Security metrics are quantitative measurements that show whether a security program is working, track progress toward defined goals, and translate technical activity into terms leadership can act on. Good metrics tie back to business outcomes and decisions, not to raw counts of alerts or blocked emails that look busy but say little about real risk.
NIST’s current guidance is SP 800-55, which was split into two volumes in December 2024, replacing the 2008 Rev. 1. Volume 1 covers identifying and selecting measures, and Volume 2 covers building a measurement program.
How do security metrics work?
NIST draws a line between measures and metrics. A measure is a single quantified result, like “212 of 230 servers carry the standard configuration.” A metric uses measures and assessment results to track progress over time and support decisions, like “standard configuration coverage rose from 78 percent to 92 percent this quarter, against a 95 percent target.”
SP 800-55 Volume 1 sorts measures into four types, and a mature program moves through them in order.
| Type | What it shows | Example |
|---|---|---|
| Implementation | Whether a control is in place | Percentage of systems with an approved security plan |
| Effectiveness | Whether the control does its job | Percentage of phishing simulation reports versus clicks |
| Efficiency | How fast controls give feedback and issues get fixed | Mean time to remediate critical vulnerabilities |
| Impact | Effect on mission, cost, and trust | Security spend as a share of budget, losses avoided, regulatory fines |
Implementation measures come first and never fully retire, because they record what exists. NIST notes that impact measures give executives the most direct view of security’s value, but they need the most analysis before a non-technical audience can use them.
Each metric needs a defined owner, a data source, a collection frequency, a target, and a formula. Without those, two teams will report the same number differently and nobody will trust either.
What is the difference between a KPI and a KRI?
The CISSP outline names both. A key performance indicator (KPI) looks backward at how well a process performed, such as the share of systems patched within the 30-day policy window. A key risk indicator (KRI) looks forward at how exposure is changing, such as the count of internet-facing systems with known exploited vulnerabilities older than 14 days. KPIs tell you whether the team is doing its job. KRIs warn that risk is heading past the organization’s Risk Appetite before a loss happens.
Many numbers can be either, depending on the threshold. Patch latency is a KPI when you grade the operations team and a KRI when you set a trigger that escalates to the CISO.
What is an example of a useful security metric?
Useful examples include mean time to detect (MTTD) and mean time to respond (MTTR), the share of systems patched within policy, critical vulnerabilities still open past their remediation deadline, and phishing simulation click and report rates measured before and after training.
The common trap is vanity metrics. Reporting that the firewall blocked ten million events sounds impressive but tells an executive nothing about exposure, since blocked traffic is the control working as designed. Compare that with “critical vulnerabilities older than 30 days on internet-facing systems fell from 41 to 6 after we added a second patch window.” A board can act on that. A number earns its place only if someone would make a different decision when it moves.
How are security metrics tested on the CISSP?
Metrics sit in Domain 6: Security Assessment and Testing. Objective 6.3, collect security process data, lists key performance and risk indicators along with account management, backup verification data, and training and awareness. Objective 6.4 covers analyzing results and reporting them. Domain 1 adds continuous monitoring and measurement under objective 1.9, and objective 1.12 asks for program effectiveness evaluation of awareness training.
Expect scenario questions that ask which metric a board would care about. The managerial answer ties the metric to business risk and an agreed target, not to activity volume or tool output. Also watch for questions about what to do first. You can’t measure effectiveness until implementation is known, and a metric without a baseline can’t show improvement. When a Security Audit finds a control missing, the implementation measure is the right starting point.
NIST’s guidance is SP 800-55 Volume 1, Identifying and Selecting Measures.
Related terms: Phishing, Security Operations, Governance