Join our Newsletter — 33% off our NHI Course
Home› Guides› Zero Trust Programme Health Check: 10 Metrics to…
Template Zero Trust

Zero Trust Programme Health Check: 10 Metrics to Track Progress

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 14 November 2026 12 min read
On this page

Most zero trust programmes report activity, such as tools bought and pilots run, rather than how much of the estate is actually governed by per-request, identity-based access decisions. This health check gives ten programme progress metrics, at least one per pillar (identity, devices, networks, applications and workloads, data) plus the cross-cutting visibility capability, each with a formula, a data source type and green, amber and red thresholds. It is run quarterly by the zero trust programme lead and reported to the security steering group. It does not place you on a maturity stage: that is the job of the Zero Trust Maturity Model Guide and the CISA Zero Trust Maturity Model Guide. It also does not check how the ZTNA broker or VPN gateway is configured; that is the ZTNA and remote access VPN health check. This page measures coverage: how far the programme has moved the estate, pillar by pillar, between assessments.

What this health check covers

The ten metrics follow the tenets in NIST SP 800-207: no implicit trust from network location or asset ownership, all communication secured regardless of location, access granted per session with the fewest privileges needed, dynamic policy using the state of identity and device, and continuous monitoring of asset posture. Each metric turns one tenet into a ratio you can collect from systems you already run. The identity design behind several of them is explained in the Zero Trust Identity Guide; this page measures it and does not repeat it.

Pillar or capabilityMetricsTenet measured (NIST SP 800-207)
IdentityZTP-03 MFA coverage, ZTP-08 continuous authentication share, ZTP-09 least privilege scoreStrictly enforced authentication and authorisation; per-session, least privilege access
DevicesZTP-02 device posture enforcementMonitor and measure the integrity and posture of all assets; dynamic policy
NetworksZTP-04 microsegmentation coverage, ZTP-05 legacy VPN users, ZTP-06 east-west policy coverage, ZTP-07 encrypted internal trafficNo trust from network location; all communication secured
Applications and workloadsZTP-01 applications behind a policy enforcement pointAccess to individual resources on a per-session basis
Visibility and analytics (cross-cutting); dataZTP-10 telemetry coverageCollect as much information as possible about the state of assets and requests

Key takeaways

  • Measure coverage of the estate, not tool deployment: every metric has the full in-scope population as its denominator, so an unmanaged application or device lowers the score.
  • Non-human identities count: service accounts, workloads and agents sit inside ZTP-03, ZTP-07, ZTP-08 and ZTP-09; a programme that measures only staff sign-ins overstates progress.
  • Legacy access is a metric of its own: users who still hold network-level VPN access show how much implicit trust remains, whatever else has been deployed.
  • Report pillars side by side: pillars move at different speeds, so present ten numbers and their trend, never one blended score.
  • Thresholds are starting points: the values below are NHI Mgmt Group recommendations; record your own target per metric and the reason for any change.

How to use this template

Copy the metrics table and the findings log into your tracker. Fix the scope first: list the in-scope applications, devices, network segments, human and non-human identities, and record the count for each, because those counts are the denominators. Then collect each figure from the data sources described, score it against the thresholds and log every amber or red result as a finding with an owner and due date. Keep each quarterly run so the trend can be reported. Re-baseline the denominators each quarter: a score that rises because the inventory shrank is not progress.

Cadence and sign-off

ReviewFrequencyOwnerApproverEvidence kept
Scope and denominator refreshQuarterly, before the KPI runZero trust programme officeZero trust programme leadApplication, device, segment and identity inventories with counts
Full KPI run (ZTP-01 to ZTP-10)QuarterlyZero trust programme leadCISO or security steering group chairKPI workbook, findings log, steering minutes
Legacy VPN and exception review (ZTP-05)MonthlyNetwork and identity engineeringZero trust programme leadVPN group export, exception register with expiry dates
Board or risk committee summaryHalf-yearlyCISOBoard or delegated committeePillar trend chart, top findings, decisions minuted
Independent re-performanceYearlyInternal auditAudit committeeRe-performed samples and variance notes

Metrics

NHI Mgmt Group recommended thresholds; adjust to your risk appetite. "In-scope application" means an application in your inventory used by staff, suppliers or services. "Policy enforcement point" (PEP), as used in NIST SP 800-207, means a component that allows, denies or ends a session based on a policy decision: an identity-aware proxy, a ZTNA broker, an identity provider with conditional access, an API gateway or a service mesh sidecar.

IDMetricFormula (numerator / denominator)Data source typeGreenAmberRed
ZTP-01Applications behind a policy enforcement pointIn-scope applications reachable only through a PEP that evaluates identity and device per session / in-scope applicationsApplication inventory; identity provider application registrations; ZTNA and proxy published application lists; firewall rules showing direct paths≥ 90%60% to below 90%< 60%, or any crown jewel application reachable directly
ZTP-02Device posture enforcementAccess policies for in-scope applications where a device compliance or posture signal is evaluated and a failure blocks or restricts access / access policies for in-scope applicationsConditional access and PEP policy exports; device management compliance policies≥ 90%70% to below 90%< 70%
ZTP-03MFA coverage, phishing-resistant share(a) Human sign-ins to in-scope applications that satisfied MFA / all human sign-ins; (b) privileged, remote and supplier sign-ins using FIDO2/WebAuthn or PKI / privileged, remote and supplier sign-insIdentity provider sign-in logs with authentication method(a) ≥ 99%; (b) ≥ 95%(a) 95% to below 99%; (b) 70% to below 95%(a) < 95%; (b) < 70%
ZTP-04Microsegmentation coverageWorkloads in scope (servers, virtual machines, containers) assigned to a segment with an enforced allow-list policy / workloads in scopeSegmentation platform, host firewall or cloud security group exports; CMDB workload list≥ 80%; 100% for crown jewel systems50% to below 80%< 50%, or any crown jewel system unsegmented
ZTP-05Legacy VPN usersActive accounts holding network-level VPN access (subnet or full tunnel) / active accounts with any remote access; plus count of those unused for 30 daysVPN group membership; VPN connection logs; directory account status≤ 10% and falling, zero unused10% to 30%, or unused accounts with a removal date> 30%, any unused account without a removal date, or any VPN access without MFA
ZTP-06East-west policy coverageObserved internal flows between workloads that match an explicit allow rule / observed internal flows (default-allow and "any" rules count as unmatched)Flow logs (cloud VPC flow logs, NetFlow, segmentation platform flow maps); rule exports≥ 90%60% to below 90%< 60%
ZTP-07Encrypted internal trafficInternal service-to-service and user-to-application connections using TLS 1.2 or later, or mutual TLS / internal connections observedFlow logs with port and protocol; TLS inspection or sensor data; service mesh mTLS mode reports≥ 95%80% to below 95%< 80%, or any cleartext authentication protocol in use
ZTP-08Continuous authentication shareSessions to in-scope applications where the PEP re-evaluates risk or device signals during the session and can revoke it (CAEP or equivalent signal) / sessions to in-scope applicationsIdentity provider session and token logs; Shared Signals receiver configuration≥ 60% and rising25% to below 60%< 25%, or session revocation untested
ZTP-09Least privilege scoreHuman and non-human identities with no standing privileged assignment and no unused entitlement older than 90 days / identities in scopeIGA and PAM exports; cloud entitlement reports; access activity data≥ 90%70% to below 90%< 70%
ZTP-10Telemetry coverageData sources needed for access decisions and investigation (identity provider, PEPs, endpoint detection, flow logs, data access logs for sensitive stores) sending to the central log platform in the last 24 hours / required data sourcesLog platform source list; required source register≥ 98%90% to below 98%< 90%, or any PEP not logging decisions

Notes on the formulas

  • ZTP-01 counts exclusivity: an application published through a proxy that is also reachable over the flat network or VPN does not count. Check firewall rules for direct paths, which is the coverage question; broker configuration belongs to the ZTNA health check.
  • ZTP-03 follows CISA guidance: CISA describes FIDO/WebAuthn as the only widely available phishing-resistant authentication, with PKI smart cards as the other strong option, and advises starting with administrators, help desk and the services that already support FIDO. See the Phishing-Resistant MFA Guide.
  • ZTP-04 and ZTP-06 differ: ZTP-04 asks whether a workload has a segment policy at all; ZTP-06 asks whether the traffic actually observed is explained by explicit rules. A high ZTP-04 with a low ZTP-06 means segments exist with permissive rules. NSA's network and environment pillar guidance treats data flow mapping as the starting point for both.
  • ZTP-07 includes machine traffic: NIST SP 800-207 states that all communication is secured regardless of network location, and OMB M-22-09 required US federal agencies to enforce HTTPS for all web and API traffic, including internal traffic. Count service accounts and workloads, not only browsers. See the mTLS Guide.
  • ZTP-08 needs a test: a session counts only if revocation on a risk or device change has been tested in the quarter. OpenID CAEP 1.0 (published 29 August 2025) defines event types such as session-revoked, credential-change and device-compliance-change. See CAEP and Continuous Access Evaluation.
  • ZTP-09 includes machines: service accounts, API keys and workload roles with standing administrator rights are often the largest gap. See the Least Privilege Techniques Guide and the Identity Security Metrics Guide.

Reading the results together

  • ZTP-01 high, ZTP-05 high: applications are published through a PEP, but the old network path was never closed. Retire VPN groups as each application moves.
  • ZTP-03 high, ZTP-02 low: strong authentication from any device. A stolen session on an unmanaged device still works.
  • ZTP-04 high, ZTP-10 low: segmentation without telemetry; blocked lateral movement will go unnoticed and rules will drift.
  • ZTP-08 low with ZTP-03 high: trust is decided once at sign-in and held for the life of the token, which is the implicit trust the programme set out to remove.

How to collect the data

  1. Fix the denominators: export application, device, workload and identity inventories (including service accounts and workload identities) and agree the in-scope list with owners.
  2. Export access policy: identity provider conditional access, ZTNA and proxy published applications, API gateway routes, for ZTP-01, ZTP-02 and ZTP-08.
  3. Pull sign-in logs: 30 days of identity provider sign-ins with authentication method and device state for ZTP-03.
  4. Pull network evidence: segmentation and firewall rule exports and 7 to 30 days of flow logs for ZTP-04, ZTP-06 and ZTP-07; VPN group membership and connection logs for ZTP-05.
  5. Pull entitlement data: IGA, PAM and cloud entitlement exports with last-used dates for ZTP-09.
  6. Check the log platform: compare active sources with the required source register for ZTP-10.
  7. Trace samples: pick five applications and five service accounts and walk each through every relevant metric. A metric that cannot be traced is reported amber at best.

Red flags that need action now

  • An active VPN account with network-level access that has not been used for 30 days or does not require MFA (ZTP-05). The Colonial Pipeline attack began with a legacy VPN account that was no longer used, still active and not protected by MFA.
  • A crown jewel application reachable directly from the internal network or VPN, bypassing the PEP (ZTP-01).
  • Privileged administrators signing in without phishing-resistant MFA (ZTP-03).
  • A default-allow or "any to any" rule between production segments (ZTP-06).
  • Cleartext authentication or management protocols observed on internal networks (ZTP-07).
  • Service accounts or workload roles with standing administrator rights and no owner (ZTP-09).
  • A PEP that does not log its allow and deny decisions centrally (ZTP-10).

Findings log

Finding IDFindingMetricSeverityOwnerDue dateStatus
HC-ZTP-01[Number] in-scope applications still reachable directly over VPNZTP-01, ZTP-05[High / Medium / Low][Role][Date][Open / In progress / Closed / Risk accepted]
HC-ZTP-02[Number] access policies with no device signalZTP-02[Severity][Role][Date][Status]
HC-ZTP-03[Percentage] of internal flows matched only by permissive rulesZTP-06[Severity][Role][Date][Status]
HC-ZTP-04[Number] service accounts with standing privileged accessZTP-09[Severity][Role][Date][Status]
HC-ZTP-05[Finding][Metric][Severity][Role][Date][Status]

Group items with a shared root cause into one finding. A risk accepted finding carries an expiry date and the approver's name and is reopened at the next quarterly run.

Framework mapping

MetricNIST SP 800-53 Rev. 5NIST CSF 2.0
ZTP-01, ZTP-02, ZTP-08AC-17 Remote Access; SI-4 System MonitoringPR.AA-05
ZTP-03IA-2(1) and IA-2(2) Multi-factor AuthenticationPR.AA-03
ZTP-04, ZTP-05, ZTP-06SC-7 Boundary Protection; SC-7(5) Deny by Default, Allow by ExceptionPR.IR-01
ZTP-07SC-8 Transmission Confidentiality and IntegrityPR.DS-02
ZTP-09AC-6 Least Privilege; AC-2 Account ManagementPR.AA-05
ZTP-10CA-7 Continuous Monitoring; SI-4DE.CM-01, DE.CM-09
The health check itselfPM-6 Measures of Performance; CM-8 System Component InventoryGV.OV-03, ID.AM-01, ID.AM-02

Remediation map

Finding typeMetricsWhere the fix is explained
Applications outside a PEP; legacy VPNZTP-01, ZTP-05ZTNA Guide; ZTNA vs VPN
Weak or missing MFA, no device signalZTP-02, ZTP-03Phishing-Resistant MFA Guide; Zero Trust Identity Guide
Segmentation and east-west gapsZTP-04, ZTP-06Network Segmentation Guide; Microsegmentation vs ZTNA
Cleartext internal trafficZTP-07mTLS Guide; Service Mesh Security Guide
Sign-in-only trustZTP-08CAEP and Continuous Access Evaluation; Shared Signals Framework Guide
Standing privilege, machine identitiesZTP-09Least Privilege Techniques Guide; Workload Identity Federation Guide
Missing telemetryZTP-10SIEM Log Source Onboarding Checklist

References

Related NHI Mgmt Group resources: Zero Trust Identity Guide, CISA Zero Trust Maturity Model Guide, ZTNA and remote access VPN health check, Identity Security Metrics Guide.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 14 November 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org