The room and the record¶

Two practices that address the same underlying condition. The brief does not stay where it was intended to stay, and the account of what was said does not stay accurate.

The room is not just the room¶

A brief prepared for institutional leadership travels to oversight bodies. A restricted report is summarised in a follow-on debrief, and the summary circulates further. An informal channel between a senior official and a principal’s representative carries versions of the analysis that were never written down. A paper prepared for a closed technical session informs a public position statement.

Not exceptional. This is how information moves in institutions with overlapping lines of authority, competing interests, and dense informal networks. Treating the analysis as if it exists independently of its institutional context is a mistake. It does not.

Threat-modelling a briefing¶

Before a brief circulates, map the room, not just the formal attendees but the actual routing.

Who is formally in the room, and what position does each hold. Who sees the paper afterwards: immediate recipients, their staff, counterpart offices, institutional officials who will file it. What gets discussed informally after the session ends.

Then: what are the interests at stake for each audience the brief is likely to reach. A technical finding that is neutral for institutional leadership may be significant for an oversight body with a particular stake in the outcome. A confidence tier of “it appears” reads differently for an audience looking for grounds to resist a policy than for an audience looking for grounds to advance one.

The threat model also covers the hostile reading, meaning a reading by someone with an opposing position, looking for what the document concedes. A well-constructed brief can be used to support positions the advisor did not intend to support. Knowing what the hostile reading produces before the brief circulates is part of the work.

The reception environment as part of the analysis¶

A recommendation delivered to the right body at the right moment in the institutional cycle is a different document from the same recommendation delivered off-cycle. Content is identical. Effect is not.

Timing within the institutional cycle, the lead-up to a formal review, the period immediately after a significant incident, the window before a mandate renewal or assessment, can affect what a brief accomplishes. Ignoring the reception environment tends to produce analysis in the abstract, which is less useful than analysis that accounts for where it will land.

Framing a brief for its routes¶

If a brief is likely to reach audiences beyond the stated recipients, and in mandate-driven institutions it often will, the framing carries that further travel. Not by making the analysis less accurate, but by being explicit about what it does and does not claim, what evidence it rests on, and where the limitations are. A brief that reads as an overreach in the hands of an unintended audience can do more damage than a less confident brief that was well-targeted.

The threat model of the briefing is also relevant to what is left out. Some findings are accurate and consequential but would be misused if they reached certain audiences at certain moments. Deciding what to include, when, and for which audience is not the same as suppressing findings. It is sequencing them appropriately, which requires knowing the institutional terrain before the brief goes out.

Documenting advice contemporaneously¶

Decisions travel through long chains. An advisor’s recommendation, by the time it reaches the person who acts on it, may have passed through several intermediaries, each of whom interpreted it through their own understanding of the institutional position. The recommendation may have been summarised, abbreviated, or framed in the context of a subsequent conversation the advisor was not part of.

Advisors get misremembered. Sometimes by accident: the chain is long, memory is imperfect, and the version that arrived at the decision-maker was a reasonable paraphrase that lost something important. Sometimes not by accident: the account of what was advised is politically convenient to reshape, and nobody with an interest in accurate reconstruction is in a position to provide it.

A short, dated note in the advisor’s own files of what was asked, what was said, on what evidence, and what caveats were stated is cheap insurance. Sometimes it is the only honest reconstruction available when the institutional record has been shaped by the interests of the parties who controlled it.

The note¶

Not a retrospective write-up. The same day, or as close to it as possible: a note for the advisor’s own records. What was asked. What was said. What the evidence was. What the caveats were, including caveats stated verbally that did not appear in the written brief. Whether the advice was accepted, modified during the discussion, or set aside.

Short is fine. The point is the timestamp and the accuracy. A paragraph written the same day is more reliable than a page written six months later, and both are more reliable than memory at an inquiry.

A caveat that did not travel¶

A senior official requests an assessment of a proposed technical arrangement, urgently, before a scheduled meeting. The assessment is provided verbally in a short conversation, with a one-page summary to follow. The conversation includes an explicit caveat: the assessment rests on the technical specifications provided, and if those specifications are revised, the assessment may not hold.

The one-page summary goes to the meeting. The caveat is not in the summary because the context of the verbal conversation is assumed to have been communicated through the official. Some months later, the technical specifications are substantially revised. The arrangement is implemented on the basis of the earlier assessment. When a problem emerges, the advisor is identified as having provided a clean assessment with no significant reservations.

The contemporaneous note, written the afternoon of the original conversation, records the caveat, the condition under which the assessment was valid, and the fact that the summary document did not carry it. That note is the only accurate reconstruction of what was said. It is not a guarantee that the record will be corrected. But it is available, with a timestamp, when the conversation about what was said eventually happens.