Clause 9 · Performance evaluation

9.1 — Monitoring, measurement, analysis and evaluation

The requirement that turns "we have controls" into "we know whether our controls are working," with numbers to back it up.

Mandatory requirementDocumented information required: Yes

In plain language

In plain language: sub-clause 9.1 requires you to decide, in advance, exactly what security-relevant things you will measure, how you will measure them, when, and who is responsible — then actually evaluate the results, rather than collecting data nobody looks at.

Position in the standard

9.1 · Monitoring, measurement, analysis and evaluation ← you are here
See also: 9.2 · 9.3

Why this requirement exists

Controls degrade silently. A backup job that worked at implementation can start failing months later with nobody noticing until a restore is needed. This requirement exists so that failure gets caught by a metric, not by an incident.

Scenario: a company rolled out MFA at 100% coverage a year ago and considers the control "done." No one has re-checked coverage since. In the meantime, a dozen new accounts were created without MFA enforced by default. Without a monitoring metric tracking MFA coverage monthly, that gap goes unnoticed until an auditor — or an attacker — finds it.

What the standard expects

The standard expects you to determine what needs monitoring and measuring — including information security processes and controls; the methods for monitoring, measurement, analysis, and evaluation needed to ensure valid results; when monitoring and measuring shall be performed; who shall do it; when the results shall be analyzed and evaluated; and who shall analyze and evaluate them. You must keep documented evidence of the results, and — beyond just collecting numbers — actually evaluate the information security performance and the effectiveness of the ISMS as a whole.

In practice

  • Build a short list of metrics tied to your riskiest controls — MFA coverage, patch compliance rate, backup success rate, phishing test click rate, access review completion — rather than measuring everything.
  • For each metric, define a target, a measurement frequency, and a named owner who reviews it.
  • Set a threshold that triggers action — a metric that just gets logged without a reaction defeats the purpose.
  • Feed monitoring results into management review (9.3) so leadership sees trends, not just a single snapshot.

Evidence the auditor will ask for

  • A documented list of metrics with methods, frequency, and owners.
  • Actual measurement records or dashboards over time, not a one-off snapshot taken just before the audit.
  • Evidence the results were analyzed and evaluated — meeting notes, reports, or a documented review, not just raw numbers.

Common pitfalls

  • Measuring dozens of metrics nobody reviews, instead of a handful that are actually acted on.
  • Producing a single metrics snapshot right before the audit instead of showing a real, ongoing measurement history.
  • Collecting data with no defined evaluation step — numbers with no analysis attached.

Related requirements

Parent link: Clause 9 · Performance evaluation

Frequently asked questions

How many metrics do we need?
There is no fixed number. A focused set of 5-10 metrics tied to your highest-priority risks and controls is more defensible than a long list nobody actually reviews.
Can monitoring be automated?
Yes, and it usually should be — automated dashboards pulling live data from your systems are both more reliable and easier to evidence than manual, periodic checks.

Turn control monitoring into a metric, not an incident report.

Sentrix tracks your key security metrics continuously and surfaces drift before it becomes a finding.