Key Risk Indicators (KRIs) are the answer, but only when they’re chosen deliberately, defined precisely, and weighted in context. This reference library is built for practitioners who already have security controls and want to extract meaningful risk signal from them. Every domain maps explicitly to one or more CIS Controls v8.1, giving you a clear line from KRI to framework coverage.

A KPI tells you if your security program is functioning. A KRI tells you if your organization is exposed. Both matter. Only one belongs in front of a board or a cyber insurer.

How to use this library

Each KRI entry answers three questions:

  • What you’re measuring, the specific metric and its unit
  • Why it creates or reduces risk exposure, the business consequence
  • Threshold signals, the values that indicate green, amber, or red posture

Each domain header identifies which CIS Controls v8.1 it satisfies. A full mapping table appears in the appendix at the end of this document.

Weighting guidance is relative and contextual. A fintech’s MFA adoption rate carries more weight than a traditional manufacturer’s. A manufacturer’s OT patch latency may be existential in ways that don’t appear in a SaaS company’s risk model. Use the base weights as a starting point, then apply your own context multipliers.

1. Asset & Software Inventory

Base weight · Critical CIS 1, 2

Do you know what you own, hardware and software, and is all of it authorized?

You cannot protect assets you don’t know exist. Asset and software inventory sit at the foundation of every other control in this library. Gaps here compound across every other domain: you can’t patch what you can’t see, you can’t monitor what isn’t inventoried, and you can’t respond to incidents on systems that aren’t tracked.

Hardware asset inventory completeness

MeasurePercentage of known active assets confirmed in the authorized enterprise asset inventory, reconciled against network discovery scans.

WhyAsset inventory is the prerequisite for patching, monitoring, access control, and incident response. Unknown assets are unprotected assets. Discovery scans routinely surface 10–20% more devices than organizations believe they have.

Green>98% reconciled; unauthorized assets remediated within 24 hours
Amber95–97% reconciled, or 24–72 hours to resolve unauthorized assets
Red<95% reconciled, or unauthorized assets persisting >72 hours without exception

End-of-life asset rate

MeasurePercentage of assets running hardware or operating systems past vendor end-of-life (EOL) with no active support contract.

WhyEOL assets receive no security patches. A single unpatched EOL system in a critical segment is a permanent, unfixable exposure that grows worse every day.

Green<1% EOL; exceptions documented with compensating controls
Amber1–5% EOL with documented risk acceptance
Red>5% EOL, or any EOL asset in a critical or internet-facing role

Software inventory coverage

MeasurePercentage of installed software across the environment documented in the authorized software inventory.

WhyUnauthorized software is a primary malware delivery vector. Software you don’t know about can’t be patched, monitored for vulnerabilities, or removed when compromised. Shadow IT at the software layer is as dangerous as at the hardware layer.

Green>95% authorized; unauthorized software removed within 48 hours
Amber90–94%, or 48–96 hours to remediate
Red<90%, or no active unauthorized software detection

End-of-life software rate

MeasurePercentage of software instances running versions past vendor end-of-support.

WhyUnsupported software accumulates unpatched CVEs permanently. Even software that appears to function normally is a ticking vulnerability clock once it crosses the EOL threshold.

Green<2% EOL; documented compensating controls and remediation timelines
Amber2–5% with active migration plans
Red>5%, or any EOL software in production systems handling sensitive data

2. Data Protection

Base weight · High (Critical for regulated industries) CIS 3

Where is sensitive data, and is it actually protected?

Data protection KRIs are particularly relevant for regulatory exposure. Breaches are measured in records, regulators fine based on data classification, and insurers underwrite based on what you hold and how you protect it.

Data classification inventory coverage

MeasurePercentage of data stores with a documented sensitivity classification (public, internal, confidential, restricted).

WhyYou cannot apply the right protection controls to data you haven’t classified. Classification is the prerequisite for encryption, access control, DLP, and breach notification decisions. Unclassified data is implicitly treated as low risk even when it isn’t.

Green>90% of known data stores classified
Amber75–89%
Red<75%, or no formal classification scheme in place

Sensitive data encryption coverage

MeasurePercentage of data stores containing confidential or restricted data encrypted at rest.

WhyUnencrypted sensitive data converts any unauthorized access event into a reportable breach. Encryption at rest is the last line that converts a loss event into a non-event from a regulatory perspective. It is also one of the first things a cyber insurer verifies at renewal.

Green100% of classified stores encrypted at rest
Amber95–99% with active remediation
RedAny classified data store unencrypted, regardless of access controls

Data in transit encryption rate

MeasurePercentage of sensitive data transfers using current encryption standards (TLS 1.2 minimum; TLS 1.3 preferred).

WhyPlaintext transmission of sensitive data over any network, internal or external, is an interception risk. Deprecated protocols (SSL, TLS 1.0/1.1) provide false assurance.

Green100% of sensitive transfers on TLS 1.2+
AmberAny deprecated protocol with documented upgrade timeline
RedAny unencrypted sensitive transmission, or deprecated protocol without remediation plan

Cloud storage misconfiguration rate

MeasureNumber of cloud storage resources (S3 buckets, Azure blobs, GCS buckets) publicly accessible without authentication.

WhyPublicly accessible cloud storage remains one of the most common causes of data exposure, and among the most embarrassing to explain to regulators or customers. Many exposures persist for months before discovery because no one is monitoring for them.

GreenZero public storage with sensitive data; intentional public storage documented
AmberPublic storage without sensitive data, with documented business justification
RedAny public storage containing classified or sensitive data

DLP policy bypass rate

MeasureRatio of sensitive data transfer attempts bypassing DLP controls versus total detected attempts.

WhyHigh block rates with low bypass rates signal effective enforcement. High bypass rates, even with active DLP tooling, indicate data leaving channels your tools cannot inspect. Track trends, not point-in-time values.

Green<0.5% bypass; trend stable or improving
Amber0.5–2%
Red>2%, or upward trend over 60+ days

3. Secure Configuration Management

Base weight · High CIS 4

Are your systems hardened, or are they running out of the box?

Default configurations are designed for compatibility and ease of use, not security. Every asset deployed without a hardened baseline carries unnecessary attack surface, default credentials, unnecessary services, permissive network rules, and unreviewed settings that attackers have already catalogued. These KRIs measure whether your configuration controls are real or theoretical.

CIS Benchmark compliance score

MeasurePercentage of assets passing automated CIS Benchmark configuration checks for their asset class (servers, workstations, network devices, cloud accounts).

WhyCIS Benchmarks are the most widely used hardening standard. Low compliance scores indicate assets deployed with default or overly permissive configurations, the attack surface attackers exploit before they need to exploit vulnerabilities.

Green>90% of assets passing applicable benchmark profile
Amber75–89%
Red<75%, or no automated configuration compliance scanning

Configuration drift rate

MeasurePercentage of assets that have deviated from their approved configuration baseline since last verified check.

WhyConfiguration drift is insidious, it happens gradually through routine changes, software installs, and administrative shortcuts. Drift accumulates silently until it becomes a security incident. Assets that drift from baseline are running configurations no one has reviewed for security implications.

Green<5% drift; deviations reviewed and remediated or documented
Amber5–15% with active remediation
Red>15% drift, or no baseline deviation detection

Default credential elimination rate

MeasurePercentage of deployed assets (network devices, servers, applications, IoT) confirmed to have default credentials changed.

WhyDefault credentials are published by manufacturers and indexed by search engines. Shodan and similar tools allow anyone to find internet-accessible devices with default credentials in minutes. This is not a sophisticated attack, it is the equivalent of a locked door with the key taped to it.

Green100% of managed assets with non-default credentials; verified by automated scanning
Amber95–99%, with manual verification for remaining
RedAny internet-facing asset with default credentials; or no default credential scanning

Unnecessary service exposure rate

MeasurePercentage of assets running services, ports, or protocols not required for their documented business function.

WhyEvery unnecessary service running on an asset is attack surface that doesn’t need to exist. Unnecessary services accumulate through convenience installations, legacy software, and missed decommissioning steps.

Green<3% of assets with undocumented services running
Amber3–10%
Red>10%, or any internet-facing asset with undocumented services

4. Identity & Access Management

Base weight · Critical CIS 5, 6

Who can get in, and how well is that controlled?

Identity is the primary attack vector in the majority of breaches. Attackers don’t break in, they log in. These KRIs surface credential and access exposure before it becomes an incident.

MFA enrollment rate

MeasurePercentage of users and accounts with active multi-factor authentication enrolled.

WhyEach un-enrolled account is a credential-stuffing target. One compromised account in a privileged role is a breach, not a security event, a breach.

Green>95% enrolled; 100% for privileged and remote-access accounts
Amber85–94%
Red<85%, or any privileged account without MFA

Privileged account sprawl

MeasureCount of admin and root accounts versus a defined and approved baseline.

WhyOver-provisioned admin accounts expand blast radius. Each unneeded privilege is a weapon waiting to be found. Tends to grow silently during rapid hiring and migration projects.

GreenAt or below baseline; privileged accounts reviewed quarterly
Amber10–25% above baseline
Red>25% above, or undocumented admins in production

Dormant account rate

MeasurePercentage of accounts with active credentials and no activity in more than 90 days.

WhyFormer employees, contractors, and service accounts that were never properly offboarded. A classic attacker persistence vector, dormant accounts are reactivated because no one is watching them.

Green<1% of total accounts; automated detection and disablement workflow
Amber1–3%
Red>3%, or any dormant account with privileged access

Least privilege compliance rate

MeasurePercentage of user accounts confirmed to have only the permissions required for their documented role, verified through access reviews.

WhyPermission creep is universal. Users accumulate access over time, through role changes, project assignments, and convenience grants, and those permissions are almost never revoked. The result is a user population with far broader access than anyone intended or reviewed.

Green>85% validated through formal access review in the last 12 months
Amber70–84%
Red<70%, or no formal access review in the last 12 months

Service account audit coverage

MeasurePercentage of service accounts with a documented owner and a recorded last-review date.

WhyUnmanaged service accounts are the accounts attackers pivot through. No one notices unusual activity because no one is watching, and no one owns remediation when something goes wrong.

Green>90% documented and reviewed within 12 months
Amber75–89%
Red<75%, or any privileged service account with no documented owner

5. Continuous Vulnerability Management

Base weight · Critical CIS 7

How long does exploitable exposure persist?

Time is the fundamental variable in vulnerability risk. The question isn’t whether you have vulnerabilities, you do. The question is how long they remain exploitable.

Mean time to remediate (MTTR), critical CVEs

MeasureAverage number of days from discovery to confirmed remediation for critical-severity vulnerabilities.

WhyAttackers weaponize critical CVEs within hours of public disclosure in many cases. Every day above your SLA is measurable, quantifiable exposure. MTTR is the single most underreported metric in vulnerability programs.

Green<15 days
Amber15–30 days
Red>30 days, or no defined remediation SLA

Unpatched critical vulnerabilities >30 days

MeasurePercentage of assets carrying unpatched critical vulnerabilities beyond the 30-day remediation window.

WhyA direct measure of SLA compliance and remediation capacity. High rates indicate resource gaps or an organization silently accepting unacknowledged risk. This is what your cyber insurer is looking at.

Green<5% of assets
Amber5–15%
Red>15%, or any critical business system in this category

Internet-facing asset patch rate

MeasurePercentage of internet-exposed assets fully patched at the current remediation cycle.

WhyExternal attack surface is higher priority than internal. An unpatched internet-facing server isn’t a risk, it’s an invitation. Perimeter assets should be tracked separately from internal ones.

Green>98%
Amber92–97%
Red<92%, or any critical internet-facing asset unpatched >14 days

Known Exploited CVE exposure

MeasureNumber of assets running software with CVEs listed on the CISA Known Exploited Vulnerabilities (KEV) catalog.

WhyCISA’s KEV list tracks vulnerabilities with confirmed exploitation in the wild. These aren’t theoretical risk, they’re being actively used by threat actors. Any exposure here is direct operational risk, not backlog.

GreenZero
Amber1–5 non-critical assets with active remediation
RedAny critical or internet-facing asset, or >5 assets total

Vulnerability scan coverage rate

MeasurePercentage of in-scope assets scanned for vulnerabilities within the defined scan frequency window.

WhyA vulnerability you haven’t scanned for is a vulnerability you don’t know you have. Coverage gaps in scanning, often caused by network segmentation, agent deployment failures, or undiscovered assets, create blind spots that persist until something goes wrong.

Green>98% of in-scope assets scanned within defined frequency
Amber90–97%
Red<90%, or critical systems missed in last scan cycle

6. Audit Log Management

Base weight · High CIS 8

Can you reconstruct what happened, before, during, and after an incident?

Audit logs are the foundation of detection, investigation, and forensics. Without comprehensive, tamper-resistant, centrally retained logs, you are flying blind, during an incident and during the inevitable post-incident review. Regulators, insurers, and auditors all require documented evidence of logging capability. These KRIs tell you if your logging infrastructure is actually working.

Critical asset log coverage rate

MeasurePercentage of critical assets (servers, network devices, authentication systems, cloud control planes) with active, centralized log forwarding confirmed.

WhyA log that exists only on the device it came from is useless in a breach, attackers routinely clear local logs during intrusions. Centralized logging is the baseline. Coverage gaps mean incidents in those systems will be uninvestigable.

Green>98% with confirmed centralized log forwarding
Amber90–97%
Red<90%, or any authentication or privileged-access system without centralized logging

SIEM ingestion completeness

MeasurePercentage of expected log sources actively confirmed as ingesting into SIEM within the last 24 hours.

WhyLog sources fail silently. A SIEM that appears healthy may have stopped receiving logs from critical systems days or weeks ago, and no alert fires because the absence of data isn’t an event. This metric requires active monitoring of expected-source inventory versus confirmed-receiving sources.

Green>99% actively ingesting; source-loss alert within 1 hour
Amber95–98%, or source-loss detection >4 hours
Red<95%, or no active monitoring for source loss

Log retention compliance rate

MeasurePercentage of log categories retained for the full period required by policy, regulatory requirement, or contractual obligation.

WhyRegulatory investigations and forensic analysis routinely require log data going back 12–24 months. Retention gaps discovered during an investigation cannot be retroactively filled. Common causes: storage cost pressure, misconfigured rotation policies, and system migrations with no log transfer.

Green100% of required log categories retained for full period
Amber95–99% with gaps documented and under remediation
RedAny required category below threshold, or no defined retention policy

Audit log integrity rate

MeasurePercentage of critical log streams with tamper detection or integrity verification controls enabled (immutable storage, cryptographic hashing, write-once destinations).

WhyLogs that can be modified are not evidence. Sophisticated attackers target logging infrastructure specifically to cover tracks. Log integrity controls ensure that if logs exist, they can be trusted.

Green>95% of critical streams with integrity controls confirmed
Amber80–94%
Red<80%, or authentication and privileged logs without integrity controls

7. Email & Web Browser Protections

Base weight · High CIS 9

Are your highest-volume attack vectors actually defended?

Email and web browsing are responsible for the overwhelming majority of initial access events, phishing, malicious downloads, drive-by compromise, and business email compromise. Technical controls in this domain reduce the blast radius of human behavior: even a user who clicks a malicious link may be protected by web filtering; a phishing email blocked at the gateway never reaches anyone to click.

Email authentication protocol enforcement

MeasurePercentage of owned email-sending domains with SPF, DKIM, and DMARC configured and enforced at policy=reject or policy=quarantine.

WhySPF, DKIM, and DMARC prevent domain spoofing, the technical mechanism behind most business email compromise (BEC) attacks. Without DMARC at enforcement policy, anyone can send email that appears to come from your domain. BEC losses now exceed ransomware losses annually.

Green100% of domains with SPF, DKIM, and DMARC at reject/quarantine
AmberAll configured but one or more at policy=none
RedAny sending domain without DMARC, or DMARC universally at policy=none

Email security gateway effectiveness

MeasureRate of malicious emails reaching end-user inboxes past email security gateway controls (phishing, malware, BEC).

WhyYour email gateway is the first line of defense for the highest-volume attack vector. Gateway bypass rates, measured through threat intelligence feeds or post-incident analysis, tell you how much you’re relying on your users as a backup control.

Green<0.1% bypass rate; tracked with threat-intel correlation
Amber0.1–0.5%
Red>0.5%, or no gateway effectiveness tracking

Web content filtering coverage

MeasurePercentage of users with active DNS or proxy-based web content filtering applied to all browsing sessions, including remote and mobile.

WhyWeb filtering blocks malicious download sites, command-and-control infrastructure, and phishing pages before they load. Coverage that only applies in-office is increasingly irrelevant, remote and mobile users need the same protection.

Green>98% with filtering for all sessions regardless of location
Amber90–97%, or on-premises only
Red<90%, or no filtering for remote users

Browser / email client patching rate

MeasurePercentage of endpoints running current or N-1 supported versions of email clients and web browsers.

WhyBrowsers and email clients are the most actively exploited client-side applications. They receive frequent security updates because vulnerabilities are found and weaponized continuously. Running outdated versions is a persistent, measurable risk.

Green>95% on current or N-1
Amber85–94%
Red<85%, or any endpoint more than two major releases behind

8. Malware Defenses & Endpoint Security

Base weight · High CIS 10, 1 (partial)

Are your endpoints a controlled boundary or an open field?

Endpoints are the most common initial access point. These KRIs measure the actual enforcement of your endpoint policy, not whether a tool is purchased, but whether it is deployed, active, and working.

EDR agent coverage

MeasurePercentage of endpoints with an active, reporting endpoint detection and response agent.

Why“We have CrowdStrike” is not a risk control. 85% coverage means 15% of endpoints are invisible to your detection capability. Gaps tend to cluster on older hardware and remote offices, the same assets most likely to be targeted in attacks designed to avoid detection.

Green>98% with active, reporting agent
Amber90–97%
Red<90%, or gaps concentrated in critical or internet-connected classes

Anti-malware solution effectiveness

MeasureDetection and prevention rate for known malware categories, measured through threat intelligence correlation or red team testing.

WhyAnti-malware coverage is meaningless without effectiveness data. An agent deployed but running in audit-only mode, or with outdated signatures, or misconfigured to exclude critical directories, provides false assurance without real protection.

GreenAll agents in prevention mode; current signatures; effectiveness tested periodically
AmberDeployed but not formally tested; or a subset in audit-only
RedAgents in audit-only at scale; or signatures >7 days out of date

Full-disk encryption rate

MeasurePercentage of endpoints with full-disk encryption enabled and verified through MDM or endpoint management.

WhyEvery unencrypted device that leaves a building, lost, stolen, or abandoned, is a potential reportable breach event. Pure downside risk with a simple technical control.

Green>99% with encryption confirmed via management tooling
Amber95–98%
Red<95%, or any executive or privileged-user device unencrypted

Host-based firewall compliance rate

MeasurePercentage of endpoints with host-based firewall enabled and configured per approved policy.

WhyHost-based firewalls provide a last line of network defense directly on the endpoint, particularly important for remote workers and laptops that operate outside the corporate perimeter. A disabled or misconfigured host firewall is an undetected gap.

Green>98% enabled and policy-compliant
Amber90–97%
Red<90%, or no centralized enforcement

Unmanaged device rate

MeasurePercentage of devices detected on the network that are not registered in MDM or asset inventory.

WhyShadow IT is invisible risk. Personal devices, lab hardware, and unregistered IoT all represent uncontrolled attack surface. You cannot patch, monitor, or respond to devices you don’t know exist.

Green<2% unmanaged; automated quarantine for unknown devices
Amber2–5%
Red>5%, or any unmanaged device on sensitive network segments

9. Data Recovery

Base weight · High CIS 11

When ransomware hits, what actually survives?

Data recovery capability is the ultimate test of resilience. Ransomware groups have made backup destruction their primary tactic, they know that organizations without recoverable backups have no negotiating position. These KRIs measure whether your recovery capability is real or theoretical.

Backup coverage rate

MeasurePercentage of critical systems and data stores with automated backups running on the defined schedule.

WhyBackup coverage gaps are invisible until you need them. Systems added without going through a formal provisioning process frequently get missed in backup schedules, often the same systems that would be most painful to lose.

Green100% of critical systems backed up per schedule; automated coverage monitoring
Amber95–99% with gap discovery and remediation
Red<95%, or critical systems without confirmed coverage

Backup integrity verification rate

MeasurePercentage of critical data backups successfully tested for restoration within the last 90 days.

WhyUnverified backups fail when you need them most. Ransomware groups specifically target backup infrastructure. An untested backup is an assumption, not a recovery option, and discovering it’s broken during an incident is the worst time to find out.

Green>95% verified restorable; immutable or air-gapped copy confirmed
Amber80–94% verified, or air-gap not confirmed
Red<80% verified, or no restoration testing in last 90 days

Recovery Time Objective (RTO) validation

MeasurePercentage of critical systems with documented RTOs validated through a tested recovery exercise within the last 12 months.

WhyAn RTO of “4 hours” that has never been tested is not a recovery commitment, it’s a guess. Untested RTOs almost universally prove optimistic when actually needed.

Green>90% of critical systems validated through exercise
Amber70–89%
Red<70%, or no RTO validation exercises in the last 12 months

10. Network Infrastructure & Monitoring

Base weight · High (Critical for manufacturing, OT, infrastructure) CIS 12, 13

Is your network infrastructure controlled, and are threats being detected on it?

Unexplained external exposure

MeasureDelta between open internet-facing ports and services versus the approved and documented baseline.

WhyEvery service exposed to the internet that isn’t in your approved inventory is a potential beachhead. Shadow services accumulate after migrations, developer environments, and vendor integrations. The delta is your unknown external attack surface.

GreenZero delta; unauthorized exposure remediated within 24 hours
Amber1–3 unexplained services with low-risk profile and active remediation
RedAny unexplained service on a critical port, or delta >7 days

Firewall rule hygiene score

MeasurePercentage of firewall rules flagged as any/any, missing logging, or missing a documented owner.

WhyPermissive firewall rules accumulate over years. Each undocumented “any/any” rule is an unintentional permission no one remembers authorizing and no one wants to remove for fear of breaking something. A leading indicator of control decay.

Green<2% flagged; all under active review
Amber2–8%
Red>8%, or any any/any rule on sensitive or critical segments

Remote access authentication strength

MeasurePercentage of VPN and remote access sessions authenticated with MFA.

WhyRemote access without strong authentication is the primary ransomware delivery mechanism. A single unprotected VPN credential is a company-wide risk. This metric should be 100%.

Green100%
Amber95–99% with documented exceptions
RedAny session without MFA, or exceptions without documented approval

Network monitoring coverage rate

MeasurePercentage of network segments with active intrusion detection, traffic analysis, or network detection and response (NDR) coverage.

WhyPerimeter firewalls prevent some attacks; network monitoring detects the ones that get through. Most organizations have perimeter coverage but significant blind spots inside the network, particularly in OT segments, cloud VPCs, and legacy network zones.

Green>90% of segments with active monitoring and alerting
Amber75–89%
Red<75%, or critical segments (cloud, OT, sensitive-data) unmonitored

DNS filtering coverage

MeasurePercentage of endpoints and users with DNS-level filtering applied, blocking known malicious domains.

WhyDNS filtering blocks malware command-and-control, phishing domains, and malicious downloads at the resolution layer, before any content reaches the endpoint. One of the highest-value, lowest-friction controls available.

Green>98% of endpoints filtered for all network paths
Amber85–97%, or gaps in remote/mobile coverage
Red<85%, or no DNS filtering applied

11. Human Risk & Security Awareness

Base weight · Medium CIS 14

Are your people a vulnerability or a control?

Human risk KRIs carry medium baseline weight because technology controls should bound the blast radius of human error. In organizations with immature tooling, they move up significantly. Track human risk metrics, but don’t let them substitute for fixing the technical controls that make human error consequential.

Phishing simulation click-through rate

MeasurePercentage of targeted users who click a link in a simulated phishing exercise.

WhyA leading indicator of security awareness maturity and training effectiveness over time. Track the trend across quarters, not the point-in-time number.

Green<5% click rate; trending downward
Amber5–15%
Red>15%, or trending upward over consecutive quarters

High-risk user training completion

MeasurePercentage of repeat clickers and privileged users current on targeted security awareness training.

WhyAggregate completion rates mask the users who matter most. Repeat clickers and users with privileged access are your highest-consequence human risk population. Track them separately and assign targeted training.

Green>95% of the high-risk cohort current
Amber80–94%
Red<80%, or no high-risk cohort defined

Security incident report rate

MeasureRate at which employees report suspicious emails, events, or anomalies to the security team.

WhyLow reporting rates indicate low awareness or a culture where people don’t feel safe reporting mistakes. Both suppress your ability to detect social engineering early, before it becomes an incident.

GreenTrending upward YoY; reports reviewed with feedback provided
AmberFlat with no clear trend
RedTrending down, or no formal reporting mechanism

12. Third-Party & Supply Chain Risk

Base weight · High CIS 15

How exposed are you through vendors you trust?

Your security posture is bounded by the weakest link among your vendors. Supply chain attacks have become the preferred vector for targeting organizations that are otherwise well-defended. Most TPRM programs generate questionnaire responses, not risk signals.

Critical vendor assessment currency

MeasurePercentage of Tier 1 (critical-dependency) vendors with a completed security assessment within the last 12 months.

WhyAn assessment from three years ago is not a risk control. Vendor posture changes, leadership turns over, security programs atrophy, cloud architectures shift. Currency of assessment is as important as the assessment result.

Green>90% of Tier 1 current; Tier 1 defined and documented
Amber75–89%
Red<75%, or any Tier 1 >24 months stale

Vendor data access + MFA enforcement

MeasurePercentage of vendors with access to sensitive customer or operational data that can confirm MFA enforcement for their own personnel.

WhyA vendor with access to your customer data who doesn’t enforce MFA internally is your problem. Their breach becomes your breach, and your regulatory notification obligation.

Green100% of data-access vendors confirmed
Amber90–99% with remediation plan for gaps
RedAny data-access vendor unable to confirm MFA

Vendor contract security clause coverage

MeasurePercentage of vendor contracts with material access to sensitive data or critical systems that include defined security requirements, breach notification obligations, and audit rights.

WhyContractual security clauses are the mechanism for enforcing vendor accountability. Without them, you have no recourse when a vendor’s breach exposes your data and no notification timeline to plan around.

Green>95% of material vendor contracts with security requirements
Amber80–94%
Red<80%, or critical data-access vendors lacking contractual obligations

Vendor concentration risk

MeasurePercentage of critical business functions dependent on a single vendor with no documented fallback.

WhyThe CrowdStrike incident showed how a single vendor update can halt global operations. Concentration in critical infrastructure is systemic risk, not individual vendor risk.

GreenNo undocumented single-vendor critical functions
Amber1–2 single-vendor functions with active mitigation plans
RedAny undocumented single-vendor dependency in a critical function

13. Application Software Security

Base weight · High (Critical for software product companies, SaaS, fintech) CIS 16

Are you building security in, or bolting it on?

Application vulnerabilities are the entry point for a significant proportion of web-based breaches. Organizations that develop software, customer-facing applications, internal tools, or APIs, carry application security risk that doesn’t appear in infrastructure-focused security programs. These KRIs measure whether your software development process is producing secure code or quietly accumulating technical security debt.

SAST / DAST pipeline coverage

MeasurePercentage of active applications with automated static (SAST) and dynamic (DAST) security testing integrated into the CI/CD pipeline.

WhySecurity testing that runs in the CI/CD pipeline catches vulnerabilities before they reach production. Manual security reviews at release gates are too slow, too inconsistent, and too easy to skip under delivery pressure.

Green>90% SAST in pipeline; >75% DAST
Amber70–89% SAST; DAST inconsistently applied
Red<70% SAST, or SAST/DAST not integrated into pipeline

Third-party component vulnerability rate

MeasurePercentage of applications with known high or critical CVEs in direct or transitive open-source dependencies, as identified by software composition analysis (SCA).

WhyModern applications are 80–90% open-source code by volume. Log4Shell, Struts, and Spring4Shell were all vulnerabilities in dependencies, not in code organizations wrote. Unmanaged open-source risk is one of the least-visible sources of exploitable vulnerability.

Green<5% with unresolved high/critical CVEs in dependencies; SCA in all pipelines
Amber5–15%, or SCA scanning present but not gated
Red>15%, or no SCA in place

Secure SDLC adoption rate

MeasurePercentage of development teams following a documented secure software development lifecycle (SDLC) with defined security gates (threat modeling, security review, penetration testing at release).

WhySecure SDLC is the organizational practice that makes application security systematic rather than ad hoc. Teams without a defined secure SDLC make security decisions inconsistently, or skip them entirely under release pressure.

Green>80% of teams with documented secure SDLC and verified gates
Amber50–79%
Red<50%, or no formal secure SDLC

Web Application Firewall (WAF) coverage

MeasurePercentage of internet-facing web applications protected by a WAF in blocking mode.

WhyWAFs provide a compensating control layer for known web application attack classes (OWASP Top 10) while underlying vulnerabilities are being remediated. A WAF in detection mode provides visibility but not protection.

Green100% of internet-facing apps behind WAF in blocking mode
AmberWAF present but in detection/monitor mode on some
RedAny internet-facing application without WAF coverage

14. Incident Response Management

Base weight · High (Weighted heavily by cyber insurers) CIS 17

How fast can you detect, contain, and recover?

Prevention fails. The question is not if you’ll have an incident but how quickly you’ll know, and how much damage accumulates before you contain it. IR readiness KRIs are the ones cyber insurers weight most heavily, and the ones that most directly predict financial loss magnitude.

Mean time to detect (MTTD)

MeasureAverage hours from a security event occurring to the security team becoming aware of it.

WhyIndustry median dwell time is still measured in weeks. Every hour an attacker operates undetected is lateral movement, data exfiltration, and persistence establishment.

Green<4 hours for high-severity events
Amber4–24 hours
Red>24 hours average, or no MTTD measurement

Mean time to contain (MTTC)

MeasureAverage hours from initial detection to confirmed containment of a security incident.

WhyDetection is meaningless without timely response. A fast MTTD paired with a slow MTTC still produces large breach scope. Containment speed is the primary driver of breach cost.

Green<2 hours for critical incidents
Amber2–8 hours
Red>8 hours average, or no MTTC measurement

Tabletop exercise recency

MeasureDays since the last formal incident response tabletop exercise with documented outcomes and action items.

WhyIncident response is a muscle. An IR plan that hasn’t been rehearsed is a plan that won’t function under pressure. Regulators under SEC Cyber Rules, DORA, and NIS2 increasingly require documented exercise cadence.

Green<180 days; documented outcomes and action items remediated
Amber180–365 days
Red>365 days, or no exercise on record

IR plan currency

MeasureDays since the incident response plan was last reviewed, updated, and approved by relevant stakeholders.

WhyAn IR plan written before your current cloud architecture, vendor relationships, or regulatory obligations was written for a different organization. Outdated IR plans fail in ways their authors didn’t anticipate.

GreenReviewed within 12 months; contacts and escalation paths verified
Amber12–24 months since last update
Red>24 months, or no formal IR plan documented

15. Penetration Testing

Base weight · Medium (Increases post-architecture-change or at IG2/IG3) CIS 18

Have you verified your defenses by trying to break them?

Penetration testing validates that your security controls work as intended, not just that they exist. Internal assessments and automated scanning miss entire categories of attack: chained vulnerabilities, misconfiguration combinations, social engineering paths, and logic flaws that no scanner identifies. Pen testing is the reality check for your security program.

External penetration test recency

MeasureDays since last external network penetration test conducted by a qualified third party.

WhyExternal pen tests simulate the most common attacker perspective, an adversary on the internet attempting to gain a foothold. Annual testing is the minimum; organizations with frequent architecture changes or high-value targets need more frequent cadence. This is one of the most commonly asked questions at cyber insurance renewal.

GreenWithin 12 months; scope includes all internet-facing assets
Amber12–18 months
Red>18 months, or scope excludes material internet-facing systems

Web application penetration test recency

MeasureDays since last web application penetration test for customer-facing or critical internal applications.

WhyWeb application pen tests go beyond what automated scanners find, they identify business logic flaws, authentication bypasses, authorization failures, and chained vulnerabilities that SAST and DAST miss. Application architecture changes make previous test results stale faster than infrastructure changes.

GreenTested within 12 months; or within 3 months of a major release/architecture change
Amber12–18 months for stable applications
Red>18 months, or untested apps released in the last 12 months

Critical finding remediation rate

MeasurePercentage of critical and high-severity penetration test findings from the most recent test that have been confirmed remediated.

WhyA penetration test that produces a report that sits unactioned is not a security control, it’s an expensive document. Finding remediation rate measures whether the program produces outcomes, not just reports.

Green>90% of critical within 30 days; >80% of high within 60 days
Amber70–89% of critical within 30 days
Red<70% of critical remediated, or no tracking/verification process

Mean time to remediate pen test findings

MeasureAverage days from pen test finding identification to confirmed remediation, tracked by severity.

WhyFinding remediation speed tells you whether your security team treats pen test output as priority work or advisory backlog. Organizations that treat pen testing seriously close critical findings in days, not months.

GreenCritical <30 days; High <60 days; Medium <90 days
AmberCritical 30–60 days; High 60–90 days
RedCritical findings open >60 days, or no remediation SLA

How to weight and combine KRIs

Individual KRIs are signals. Risk posture is the composite. Combining them well requires a weighting methodology, not a single formula, but a structured way to account for context.

Three weighting factors

Apply all three to each KRI domain. Final weight = base weight × context multiplier × velocity factor.

Factor 1

Base weight

The inherent importance of this domain across most organizations. Asset inventory and identity management are universally critical. Penetration testing carries medium base weight because it validates other controls rather than being a control itself.

Factor 2

Context multiplier

Industry, regulatory regime, and business model. Healthcare weights data protection near-absolutely. Fintech weights IAM similarly. A software company weights application security in ways that don’t appear in a manufacturer’s risk model. A manufacturer adds weight to network segmentation and OT monitoring.

Factor 3

Velocity factor

Rate of change amplifies effective weight. A KRI trending sharply in the wrong direction deserves more attention than a stable amber metric. Movement is more informative than point-in-time state, deterioration over 30 days is a different conversation than the same level for a quarter.

Domain weight defaults by vertical

Starting weights on a 1–5 scale. Calibrate against your threat model, regulatory exposure, and what controls you actually have producing signal.

Domain CIS Controls SaaS / Fintech Healthcare Manufacturing
Asset & Software InventoryCIS 1, 2445
Data ProtectionCIS 3553
Secure ConfigurationCIS 4445
Identity & Access ManagementCIS 5, 6554
Vulnerability ManagementCIS 7455
Audit Log ManagementCIS 8453
Email & Web ProtectionsCIS 9443
Malware Defenses & EndpointCIS 10344
Data RecoveryCIS 11454
Network & MonitoringCIS 12, 13335
Human Risk & AwarenessCIS 14233
Third-Party & Supply ChainCIS 15444
Application Software SecurityCIS 16542
Incident ResponseCIS 17454
Penetration TestingCIS 18333
Cloud Security PostureCIS 4, 6542

From KRIs to a risk posture score

Once you have KRI values and weights, the aggregation model is straightforward: score each KRI green/amber/red (1/2/3), multiply by domain weight, sum the results, and normalize to a 0–100 scale. Set thresholds at the organizational level that reflect your risk appetite, not industry benchmarks, which reflect averages across organizations with very different profiles.

The score itself matters less than the direction. A risk posture improving from 62 to 71 over a quarter is the signal your board needs. A posture holding at 85 with three critical KRIs spiking red is more urgent than the aggregate suggests, which is why domain-level visibility matters as much as the composite.

Don’t average your way out of a red KRI. A single critical-weight domain in red posture, unpatched KEV vulnerabilities, zero backup verification, no MFA on remote access, is a material risk regardless of what the aggregate number says. Composite scores should surface critical outliers, not hide them.

Appendix: CIS Controls v8.1 coverage map

Every KRI domain in this library maps to one or more CIS Controls v8.1. Use this table to verify framework coverage at a glance or to trace a specific control back to the domain where it’s measured.

CIS Control Description KRI Domain
CIS 1Inventory and Control of Enterprise AssetsAsset & Software Inventory; Malware Defenses & Endpoint
CIS 2Inventory and Control of Software AssetsAsset & Software Inventory
CIS 3Data ProtectionData Protection
CIS 4Secure Configuration of Enterprise Assets and SoftwareSecure Configuration Management; Cloud Security Posture
CIS 5Account ManagementIdentity & Access Management
CIS 6Access Control ManagementIdentity & Access Management; Cloud Security Posture
CIS 7Continuous Vulnerability ManagementVulnerability Management
CIS 8Audit Log ManagementAudit Log Management
CIS 9Email and Web Browser ProtectionsEmail & Web Protections; Network & Monitoring (DNS)
CIS 10Malware DefensesMalware Defenses & Endpoint
CIS 11Data RecoveryData Recovery
CIS 12Network Infrastructure ManagementNetwork Infrastructure & Monitoring
CIS 13Network Monitoring and DefenseNetwork Infrastructure & Monitoring
CIS 14Security Awareness and Skills TrainingHuman Risk & Security Awareness
CIS 15Service Provider ManagementThird-Party & Supply Chain Risk
CIS 16Application Software SecurityApplication Software Security
CIS 17Incident Response ManagementIncident Response Management
CIS 18Penetration TestingPenetration Testing

Stop measuring by hand.

Draxis reads your existing security controls, extracts KRI values programmatically, and surfaces them as financial and operational risk signals, so you can show your board what your security stack is actually telling you about exposure, not just that it’s running.

Don't wait for the breach to read the signal →