# Four numbers carry the program: time to remediate, SLA attainment, coverage of the inventory and the age of the backlog, each with a fixed formula

MEASUREMENT, read 10 Nov 2026

**The short answer.** Report four measures, each with a written formula: mean and median time to remediate by tier, the share of items closed within SLA, the share of the inventory scanned recently, and open items past due by age band. Define the formulas once, freeze them, and publish the same cut every month.

## Define before you measure

Most arguments about program metrics are arguments about definitions. Settle three things first. The population: which findings count (a tier, a source type, a business unit). The clock: when a finding starts and when it stops, using the SLA document. The unit: finding, vulnerable asset, or remediation task. A platform that groups findings into tasks will report different counts from one that reports per finding; ServiceNow's documentation describes grouping vulnerable items into remediation tasks.

Write each formula in the reporting document and keep the query beside it. When a platform changes, the formulas are what you migrate.

Fix the reporting calendar as well. Use calendar months, cut off at the same hour each month, and take a snapshot of the open-item list so that a later correction does not rewrite history. Store each month's snapshot where the auditor can reach it. A program that can reproduce last quarter's numbers from the snapshot has better evidence than one with better numbers and no snapshot.

## 1. Time to remediate

Mean time to remediate (MTTR) is the sum of (closed date minus start date) across findings closed in the period, divided by the count closed. Report by tier, never as one blended number, and add the median and the 90th percentile; a few very old items will drag a mean. MTTR is backward looking: it counts only closed items, so a program that ignores its oldest findings looks fast. Pair it with backlog age.

For KEV items, a second view is more direct: the share closed by the catalog due date. The catalog JSON carries a dueDate on each entry, and BOD 26-04 requires agencies to report KEV status through the CDM Dashboard, bi-weekly if they are not fully automated.

## 2. SLA attainment

SLA attainment is the count of items closed on or before their due date, divided by the count of items that came due in the period. Items still open and not yet due are excluded; items open and past due count against. Report by tier. Handle exceptions explicitly: an approved exception with a valid expiry is excluded and listed; an expired one counts as past due.

Brinqa documents an SLA data model with deadlines by severity and a different SLA for assets with compliance requirements; Hackuity describes SLA compliance and MTTR in its executive view, and Zscaler describes dashboards built on KPIs and SLAs. These are vendor statements. Whatever the tool computes, reproduce one month by hand from the raw export.

## 3. Coverage

Coverage is the count of in-scope assets with a successful assessment in the last N days divided by the count of in-scope assets in the system-of-record inventory. Choose N by asset class: FedRAMP's Vulnerability Detection and Response rules, for example, say providers should check assets likely to drift at least every 14 days and others at least monthly, and non-machine resources every 3 months. Those are provider expectations, but they make workable defaults.

Report coverage by source type too: infrastructure scans, application security scans, cloud posture. A single percentage hides a class with none. Coverage is also where the Connector Census matters: a source with no working connector is a source with no coverage in the ranking.

Report age against the tier's SLA, not in absolute days. A 40-day-old P1 item is a failure; a 40-day-old P4 item with a 90-day SLA is on time. The ratio of age to SLA makes the tiers comparable and makes one chart enough for a leadership review.

## 4. Backlog age

Backlog age is the distribution of open items past their due date, in bands. A workable set is 1 to 30, 31 to 90, 91 to 180 and over 180 days past due, by tier. Show the count and the change since last month. The oldest-item age is a useful single number; it moves slowly and cannot be hidden by closing easy items.

Add the net flow: opened minus closed in the period, by tier. A backlog that grows while MTTR looks good means intake exceeds capacity.

## Reporting rhythm and traps

Publish the same four measures monthly, with a one-line explanation for each movement of more than 10 percent. Review the formulas annually, and version them.

- Do not report counts of vulnerabilities found as a success measure; the count rises when coverage improves.
- Do not average across tiers. Tier 1 and tier 4 are different populations.
- Watch the threat input as a measure of its own. The KEV Lag Ledger finds a median of 10.5 days from NVD publication to listing, so a program that re-ranks weekly will lag behind a catalog that adds entries daily.
- Check measures after a platform change against the old tool for one overlapping month.

## Sources

- CISA, BOD 26-04, Prioritizing Security Updates Based on Risk: https://www.cisa.gov/news-events/directives/bod-26-04-prioritizing-security-updates-based-risk (read 10 Nov 2026)
- CISA, Known Exploited Vulnerabilities catalog (JSON feed): https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json (read 10 Nov 2026)
- FedRAMP, Consolidated Rules for 2026, Vulnerability Detection and Response: https://www.fedramp.gov/2026/reference/rev5/c/vulnerability-detection-and-response/ (read 10 Nov 2026)
- Brinqa documentation, SLA data model: https://docs.brinqa.com/docs/sla (read 10 Nov 2026)
- ServiceNow documentation, Vulnerability Response: https://www.servicenow.com/docs/api/khub/maps/x8~ohSTeWLioGLVI5x3Vgg/topics/16Miv0EjmvVF7qlqVbs6Aw/content (read 10 Nov 2026)

Last reviewed 10 Nov 2026. Canonical: https://riskbasedvulnerabilitymanagement.com/guides/measuring-the-program/
