable paragraph in most board papers.
Authorship matters. A downside written by the recommending executive, in their own words, is worth more than one supplied by someone else, and its quantification should share a basis with the recommended case so the two compare.
Where a paper reads as though the risks were added by a different hand, an experienced director notices the change of register, and the natural inference is that the recommender didn't want them there.
One reframing helps with the discomfort. A board approving a proposal having seen a properly severe downside has made a decision that's defensible however it turns out. The downside case doesn't weaken the recommendation. It's what makes the approval durable.
Directors Read for the Shape of the Argument Before They Read Its Content
Non-executive directors are structurally disadvantaged and know it.
They see the business intermittently, through documents prepared by the people whose proposals they're assessing, without the operational knowledge to test most assertions directly.
What they develop instead is a sensitivity to the shape of a document: the tells indicating a paper was built to produce a conclusion rather than inform a decision. Those tells are independent of subject matter, which is why they're relied on.
The commonest is asymmetry. Where the upside is modelled in detail and the downside described in a sentence, or the recommended course runs three pages and the alternatives a paragraph each, the paper has shown where its author's effort went.
Effort follows intention. A director doesn't need to understand the business to notice that two halves of an argument weren't prepared with the same care.
A second is the assumption that quietly persisted. Papers on a recurring subject inherit their predecessors' assumptions, and one reasonable when adopted can survive years past the conditions that justified it.
Stating the provenance and date of each load-bearing assumption pre-empts this entirely and costs a line.
A third is the conclusion preceding the analysis. Where material below the recommendation reads as support rather than basis, the analysis reaches only the conclusions it needs and stops.
The tell is what's absent: the calculation not run, the comparison not made, the question a knowledgeable sceptic would have asked immediately.
A fourth is easy to fix: precision offered where the underlying basis doesn't support it. A figure carried to two decimal places on an estimate arrived at by judgement invites, and deserves, a question about the estimate.
| The tell | What it looks like | What it suggests | The fix | |
|---|---|---|---|---|
| 1 | Asymmetric effort | Upside modelled in detail; downside described in a sentence. Favourable assumptions sourced, unfavourable ones asserted. | The paper was built to support a conclusion rather than to test one. | Model both cases on the same basis, to the same depth, with the same sourcing. |
| 2 | The inherited assumption | A figure that has appeared in successive papers, never re-derived, its origin no longer stated. | Nobody has tested it since it was introduced, and conditions may have moved. | State the provenance and date of each load-bearing assumption. One line each. |
| 3 | The absent question | The obvious comparison, calculation or objection that a knowledgeable sceptic would raise, simply not addressed. | It was considered and the answer was unhelpful, or it was not considered at all. | Have the draft read by someone briefed to find what a hostile reader would notice. |
| 4 | False precision | Two decimal places on a figure that was arrived at by judgement. | The presentation is carrying confidence the basis does not support. | Round to the precision the basis supports, and state the basis. |
| 5 | The token alternative | Two alternatives, both obviously inferior, neither seriously analysed. | A choice was constructed rather than presented; the decision was made before the paper. | Analyse the alternatives that were genuinely available, including doing nothing. |
| 6 | Late and long | A substantial paper circulated shortly before the meeting, or a decision buried inside a long update. | Whether intended or not, the effect is to reduce the time available for challenge. | Circulate to the standing notice period; separate decision papers from information papers. |
A generalised description of how papers are commonly read; it is not a statement about any board, company or governance code, and it is not legal advice. Directors' duties and board procedure are governed by the applicable companies legislation, listing rules and the company's own constitution.
In a Board Paper, Length Is Not Thoroughness and an Annexure Is Not a Disclosure
Board packs grow, and each addition is individually justified. A regulator asked about a topic, an incident prompted a report, a director once requested a schedule and it became standing.
The aggregate is a pack nobody can read properly in the time available.
Directors then triage, deciding separately what to skip on the basis of what looks important, and a material item that doesn't look important is precisely what gets skipped.
So materiality has to be exercised by the person preparing the paper rather than delegated to the reader.
The test isn't whether an item is interesting or whether someone might want it. It's whether it could change a director's view of the decision.
What could is in the body. What couldn't, but supports something that could, is in an annexure with a specific reference from the body. What couldn't and supports nothing isn't in the pack at all.
An annexure isn't a disclosure. Material in a schedule at the back, unreferenced from the body, is available formally and unread practically.
When a decision is reviewed later, the question isn't whether the information was in the pack but whether the board was positioned to appreciate it. A significant risk disclosed only on page ninety of an annexure invites an obvious and difficult question.
The reverse failure is real too. A paper omitting material a director needs, because the author judged it too detailed, has substituted the author's judgement for the board's on a matter the board is responsible for.
Two structural habits help more than their simplicity suggests. Separating decision papers from information papers, so a decision is never buried inside an update. And a short cover page stating what is being decided, what is recommended, what it depends on, and what the principal risks are.
One paper, rewritten so the minute can be accurate
A board is asked to approve an acquisition at 84. The paper as drafted runs to eleven pages and recommends approval. Below is the same decision, restructured against the tests above. All figures are illustrative only, and what any particular board requires is governed by its own procedures and the applicable law.
- As drafted Eleven pages, one conclusion
Strategic rationale, market context, target description, synergy analysis, and a recommendation to approve at 84. Risks appear on page ten as a register of eight generic items, each with a mitigation, none quantified. No alternative is analysed. The minute this supports can say only that the board considered the paper and approved.
- Added The alternatives, including doing nothing
Three courses set out: acquire at 84; acquire the smaller competitor identified in the same review at 31; do nothing. The do-nothing case is not neutral — the paper states that the commercial arrangement with the target expires in eighteen months and is unlikely to be renewed on current terms, which is the actual reason for the timing.
- Added The load-bearing assumptions, identified and dated
Four assumptions the recommendation depends on, each with its source and the date it was last tested. One — the retention of the target's two largest customers through change of control — is disclosed as untested, because neither contract has been reviewed for a change-of-control provision. That single line is the most useful in the paper.
- Added The downside, on the same model — value at 61
The base case applied with the two largest customers lost within twelve months and synergies delivered at half the assumed rate. The same model returns 61 against a price of 84. The paper states plainly that on this case the recommendation would be not to proceed at 84, which is what makes it a downside rather than a variation.
- Added The condition that follows from it
Approval recommended subject to review of the change-of-control provisions in the two contracts before exchange, with authority delegated up to 84 and any increase reserved to the board. The condition exists because the paper identified the untested assumption — the two are connected on the page.
- Result What the minute can now record
That the board considered three courses including not proceeding; that it identified customer retention through change of control as the principal risk and was advised it was untested; that it was shown a case valuing the target at 61 in which the recommendation would have been to decline; and that it approved to a limit of 84 subject to a stated condition. Every clause traces to something on the page.
The recommendation did not change and the price did not move. What changed is that a board approving this acquisition can now demonstrate what it knew, what it weighed and what it accepted — and if the customers are lost and the value does fall to 61, the record shows a board that considered precisely that case and took the risk deliberately. The paper became about a page longer. Most of the addition was the four assumptions and the downside, and both were already known to the person who wrote it.
An Audit or Risk Committee Reads for What Management Chose Not to Say
A committee charged with audit or risk oversight reads with a different orientation from the main board, and papers written without recognising that tend to satisfy neither.
The main board decides; the committee assures itself. That the numbers are what they appear to be, that judgements rest on a reasonable basis, that the control environment functions, that risks are managed rather than listed.
Its instinct isn't whether the proposal is good, but what would have to be true for the paper to be wrong.
So where a paper involves an estimate, such as a valuation, provision or impairment assessment, the committee is interested in the range at least as much as the point.
A single figure presented without a range, without sensitivity to its principal assumptions, and without an indication of where in the range management chose to sit has withheld the information the committee exists to examine.
Presenting the range, and saying plainly where within it the reported figure falls and why, converts a question into a disclosure.
Consistency across time is the second instinct. A committee sees the same estimates repeatedly and notices when a methodology changes, or when a figure described as prudent last year is described as realistic this year.
None is improper on its own and each may have an excellent explanation, but the explanation should be in the paper rather than produced in the meeting.
A change of basis disclosed by the author reads as diligence. The same change identified by a committee member reads as something else entirely.
Committees also read for what isn't there, and they're better at it than main boards because their agendas are narrower. The estimate whose assumption wasn't disclosed, the risk reported last quarter that has silently disappeared, the internal audit matter absent from the follow-up schedule.
A change of basis disclosed by the author reads as diligence. The identical change found by a committee member reads as something else — and the merits are no longer the point.
A Paper That Only Argues for Approval Is Itself the Governance Risk
A one-sided board paper isn't merely a weaker document. It's a governance exposure in its own right, independent of whether the underlying decision was sound.
It deprives the board of the ability to demonstrate that it decided rather than ratified, and that ability is the substance of what a board is for.
A board receiving only advocacy is being asked to approve. A board that only ever approves has, structurally, stopped being a board however capable its members are.
The exposure surfaces in a predictable pattern.
A decision turns out badly for a reason that was foreseeable. Somebody reads the paper, an auditor, a regulator, a litigant, an incoming board, and finds the foreseeable reason wasn't mentioned.
The question that follows isn't whether the decision was reasonable at the time. It's whether the board was given what it needed to assess it.
At that point the paper is the entire evidence, and a paper presenting only the case for approval answers the question against the board that received it. That's the wrong party to have exposed.
There's also a live cost, before anything goes wrong.
A board that consistently receives advocacy learns to discount what it reads, and a board that discounts its papers asks more questions in the meeting, takes longer to decide, and trusts management less on the next proposal.
The remedy is structural rather than cultural, and it's within the control of whoever drafts.
Present the decision rather than the recommendation. Set out the genuine alternatives, including doing nothing. Quantify the downside on the same basis as the base case, and say whether it would change the answer.
State the conditions and the limits of the delegation. Put in the body anything that could change a view.
Write the Paper for the Review, and the Meeting Takes Care of Itself
The reorientation is the same one that applies to a credit proposal and to a public submission, arriving from a different direction.
The document's most consequential reader isn't the one in front of you.
For a credit proposal it's the officer who writes the note. For a public submission it's the file, read years later. For a board paper it's whoever reviews the decision afterwards, knowing how it turned out.
A paper written for that reader is complete, balanced, and specific about what was known and when.
A paper written for the meeting is persuasive, and persuasion is exactly what the later reader discounts.
The practical test is short enough to apply to any draft.
Could a reader who wasn't present reconstruct what the board decided, what it considered, what it relied on, and what it accepted?
If yes, the paper has done its work and the meeting will be about substance.
If no, the paper has asked the board to take a risk it cannot later demonstrate it understood, which is the exposure this article is about.
None of it requires more length, more caution, or more process. It requires deciding, before drafting, whether the document is an argument or a record.
How this work is carried out
Projectzo has prepared board and committee papers for listed companies, financial institutions and government bodies since 2010, across 22 countries. Each mandate is assigned a single senior advisor, from scoping through final delivery, and is read by a second senior reviewer before release — against the standard a governance review applies years later, not the standard of the meeting itself.
Commission a turnaround plan