Cyber Security Ratings: How They Work
How cyber security ratings are calculated, what they can and cannot tell you about a vendor, how to dispute one, and how to use ratings in third-party risk.
Cyber security ratings are scores that estimate an organization's security posture from the outside. A rating provider scans what any attacker could see on the internet, such as exposed services, known vulnerabilities, leaked credentials and email configuration, and turns those observations into a single number or grade. No access to the rated company is needed, which is why ratings are popular for judging suppliers at scale.
That same property is their main limit. A rating sees the front door, not the house. This guide explains how ratings are built, what they miss, how to challenge one, and how to use them properly inside a third-party risk program alongside TPRM software and questionnaires.
How are cyber security ratings calculated?
Every provider uses its own model, but the inputs fall into a few families.
| Input family | What is observed | What it suggests | Blind spot |
|---|---|---|---|
| Exposed assets and open ports | Internet-facing hosts, services, admin interfaces | Size of the attack surface, possible misconfiguration | Assets wrongly attributed to the company, or missed entirely |
| Vulnerabilities | Software versions with known CVEs on exposed services | Patch discipline | Version banners can mislead; internal systems are invisible |
| Credential leaks | Employee credentials in breach dumps and dark web sources | Phishing and account takeover risk | Old leaks may already be rotated |
| Email and DNS posture | SPF, DKIM and DMARC records, DNS hygiene | Exposure to spoofing and phishing | Says little about anything else |
| Technology and patching signals | Outdated frameworks, end-of-life software, certificates | General hygiene | Correlation, not proof of compromise |
The provider weights these families, normalizes the result against other companies, and publishes a score that is refreshed as new scans arrive. Two providers looking at the same company can reach different scores, because their asset discovery, weights and scales differ.
What can a security rating tell you, and what does it miss?
What it does well:
- Gives a consistent first view of hundreds of vendors without sending a single questionnaire
- Detects change: a new exposed database or a fresh credential leak shows up between annual reviews
- Provides evidence for a conversation with the vendor, because the findings are concrete
What it misses:
- Internal controls. Access management, backups, logging and incident response are invisible from outside.
- What the vendor means to you. A rating is about the vendor; criticality is about your dependence on it. Our guide to vendor risk rating separates the two.
- Attribution accuracy. If the provider maps the wrong IP ranges to a company, the score describes someone else's infrastructure.
- Context. A test server on the internet may be intentional and isolated; a scanner cannot tell.
Treat a rating as a smoke detector, not a building inspection. It tells you where to look, not what you will find.
What makes a security rating fair?
The industry has a reference point for this. The U.S. Chamber of Commerce publishes the Principles for Fair and Accurate Security Ratings, supported by a group of companies including rating providers. The principles are:
- Transparency: providers disclose their methodology and data types.
- Dispute, correction and appeal: rated organizations can challenge a rating and submit corrections, and disputed ratings are flagged until resolved.
- Accuracy and validation: ratings are data-driven or labeled as expert opinion, and methods are validated.
- Model governance: customers get reasonable notice before methods or data sets change.
- Independence: commercial relationships do not affect a rating.
- Confidentiality: information shared during disputes is protected, and data that could enable a compromise is not passed to third parties.
Use these as questions when you choose a provider. A rating you cannot explain to a vendor is a rating you cannot act on.
How should you use ratings in third-party risk?
Ratings work best as one stage in a tiered process, not as the process itself:
- Tier vendors by criticality first. Ratings do not change a vendor's tier.
- Take a baseline at onboarding and record it with the due diligence file.
- Monitor continuously for the top tiers and periodically for the rest.
- Set thresholds that trigger a reassessment, such as a significant score drop or a critical finding, not an automatic termination.
- Share findings with the vendor and ask for a remediation date.
- Combine with inside-out evidence. A questionnaire such as the SIG questionnaire covers the controls a scan cannot see.
The combination matters for compliance. The NIST Cybersecurity Framework 2.0 subcategory GV.SC-07 expects supplier risks to be assessed, responded to and monitored over the course of the relationship. Continuous ratings are good evidence for "monitored"; they are weak evidence for "assessed".
Ratings, questionnaires or penetration test reports?
Each method answers a different question, which is why mature programs use all three in proportion to the tier.
| Method | View | Best for | Limit |
|---|---|---|---|
| Security rating | Outside-in, continuous | Watching many vendors for change | Cannot see internal controls |
| Questionnaire | Inside-out, self-reported | Policies and controls at a point in time | Depends on honest, informed answers |
| Penetration test report | Expert-tested, external or internal | Showing what is actually exploitable at a critical vendor | Point in time, and scoped by the vendor |
A sensible default: ratings for every vendor, questionnaires for the top three tiers, and a recent independent test report for critical vendors. When the three disagree, for example a clean questionnaire next to a falling rating, that disagreement is the most useful finding you have.
What do NIS2 and DORA say about ratings?
Neither text mentions security ratings. Both ask for things ratings can support.
NIS2 Article 21(2)(d) requires supply chain security, and Article 21(3) asks entities to take into account the vulnerabilities specific to each direct supplier and the overall quality of their cybersecurity practices. An outside-in rating is one input to that judgment.
DORA Article 28(4) requires financial entities, before contracting, to identify and assess all relevant risks and to undertake due diligence on prospective ICT third-party service providers. A rating can open that due diligence; it cannot complete it.
How do you dispute or improve your own rating?
If you are the vendor being rated, start with attribution. Ask the provider which assets it attributes to you and remove anything that is not yours. Then work through what is genuinely yours:
- Close or protect exposed administrative interfaces and unused services
- Patch internet-facing software with known vulnerabilities first
- Rotate credentials found in leaks and enforce multi-factor authentication
- Publish correct SPF, DKIM and DMARC records
- Ask for a rescan, and keep the before-and-after evidence for customer questionnaires
Continuous attack surface management on your own estate catches the same issues before a rating provider, or a customer, does.
How does ResilientX rate vendors?
The ResilientX security rating is a continuous outside-in score built from exposed assets and open ports, vulnerabilities, credential leaks, patch status and technologies, and email and phishing posture, with the weight of each factor shown. It shows how scores are calculated and what drives risk, and weights and thresholds can be adjusted by vendor tier.
When a score crosses a threshold, the platform can open remediation requests, trigger reassessments and notify owners, with an audit trail. Automate third-party risk with ResilientX TPRM to put ratings, tiers and questionnaires in one workflow.
Frequently asked questions
Are cyber security ratings accurate?
They are accurate about what they can see, and silent about the rest. The biggest error source is attribution: assets assigned to the wrong company. The second is context, such as a deliberately exposed test system. Ask your provider how it discovers and attributes assets, and check a few vendors you know well against their scores before relying on the method at scale.
Can a vendor have a good rating and still be risky?
Yes. A clean external surface says nothing about weak access control, missing backups or an untested incident plan. A vendor that hosts your core platform is critical whatever its score. That is why ratings sit after criticality tiering and next to questionnaire evidence, never in place of either.
How often do security ratings change?
As often as the underlying scans. Exposed services, new vulnerabilities and fresh credential leaks can move a score within days, while email configuration rarely changes. For critical vendors, alert on significant changes instead of reviewing the number on a calendar, and treat each alert as a prompt to ask the vendor what happened.
The short version
Cyber security ratings are a fast, cheap outside view of vendor security. Use them to monitor and to start conversations, tier vendors by criticality before you look at them, and pair them with inside-out evidence before you make decisions.
Automate third-party risk with ResilientX TPRM — book a demo and see your supplier list scored, tiered and monitored in one place.