Ninety minutes is a discipline

Every decision document makes a claim on the scarcest budget an executive team manages: its attention. That claim is almost never priced. Page counts grow because adding is safer than cutting, because every contributor's section must survive review, and because length still passes, in some rooms, for diligence.

Our answer is a number and a rule. The number is ninety minutes: the executive read of any Stredge Partners decision document is designed to be complete, evidence included, inside ninety minutes. The rule is what the number does — it is a design constraint on the author, not a reading instruction to the reader. A ninety-minute ceiling forces evidence density, with every claim graded, every threshold on the page and every section earning its place, because nothing else fits. And it encodes a judgment: a document that cannot change a decision in ninety minutes will not change it in nine hours, because the constraint that binds is attention, not information.

This note grades its own evidence carefully, because the evidence here is thinner than it looks. What research establishes about executive attention is real but old and small; what governance institutes report about board packs is survey perception; and no study we could verify establishes ninety minutes as a threshold. The number is ours. The discipline it enforces is the argument.

The evidence divides honestly into three shelves: what has been measured, what has been surveyed, and what does not exist.

What has been measured. The founding observational study of executive attention remains Mintzberg's structured observation of chief executives at work, replicated three decades later. The point for document design is not any single number; it is that sustained executive attention is a scarce, interrupted resource, measured as such for fifty years, and that the longest attention blocks executives reliably give anything sit near the hour, not the afternoon.

What has been surveyed. The governance institutes have looked at the artifact itself, and the people who assemble the board pack report that the pack overclaims the board's attention. That is perception, not measurement, and the research was conducted with a board-portal software provider, which we flag rather than hide. But as evidence of a governance profession's own verdict on document length, it points one way.

The UK regulator completes the shelf from the demand side: the Financial Reporting Council's Guidance on Board Effectiveness treats the quality and timing of board papers as a named factor in decision-making, and observes that well-informed, high-quality decisions do not happen by accident. Official guidance, and deliberately unquantified — it tells boards to design their information and declines to tell them the number.

What does not exist. We searched for peer-reviewed evidence that a specific reading duration, ninety minutes or any other, marks the point past which decision documents stop changing decisions. We did not find any that meets our standard, and we are not going to cite around the gap: the ninety-minute figure is not a research result. It is a firm position, graded SEG-1, and it earns its keep differently, as a constraint whose consequences can be inspected. Under a ninety-minute ceiling an author cannot defer the argument to an appendix, cannot let two claims share one ungraded source, and cannot write the decision into paragraph forty. The ceiling does not guarantee a good document; it makes a bad one visible in one pass.

The constraint binds hardest in our corridor. A board deciding between Brussels and Abidjan reads everything twice, because in the WAEMU–euro corridor the decision estate is bilingual by construction: the French edition for the seats that own the consequence under a French-language legal regime, the English edition for the holding board and its lenders. An advisory habit that ships two-hundred-page packs in one language ships four hundred here, and those pages compete with reading that cannot be declined — the uniform acts, the central bank's communiqués, the regulator's calendar. Discretionary reading in this corridor is taxed at the border; density is not style, it is the tariff. The ceiling also disciplines the bilingual build, because two editions stay decision-identical only if the decision core is small enough to freeze before the second authorship begins: the thesis, the graded claims, the thresholds.

Evidence cards

What the ceiling buys is one decision per document, evidence carried as graded cards rather than an annex the reader is trusted not to open, limitations printed where the claim is made, and a closing section a chair can turn to first. What it refuses is the comfort of comprehensiveness and the quiet transfer of the author's editing work to the reader's calendar. The boundary is narrow: some decisions require reference documents that exceed any single sitting, and the rule governs the executive read, not the archive.

Sources and limitations

Sources and limitations. The measured shelf rests on Mintzberg's 1973 structured observation and Tengblad's 2006 replication in the Journal of Management Studies; the surveyed shelf on the Chartered Governance Institute UK & Ireland's 2017 board-pack research; the official shelf on the FRC's Guidance on Board Effectiveness (2018), which names paper quality as a decision factor and quantifies nothing. Each carries its caveat above, and the survey's commercial association is flagged. The ninety-minute figure itself is the firm's position, graded SEG-1: we found no peer-reviewed threshold for the reading duration past which decision documents stop changing decisions, and this note does not pretend otherwise. The boundary is explicit: the ceiling governs the executive read of decision documents; it is not a claim about reference material, regulatory filings or technical annexes, and it is not evidence that shorter documents produce better outcomes — no such outcome study is cited here.

1. The Chair sets the attention budget in the commissioning note, before the paper is written. A budget set after the document arrives is not a budget; it is a surrender negotiated page by page.

2. The Company Secretary asks the author of the last board paper what it would keep if it had ninety minutes, and circulates that version instead. Ask the author, not the board. The difference between the two documents is the cost being transferred to the room.

3. The paper's sponsor records who was designed to finish it, and whether anyone did. If completion was never a design goal, the document was written for the file, not for the decision.

4. In a bilingual estate, the Managing Partner checks which edition was designed for the reading time its seats actually have. If one edition is a courtesy copy, the attention budget of the consequence-bearing seat was the one cut first.