A board cyber report fails in one of two ways. Either it is an activity log, full of blocked attacks and training completion, or it is a heat map with no history, where every square is amber and nobody can say whether amber is better or worse than last quarter.

Both fail for the same reason. A director's job is to set direction, agree how much risk the business will carry, and hold management to it. A report that does not show risk against an agreed limit gives them nothing to do except ask "are we secure?", which is the question the report was supposed to answer.

The UK Cyber Governance Code of Practice, published in April 2025, states the board's side of this plainly: require formal reporting at least quarterly, set suitable metrics to track, and agree tolerances for each. That one sentence is the design brief for the whole report. Metrics, with tolerances, that the board agreed. Everything below follows from it.

Preconditions

Do not build the report until three things exist, because the report is a view onto them and cannot substitute for them.

  • A written risk appetite statement for cyber, approved by the board or a committee. NIST CSF 2.0 lists this as a Govern outcome (GV.RM-02): risk appetite and tolerance statements established, communicated, and maintained. Without it, no metric has a threshold.
  • A risk register with named owners. The top risks in the report must be traceable to entries someone in management owns.
  • A named committee. Audit, risk, or a dedicated technology committee. US public companies already have to describe this in their annual filing under Regulation S-K Item 106(c), including the processes by which the board or committee is informed about cyber risk. The report is that process, so it should match what the filing says.

If the appetite statement does not exist, the first quarterly report is the proposal for one. That is a better first meeting than a deck of numbers with no thresholds.

Phase table

PhaseWhenOutputOwner
Agree the frameOnce, then yearlyAppetite statement, 6 to 10 metrics, a tolerance for eachCISO with the committee chair
CollectWeeks 1 to 2 of the quarter endMetric values, register changes, incident summarySecurity operations, GRC
DraftWeek 3Five-section report, two pages plus appendixCISO
Pre-briefWeek 3Committee chair walkthrough, no surprisesCISO
PresentCommittee meetingDecisions recorded in the minutesCISO, committee
Close the loopWithin two weeksActions logged with owners and datesCISO

The five sections

1. Risk position against appetite

Lead with the five to eight risks that matter most, each shown as current exposure against the tolerance the board agreed, with the direction since last quarter. One line each. If a risk is outside tolerance, say so in the first sentence of the report, not on slide nine.

This is the section that turns the report from information into oversight. A director reading "credential compromise exposure is outside tolerance for the second quarter running" knows exactly what to ask next.

2. Metrics with tolerances

Six to ten metrics, each with a value, a tolerance, and a trend over at least four quarters. A metric belongs here only if it passes three tests:

  • A decision follows from it moving. If the number doubled, would anyone change a budget, a priority, or a deadline? If not, cut it.
  • Management controls it. Attack volume is set by attackers. Time to patch internet-facing critical vulnerabilities is set by you.
  • It is defined the same way every quarter. A metric whose definition changed is a new metric, and the trend line has to restart.

Metrics that tend to pass: coverage of multifactor authentication on privileged and remote access, time to remediate critical exposed vulnerabilities against your own stated target, percentage of critical systems with tested restore, open high-severity audit findings past due, third parties with access to sensitive data that have a current review, and time to revoke access for leavers. Metrics that tend to fail: blocked emails, blocked attacks, alerts processed, and training completion on its own.

3. Decisions needed

The section most reports leave out. List what you need the board to decide or note this quarter: a risk acceptance above your authority, a tolerance that should change, a budget reallocation, a deadline that will slip. Each item gets a recommendation and a deadline.

A report with no decisions in it for four quarters in a row is a signal that the tolerances are too loose or that risk is being accepted below the level the board thinks it is.

4. Incidents and near misses

What happened, what it cost or could have cost, what changed as a result. Keep it to material and near-material events, and say which ones were tested against your disclosure process. For US registrants, the board will want to know how the materiality determination was made, because a material incident has to be disclosed on Form 8-K within four business days of that determination.

Include exercise results here too. The UK code asks boards to confirm the incident plan is exercised at least annually and that lessons feed back into the plan. One line per exercise: date, scenario, what broke, what was fixed.

5. What changed

Regulatory changes, major vendor or architecture changes, and anything that moved a risk rating. Short. This is the section that explains why section 1 moved.

Put the detail, methodology, and metric definitions in an appendix. Directors who want them will read them, and the ones who do not should not have to scroll past them.

What a good page one looks like

Page one should be readable in three minutes and survive being the only page read. A workable shape:

ItemContent
HeadlineOne sentence: overall position against appetite and the direction
Outside toleranceNamed risks, how long they have been outside, the plan
Decisions neededUp to three, each with a recommendation
Material eventsCount and one-line description, or "none"
Next quarterThe one change the committee should expect to see

If page one cannot be written, the frame is not ready. That is useful information, and the fix is in the preconditions, not the slides.

Pre-briefing the chair

Walk the committee chair through the report before the meeting. The point is not to rehearse. It is to make sure nothing in the room is a surprise to the person running the agenda, and to find out which question the chair intends to ask so you can answer it on paper.

A risk going outside tolerance should never reach the chair for the first time in the meeting. If it moved between meetings and it matters, it should already have reached them between meetings, which is a separate escalation path covered elsewhere on this desk.

Failure modes

The metric zoo. Thirty metrics because each team wanted theirs included. A board cannot hold thirty tolerances in mind. Cap the set and rotate supporting metrics into the appendix.

The permanent amber. Ratings that never move, because moving a rating to red invites scrutiny and moving it to green invites a budget cut. If a rating has not moved in a year, either the risk is stable and should drop out, or the rating method is not measuring anything.

The vendor benchmark as headline. An industry survey figure is context, not a measurement of your organization. Use it next to your own number, labelled as a survey, never instead of it.

Changing definitions quietly. If the vulnerability metric now covers a different asset set, say so and restart the trend. A director who notices the change before you mention it will discount everything else on the page.

No decisions, ever. See section 3. A report that never asks for anything is a newsletter.

Skip this if

You report to a board that meets twice a year and has no committee with cyber in its remit. In that case the first job is getting cyber onto a committee's terms of reference, and the quarterly cadence follows once someone owns it. Small private companies with a founder-led board can run the same five sections as a one-page monthly update instead.

What to cut if you only have two weeks

Keep sections 1 and 3. Risk against appetite and decisions needed are the two that give the board something to do. If no appetite statement exists yet, replace section 1 with a draft statement and ask the committee to approve or amend it. That is a complete and useful first report.

Next steps

  1. Find your current risk appetite statement, or draft one for the committee chair to react to.
  2. Pick six metrics that pass the three tests above, write a one-line definition and a tolerance for each, and backfill four quarters where the data exists.
  3. Build page one before anything else.
  4. If the deck is being used to justify spend, pair it with the budget piece on this desk, and for the question every board eventually asks, see how to answer "are we secure?".
  5. Start from the board cyber metrics one-pager template; the rest of the template library is at /ciso/templates/.