Most cybersecurity reporting suffers from the same problem: it counts activity instead of measuring risk. A slide that says the firewall blocked 4.2 million events last month feels reassuring, but it tells a board nothing about whether the organisation is more or less likely to suffer a damaging breach. Good metrics answer a harder question: are we managing our cyber risk within the level the business has agreed to accept?
This guide sets out which cybersecurity metrics and key performance indicators actually inform decisions, how to distinguish the useful from the decorative, and how to report them so that executives and directors can act.
Leading versus lagging indicators
The single most important distinction in security measurement is between leading and lagging indicators, and most programs lean far too heavily on the latter.
Lagging indicators measure outcomes that have already occurred. The number of incidents this quarter, the number of records exposed, or the hours of downtime caused by an attack are all lagging. They are valuable for accountability and for understanding the cost of failure, but by definition they only tell you about the past. You cannot manage a number that has already happened.
Leading indicators measure the conditions that make incidents more or less likely. The percentage of internet-facing systems patched within your target window, the proportion of staff who have completed phishing simulation training, the share of privileged accounts protected by multi-factor authentication, and the percentage of critical assets covered by monitoring are all leading. They predict future risk, and crucially, you can act on them now.
A mature program reports both, but uses leading indicators to drive day-to-day decisions and lagging indicators to validate whether those decisions are working. If your patching leading indicator improves and your incident lagging indicator falls over the following quarters, you have evidence the investment is paying off.
The metrics that matter most
You do not need fifty metrics. You need a focused set that maps to genuine risk reduction. The following categories cover most organisations.
- Vulnerability and patch performance. Mean time to remediate critical vulnerabilities, and the percentage of critical and high vulnerabilities patched within your service-level window. Break these down by asset criticality so that a slow patch on a test server does not hide a slow patch on a payment system.
- Identity and access. Percentage of accounts with multi-factor authentication enabled, number of privileged accounts, and the time taken to revoke access when someone leaves. Stale and over-privileged accounts are a recurring root cause of serious breaches.
- Detection and response. Mean time to detect and mean time to respond to incidents. These show whether your monitoring and response capability is improving, and they are far more meaningful than a raw count of alerts.
- Coverage and configuration. Percentage of assets under endpoint protection, logging and backup. Coverage gaps are where attackers live, so a metric showing 88 per cent backup coverage immediately raises the question of which 12 per cent is unprotected.
- Human risk. Phishing simulation click and report rates, and training completion. Trend matters more than a single figure; a falling click rate alongside a rising report rate is a genuine cultural improvement.
- Third-party risk. Percentage of critical suppliers assessed within the last review cycle, and the number of suppliers with unremediated high-risk findings.
For organisations using the Essential Eight, many of these metrics map directly to maturity-level evidence, which means the same measurement effort supports both internal management and external assurance. APRA-regulated entities will recognise several as inputs to their CPS 234 control testing obligations.
Tie every metric to a target and a risk
A number on its own is noise. Each metric needs three things to be useful: a target that reflects your risk appetite, a trend that shows direction, and a clear link to the business risk it represents.
For example, “mean time to remediate critical vulnerabilities” becomes meaningful when expressed as “currently 18 days against a target of 14 days, improving from 31 days six months ago, reducing our exposure window to actively exploited flaws.” That single sentence tells an executive the current state, whether it is acceptable, the direction of travel, and why they should care. Reporting metrics without targets invites the question every director eventually asks: is that good or bad?
Reporting to the board versus the operations team
The most common reporting failure is showing the board what the security operations centre needs to see. Operational teams need granular, high-frequency data: alert volumes, queue depths, individual asset states. Boards need a small, stable set of indicators tied to risk and money.
For board-level reporting, keep to a handful of metrics and frame them around risk appetite. A useful structure is: where are we against our agreed tolerances, what is the trend, what is the most significant exposure, and what investment decision (if any) is required. Avoid jargon, avoid raw event counts, and always show trend over at least three to four periods so a single anomaly is not mistaken for a crisis.
A virtual CISO is often the person who translates between these two worlds, taking operational telemetry and turning it into a risk narrative the board can govern against. If you are building this reporting capability for the first time, our virtual CISO service and the overview of what a virtual CISO does explain how that translation works in practice.
Common measurement mistakes to avoid
- Vanity metrics. Blocked-attack counts, total alerts, or training hours delivered measure effort, not risk. Drop them from executive reporting.
- Averages that hide outliers. An average patch time of 12 days can conceal a critical system sitting unpatched for 90 days. Report distributions or worst-case figures for critical assets.
- Changing definitions. If you redefine what counts as a “critical” vulnerability between quarters, your trend becomes meaningless. Lock definitions and document them.
- Measuring only what is easy. The metrics that are simplest to extract from tools are rarely the ones that matter most. Start from the risks you care about, then find the measurement.
- No accountability. Every metric outside tolerance should have a named owner and a remediation date, or the report becomes a record of problems nobody is fixing.
Build a simple maturity view alongside the numbers
Raw metrics tell you about specific controls, but executives often also want a sense of overall direction. A lightweight maturity view, layered on top of your metrics, provides this without inventing a new measurement burden. Rather than a single vague “we are improving” claim, you can show maturity by domain: identity, vulnerability management, detection and response, third-party risk, and so on, each rated against a defined scale and supported by the underlying metrics.
This approach has two advantages. First, it lets the board see at a glance where the organisation is strong and where it is weak, which directs attention and budget. Second, it connects naturally to recognised models. For organisations using the Essential Eight, maturity levels provide a ready-made scale, and your metrics become the evidence that justifies each rating. The discipline to avoid here is grading yourself generously; maturity claims should be backed by the same hard metrics you report elsewhere, not by optimism.
Connect metrics to investment decisions
The ultimate test of a metrics program is whether it changes where money and effort go. Metrics that simply describe the status quo, however accurate, add little value. Metrics that drive decisions are the ones worth keeping.
To make this connection explicit, pair significant metrics with the decision they inform. A persistently slow patch metric on critical systems should prompt a conversation about whether the team is under-resourced, whether the tooling is inadequate, or whether legacy systems need investment to become patchable. A backup coverage gap should trigger either remediation funding or a documented, accepted risk. When every out-of-tolerance metric leads to a clear choice, fund the fix, accept the risk, or change the target, the program earns its place in governance. When metrics float free of any decision, they become a quarterly ritual that consumes effort and produces nothing.
This is also where good metrics protect the security budget. A leader who can show, with trend data, that past investment reduced exposure has a far stronger case for the next request than one offering only assertions. Metrics turn budget conversations from faith into evidence.
Getting started
If your current reporting is mostly activity counts, do not try to build a perfect program overnight. Choose three or four leading indicators that map to your biggest risks, set realistic targets, and report them consistently for two or three quarters. Once the trend lines tell a story, expand the set. The goal is not a beautiful dashboard; it is a small number of measures that change what the organisation does.
CISO Advisory works with Australian government, financial services and enterprise clients to build security metrics that boards can actually govern with. If you would like help designing a metrics and reporting framework that fits your risk appetite, get in touch or call 07 2112 8502.
Frequently asked questions
What is the difference between a leading and a lagging cybersecurity metric?
A lagging metric measures outcomes that have already happened, such as the number of incidents last quarter. A leading metric measures conditions that predict future outcomes, such as the percentage of systems patched within your service-level target. Leading indicators let you act before an incident occurs; lagging indicators tell you whether past efforts worked.
How many cybersecurity KPIs should we report to the board?
Aim for five to ten board-level metrics. Boards govern risk, not operations, so they need a small set of indicators tied to business impact and risk appetite. Keep deeper operational metrics for the security team and use the board set to show trend, exposure and whether you are inside agreed tolerances.
What is a good metric for measuring patching performance?
Mean time to remediate critical vulnerabilities, and the percentage of critical and high vulnerabilities patched within your defined window, are the most useful. They are leading indicators of exposure and map directly to Essential Eight patch application maturity. Track them by asset criticality, not just as a single organisation-wide average.
Should we benchmark our metrics against other organisations?
External benchmarks are useful for context but treat them with caution because definitions and scope vary widely between organisations. Your own trend over time is usually more meaningful than a single comparison. Where you do benchmark, prefer sector-specific sources and frameworks such as the Essential Eight maturity model.
How often should cybersecurity metrics be reported?
Operational teams should review key metrics weekly or even daily. Executive committees typically review monthly, and boards quarterly. The cadence matters less than consistency: report the same metrics the same way each period so trends are visible and a single bad month is not mistaken for a crisis.
Need this handled for your organisation?
Confidential and no obligation. We respond the same business day — on-site same day / next business day, or remote, Australia-wide. Prefer to talk now? Call us 24/7 on 07 2112 8502.