How to Read a Board Paper: What They Say, What They Mean, and What They Are Trying Not to Say

In theory, the difficult part of being a non-executive director is the weighty responsibility of it all: the complex decisions, the strategic counsel, the solemn fiduciary duty. In practice, a considerable portion of the job involves trying to work out what on earth the board paper in front of you is actually saying.

This is not an entirely new problem. But it is a persistent one. Board and committee papers have improved in recent years — fewer appendices of raw data, somewhat less corporate euphemism — yet the familiar traps remain. Papers that are too long swamp the reader. Papers that are too brief invite the reader to construct their own narrative, which is rarely the one management intended and sometimes worse. And then there are the papers that have been, let us say, thoughtfully curated: steering the reader towards a preferred conclusion while maintaining an entirely straight face.

The ability to read a board paper critically — to spot what is missing as well as what is present, to interrogate the framing as well as the facts — is one of the most valuable skills a NED can possess. It is also one of the least frequently discussed. What follows is an attempt to address that gap.

The Problem with Data

Data is supposed to be the antidote to ambiguity. In a boardroom, it rarely behaves so obligingly.

Multiple versions of the truth

It is not uncommon for a board to be working from one set of numbers while individual business units operate from entirely different datasets — often a legacy of the proliferation of systems accumulated through years of acquisitions, reorganisations and technology projects that ran slightly over budget and slightly under specification. The divergence may be innocent, but its effects are not: decisions made on misaligned data are decisions made in the dark, however brightly the dashboard appears to be lit.

A few organisations have addressed this by creating a single, enterprise-wide source of data. It is an admirable solution and a substantial undertaking. For those who have not yet managed it, a useful line of enquiry is to ask the company secretary or CFO to confirm, on the record, that the information presented to the board is consistent with that used by management across the business. An affirmative response is reassuring. A hesitant one is illuminating.

When the numbers are suspiciously good

Some board papers present data that fits the corporate narrative with a neatness that should, in itself, prompt questions. Financials that improve with metronomic consistency, month after month. Customer complaint figures that hover at a level so low they suggest either world-class service or a counting methodology worth examining. Employee satisfaction scores of a warmth that would embarrass a Swiss finishing school.

The instinct to accept such numbers — confirmation bias, in the language of behavioural psychology — is powerful and entirely human. The appropriate response is to cultivate, for the duration of the board meeting at least, an aggressively sceptical frame of mind: not hostile, but genuinely interrogative. Ask to see the raw data. Ask how the figures were compiled. Ask whether external benchmarks exist. One does not need to do this very often before management understands that good news, like everything else, requires supporting evidence.

When the narrative and the numbers diverge

A subtler problem arises when the explanatory narrative accompanying a set of numbers tells a broadly encouraging story, but the numbers themselves — read carefully, with the relevant context — tell a rather different one. The narrative is not necessarily false. It is simply more generous in its interpretation than the data strictly warrants.

Raw numbers are sometimes absent from board reports altogether, either because they have been judged too voluminous to include or because their inclusion would complicate the story being told. This is sometimes called a “snow job”: the reader is buried not under bad news but under a quantity of acceptable-looking information that makes further excavation seem unnecessary.

If there is any doubt about the relationship between a paper’s narrative and its underlying data, insisting that the full dataset be included as an appendix — or made available on request — is a reasonable and proportionate response. A NED who occasionally drills into the underlying numbers, and who is able to construct from them a story materially different from the one presented, has learned something important about the organisation’s reporting culture.

The Forward-Looking Metrics Trap

The counsel to steer by looking ahead rather than in the rear-view mirror is so widely given as to have become almost meaningless. Forward-looking metrics are the right instinct. Badly constructed ones, however, can obscure the road ahead as effectively as any amount of backward-looking data.

Consider a board paper covering progress on a significant but inherently complex project — a new product launch, say, or a major marketing campaign. The paper presents a range of metrics: digital engagement figures, focus group results, spontaneous brand recall, inbound enquiry volumes. Each metric illuminates something. None of them, individually or together, makes clear whether the project is genuinely on track — because the paper does not explain the assumed causal chain between management’s actions, those metrics, and the eventual outcome.

This matters enormously. If spontaneous recall has increased by five percentage points and digital engagement is up twelve, the natural question is: so what? How much should we spend on the next phase? Is this performance in line with, ahead of, or behind where the model assumed we would be at this stage? Without interim targets, without an explanation of the relationship between inputs and outputs, and without enough historical data to calibrate the model, the board is being asked to make decisions on the basis of movement rather than direction.

In such circumstances, the honest answer is that the board must rely on management’s experience and judgement — which may well be sound, but is an unsatisfying basis for authorising significant expenditure. The more productive response is to insist, before the next major initiative begins, on a shared understanding of the causal logic: what will be measured, what interim targets will be set, what the known areas of uncertainty are, and what would constitute evidence that the project is in trouble. If management cannot provide this level of analytical rigour at the outset, that conversation is itself valuable.

The Traffic Light Trap

Red, amber and green. Simple, intuitive, universally understood — or so it appears. In practice, RAG reporting contains at least three distinct traps, each capable of misleading a board that has not thought carefully about how the system has been designed.

The state-versus-outcome confusion

The first concerns what, precisely, the traffic lights are measuring. A project that is running behind schedule at the midpoint may be reported as green by a confident management team that fully intends to recover the position by delivery. Whether such optimism is warranted is a separate question. The problem is that a board reading green understands the project to be in good health, when the more accurate description might be “currently behind, but management believes recovery is achievable.” These are not the same thing. Reporting guidelines should specify whether RAG statuses reflect the current state of affairs or management’s assessment of the likely end state — and ideally, they should capture both.

The undefined colour problem

The second trap is definitional. Amber means different things to different people. Without explicit, agreed criteria for what triggers each colour — in quantitative terms wherever possible — a RAG report is less a system of objective signalling than an expression of the author’s temperament. One manager’s amber is another’s red; one executive team’s green is another’s cause for concern. The solution is straightforward: define the triggers, publish the definitions, apply them consistently.

The amber gravity well

The third trap is the most insidious, because it is a product of human nature rather than poor system design. When presented with three options, people tend to select the middle one. In RAG reporting, this gravitational pull towards amber means that many board and committee papers show a preponderance of amber statuses — and many of them stay amber, month after month, with a stability that appears reassuring but may simply reflect the reluctance of authors to commit to a more definitive assessment.

Worse, a static amber can mask genuine deterioration. A project moving from healthy amber towards red may spend several reporting cycles at amber before the change in direction becomes visible. By which point, the window for effective intervention has frequently closed.

A practical remedy is to replace the three-colour system with four: red, red-amber, green-amber and green. This forces matters out of the comfortable middle, compels authors to think more carefully about what they are actually reporting, and provides earlier warning of movement in either direction. It is a small change with a disproportionate effect on the quality of board information.

The Middle Option Trap

Well-constructed board papers present directors with genuine choices: a range of credible options, with an honest assessment of the merits and risks of each. This is how it should work.

In practice, authors who have already decided which course of action they favour sometimes frame the alternatives with a degree of care that borders on the theatrical. The preferred option — typically the middle one — is bracketed by two alternatives that have been selected, consciously or otherwise, because they are most easily dismissed: one too bold, one too timid, leaving the recommended course looking like the only reasonable conclusion a sensible person could reach.

The antidote is not to ignore the middle option but to interrogate the others with equal rigour. How realistic are they, really? What alternatives were considered and discarded before the paper was written? Why were they discarded? The answers to these questions frequently reveal more about management’s thinking — and about the quality of the underlying analysis — than the recommended option itself.

Groupthink and the Perils of Tidiness

There is a particular kind of board paper that is almost entirely free of tension. Every metric is moving in the right direction. Management’s proposed actions are clearly correct. The risks are noted and apparently under control. The executive and the various reporting layers below it all appear to be of one mind.

This may reflect a genuinely well-run organisation. It may also reflect the effect of templates, length restrictions, editorial conventions and the accumulating pressure of a reporting culture in which bad news travels slowly and counter-narrative information is quietly omitted at successive stages of the drafting process.

“Reporting by exception” — presenting only items that deviate from plan — is a well-intentioned reform that can exacerbate this problem. A board that sees only the exceptions loses the context that would allow it to assess whether the exceptions are genuinely exceptional, or whether they are emerging as a new norm that nobody has yet felt comfortable enough to name. Significant secular trends — gradual shifts in customer behaviour, a slow deterioration in a particular risk metric, a pattern of small operational failures — can disappear entirely from a board’s view if they fall below the exception threshold.

NEDs should ask for counter-narrative information as a matter of routine: the data that complicates the dominant story, the dissenting views, the risks that management considers manageable but that an independent observer might weigh differently. If none exists — if the organisation genuinely presents a single, uncontested picture on every significant matter — that is itself a signal worth examining.

The Curse of Vagueness

Precision in language is not pedantry. In a board paper, it is a governance obligation.

The problem with phrases such as “a number of”, “broadly speaking”, “a few”, “many”, “significant” and “insignificant” is not that they are imprecise — it is that they are imprecise in ways that allow different readers to draw entirely different conclusions from the same sentence. This divergence in interpretation is sometimes innocent; occasionally it is convenient.

The point was illustrated memorably, if unintentionally, by a former Cabinet minister who described her schedule of unauthorized meetings with foreign officials as involving “a few” encounters. The eventual tally was twelve meetings in twelve days. Whether that constitutes a few is, as she might have put it, open to interpretation. Twelve is not.

Board paper guidelines should insist on precise terminology wherever it is available: actual numbers rather than approximate ones, defined thresholds rather than subjective adjectives, specific timeframes rather than loose references to the near or medium term. It is a small discipline with a significant cumulative effect on the quality of board information.

The word “average” has a similar problem – is it mean, mode or median? The difference between each can be significant.

When Brevity Becomes Sophistry

A board paper can be too long. It can also be too short — not because it fails to meet a page count, but because it says less than the subject warrants, and relies on the reader’s disinclination to press for more.

The signs are recognisable to any experienced director. The narrative is oddly smooth, with no rough edges. Hard KPIs are absent, or present only in their most favourable form. Vague language predominates. The subtext, if one were to render it explicit, would read something like: “Nothing troubling is happening here; please proceed to the next item.”

Reading such documents with a genuinely sceptical mindset — asking what is absent rather than what is present, what a well-informed author would have included that this author has not, how this paper compares with analogous reporting seen elsewhere — is the most reliable way to identify the gaps. Spotting what is not there is a skill that improves with practice and that tends to improve the quality of future papers considerably, once management understands that absence of evidence is being treated as evidence worth investigating.

The Biggest Trap of All: Reading the Wrong Papers

All of the foregoing assumes that the papers presented to the board are addressing the right questions. This assumption deserves more scrutiny than it usually receives.

Many organisations govern their board agenda through a content calendar: a schedule of what will be discussed, when, agreed in advance and followed with reasonable consistency. In stable conditions, this is sensible governance hygiene. In rapidly changing conditions — and conditions have been changing at considerable pace for some time — it can mean that the board’s timetabled preoccupations drift away from the organisation’s most pressing challenges without anyone quite noticing.

The volume of fiduciary and regulatory reporting required of boards in many sectors compounds this problem. Compliance-driven papers — necessary, backward-looking, procedurally important — can crowd out the strategic and forward-looking discussion that should sit at the centre of a board’s work. A board that spends most of its reading time on what has happened, and most of its meeting time on regulatory obligations, is not necessarily governing badly. But it may be governing the wrong things.

One corrective is to seek out unconventional data: information that conventional management reporting does not capture and that may point to questions the board should be asking but is not. Net Promoter Scores from customers or patients. Third-party rankings and sector analyses. Analyst commentary on competitors. Social listening data. Glassdoor ratings. These sources are imperfect, partial and sometimes misleading — but they are not filtered through the organisation’s own reporting machinery, which gives them a particular value.

The deeper question — whether the board’s agenda is genuinely aligned with the organisation’s most important challenges and its stakeholders’ most pressing concerns — is one worth asking explicitly, perhaps annually, rather than allowing the content calendar to answer it by default.

How to Prevent Problems More Widely

The most effective safeguard against poor board papers is not better NED training, though that helps. It is a board paper culture in which high standards are set clearly, enforced consistently, and understood by everyone who contributes to the process.

This means reviewing reporting guidelines regularly — not merely dusting them off when a crisis makes the inadequacy of the existing approach impossible to ignore. It means ensuring that the company secretary’s secretariat has both the time and the authority to return substandard papers for redrafting. Requiring a paper to be rewritten is never popular; it is, however, considerably less painful than the alternative, which is a board that makes decisions on the basis of inadequate information and a regulator that subsequently wants to understand why.

In sectors where regulatory scrutiny is intense — financial services, healthcare, utilities and others — the quality of board information is not merely a governance aspiration. It is a material risk. A Section 166 review, or its equivalent in other regulated industries, will examine what the board was told, when it was told, and what it did as a result. The board pack is the primary evidence in that examination.

But guidelines and gatekeeping will only go so far. The most important safeguard of all is a board that stays alert — that reads its papers critically, speaks up when something does not look right, and treats the absence of bad news with the same scepticism it would apply to an abundance of it.

A board that does this well, consistently, is one of the most effective governance mechanisms any organisation can have. It is also, in the author’s experience, somewhat rarer than it should be.