Metrics

Coverage metrics

The first to look at: they tell you whether the arrangement exists.

Metric Target Source
Share of products qualified as PDEs 100 % Qualification register
Share of products classified 100 % Classification register
Share of products with an automated SBOM 100 % of active products Platform
Share of SBOMs above the quality threshold 100 % Platform
Share of products under continuous monitoring 100 % of shipped versions Platform
Share of products with a declared support period 100 % Register
Share of products with complete technical documentation 100 % before market entry Compliance tracking
Share of class II / Critical products with a notified body engaged 100 % Register

Performance metrics

They tell you whether the arrangement works.

Metric Indicative target
Median detection interval (publication → your alert) < 24 h
Median triage interval (alert → decision) < 5 days
Median remediation interval, by severity Per SLAs
Share of SLAs met > 90 %
Average age of open vulnerabilities Steadily falling
Ratio of actionable alerts to raw alerts Steadily rising
Interval achieved at the last reporting exercise < 12 h for the early warning
Interval achieved at the last retrieval exercise < 1 working day

Risk metrics

They tell you what remains exposed.

Metric Meaning
Number of components at end of maintenance inside supported products Direct regulatory liability
Number of single-maintainer components in critical products Takeover or abandonment risk
Number of licences out of policy, with no active exception Intellectual property risk
Number of expired exceptions not dealt with Process defect
Number of support periods with a limiting component ending earlier Untenable commitments
Estimated financial exposure See Exposure

Readiness metrics

Metric Source
Days remaining before each legal deadline Timeline
Back-planning progress, by wave Roadmap
Date of the last reporting exercise Exercise reports
Date of the last retrieval exercise Exercise reports
Self-assessment score by module Quarterly assessment

Metrics not to track

Misleading metric Why
Number of vulnerabilities detected It rises as the arrangement improves. Presenting it to a committee creates a perverse incentive to detect less.
Number of SBOMs generated Without coverage and quality it is a file counter.
Number of alerts closed Closing an alert without handling it removes it from the counter.
Overall security score A non-actionable aggregate that hides variation between products.

Leadership dashboard mock-up

One page, four blocks.

┌──────────────────────────────┬──────────────────────────────┐
│ COVERAGE                     │ DEADLINES                    │
│                              │                              │
│ Products classified   ██  %  │ Reporting operational     ✓  │
│ SBOM automated        ██  %  │ Full application:     D-___  │
│ Monitoring active     ██  %  │ Notified bodies       __/__  │
│ Files complete        ██  %  │ Back-planning: wave __/6     │
├──────────────────────────────┼──────────────────────────────┤
│ PERFORMANCE                  │ OPEN RISKS                   │
│                              │                              │
│ Critical remediation __ d    │ Abandoned components     __  │
│ SLAs met             __ %    │ Licences out of policy   __  │
│ Last 24 h exercise   __ h    │ Untenable support        __  │
│ Actionable alerts    __ %    │ Estimated exposure    __ €m  │
└──────────────────────────────┴──────────────────────────────┘
Decisions required from the committee: …

The keeping rule

Every metric has a named owner, a single source and an update frequency. A metric computed by hand the day before the committee is not a metric: it is an estimate.