Organisations adopting the CIS Controls rarely implement every Safeguard everywhere, and they need a defensible way to show that what they chose is reasonable. The CIS Risk Assessment Method (CIS RAM) is the free method the Center for Internet Security (CIS) publishes for exactly that decision: it helps an organisation set its own acceptable level of risk, model foreseeable threats against its Safeguards, and record why a gap is acceptable or needs treatment. This guide explains how the method works, which edition to use, how to run a first assessment and how to keep it alive afterwards. It is written for security leads, risk owners and assessors who already know the CIS Controls and want to use CIS RAM in practice. For the Controls themselves, Implementation Groups and the CIS tool family, see the CIS Controls Guide.
This is an independent guide from NHI Mgmt Group, based on CIS's public documentation as of November 2026. It is not sponsored or reviewed by CIS. The method changes between versions, so confirm details against CIS's current documents.
Key takeaways
- CIS RAM answers "is this gap acceptable?" It turns a list of unimplemented or partly implemented Safeguards into documented risk decisions, with residual risk you can defend to a regulator, auditor or court.
- v2.2 is current and follows Controls v8.1. CIS published CIS RAM v2.2 as a document family: a Core document, an edition with a companion workbook for each Implementation Group, and a document on using the VERIS Community Database.
- Acceptable means appropriate and reasonable. A risk is acceptable only if its likely impact is acceptable to every party that could foreseeably be harmed, and if the safeguard that would reduce it would not create a burden greater than the risk it addresses.
- Expectancy replaced likelihood in v2.1. The method estimates the most likely way an incident would happen, from how commonly threats are reported and how mature the relevant Safeguards are, rather than a probability over a time window.
- Criteria come before scoring. The work that makes a CIS RAM assessment defensible is agreeing impact and risk acceptance criteria with business leaders and legal counsel before any Safeguard is scored.
What CIS RAM is and where it came from
CIS describes CIS RAM as an information security risk assessment method that helps organisations implement and assess their security posture against the CIS Critical Security Controls. HALOCK Security Labs developed it with CIS. Version 1.0 appeared in 2018, version 2.0 in October 2021 alongside Controls v8, version 2.1 in 2022, and version 2.2 for Controls v8.1 in late 2025. Earlier editions for Controls v8 and v7.1 remain available for organisations still mapped to those versions.
Two ideas set CIS RAM apart from a generic heat-map exercise:
- Reasonableness as the goal: CIS frames the method around the balance between the harm an incident could cause and the burden of the safeguards that would prevent it. CIS links this to the "reasonable" security that regulators expect and the "due care" language judges use when deciding whether an organisation took adequate precautions.
- Built on DoCRA: CIS RAM implements the Duty of Care Risk Analysis (DoCRA) standard, which has three principles and ten practices and is maintained by the non-profit DoCRA Council, of which CIS is a founding member. DoCRA presents risk evaluation in terms familiar to legal authorities, regulators and security professionals at the same time.
CIS states that the original method conforms to established risk assessment standards including ISO/IEC 27005, NIST SP 800-30, OCTAVE and RISK IT, so a CIS RAM assessment can sit inside a wider ISO or NIST risk programme rather than replacing it. If you need the federal method in detail, see the NIST SP 800-30 Risk Assessment Guide.
The v2.2 document family and which edition to use
CIS RAM is not one document. CIS's landing page lists the v2.2 family below; the files are free but are downloaded through a short CIS registration form.
| Document | What it is for | Who should start here |
|---|---|---|
| CIS RAM Core | The "bare essentials": the method's principles and practices without Implementation Group scaffolding | Risk teams building their own tooling, or reconciling CIS RAM with an existing ISO or NIST method |
| CIS RAM for IG1 + companion workbook | A largely automated assessment against the IG1 Safeguards, designed for enterprises with little cyber security expertise | Small organisations, first assessments, IG1 programmes |
| CIS RAM for IG2 + companion workbook | Combines qualitative and quantitative analysis and evaluates risk per asset class; also addresses enterprises whose departments have differing risk profiles | Mid-size enterprises with staff who configure systems, applications and networks |
| CIS RAM for IG3 + companion workbook | Adds threat modelling against the five attack types in the CIS Community Defense Model: ransomware, malware, web application hacking, insider and privilege misuse, and targeted intrusions | Enterprises with security expertise facing sophisticated attackers |
| CIS RAM and the VERIS Community Database | Explains how reported incident data informs the expectancy estimates | Anyone who will defend expectancy scores to an auditor |
CIS has described the same split by experience since version 1.0: beginners model foreseeable threats against the Controls as they apply them; experienced practitioners model threats against information assets to decide how the Controls should be configured; risk experts analyse attack paths based on CIS's community attack model. Pick the edition that matches your Implementation Group target and your team's skill, not the one with the most columns. An IG1 organisation using the IG3 workbook usually produces scores nobody can defend.
Decision guidance: if you are already IG1 compliant and assessing against IG2 Safeguards, use the IG2 edition for the whole scope rather than mixing workbooks. If you run several business units with different risk appetites, the IG2 edition's per-department view saves you from running separate assessments. If your board already receives quantified risk in financial terms, keep that method for enterprise reporting and use CIS RAM to justify Safeguard choices beneath it; see the FAIR Model and Open FAIR Standard Guide for that layer.
The core process, step by step
CIS's v2.0 announcement sets out five steps that run through every edition. The workbooks supply the templates; the sequence below adds the practical work around each step.
- Develop criteria. Define the Risk Assessment Criteria (how you will express impact and expectancy) and the Risk Acceptance Criteria (where you draw the line). CIS stresses criteria that all parties can understand, developed with business leaders and legal counsel.
- Model risks. For each Safeguard or asset class in scope, evaluate the current implementation that would prevent or detect foreseeable threats. Record evidence, not opinion: configuration exports, coverage reports, tickets.
- Evaluate risks. Estimate expectancy and impact to reach a risk score, then compare it with the acceptance line. Risks below the line meet due care; risks above it require treatment.
- Recommend Safeguards. Propose Safeguards, or better implementation of existing ones, to bring unacceptable risks below the line.
- Evaluate recommendations. Analyse each proposed Safeguard as a risk in its own right: it must reduce risk to an acceptable level without creating an undue burden on the business it is meant to protect. If it fails, model an alternative, recalculate and reassess.
Step 5 is what most organisations skip and what makes the result defensible. A control that halts a critical business process may create more harm than the threat it addresses, and CIS RAM requires you to write that comparison down.
Criteria: impact, expectancy and the acceptance line
CIS's public material describes the criteria at the level below; the exact scales, wording templates and worked examples are in the edition workbooks.
| Criterion | What it expresses | Practical notes |
|---|---|---|
| Impact criteria | The magnitude of harm to the organisation's mission and objectives and to external parties: customers, regulators, vendors and the public | Write each level in plain business terms (for example, how many customers affected, which obligations breached). The v2.1 and v2.2 editions give a simplified format for stating impact magnitudes. |
| Expectancy criteria | An estimate that, if an incident occurred, it would come from the threat being analysed | Fixed definitions in the IG editions; calculated automatically from the commonality of reported threats and the maturity of the relevant Safeguard. Not a probability over a time period. |
| Risk acceptance criteria | The line between acceptable risk and risk that needs treatment | Agree it with executives and counsel before scoring. It should reflect harm to others as well as to the organisation. |
| Safeguard risk | The burden a proposed Safeguard places on the business and on the parties it serves | Express it in the same terms as your impact criteria, so harm from a control and harm from a threat can be compared directly. |
CIS states two conditions for an acceptable risk. Appropriate risk: the likelihood of an impact must be acceptable to all foreseeably affected parties. Reasonable risk: the risk posed by a safeguard must be less than or equal to the risk it protects against. A gap that harms only your own convenience may be appropriate; a gap that exposes customer data rarely is, however cheap it is to leave alone.
For setting the acceptance line in board language, see the Risk Appetite and Tolerance for Cyber Guide and the Cyber Risk Appetite Statement Template.
Rolling out a first assessment
- Scope and sponsor (week 1). Name a risk owner with authority to accept residual risk. Fix the scope: Implementation Group target, business units and asset classes. Download the matching edition and workbook.
- Criteria workshop (weeks 1 to 2). Run a half-day session with business leaders, legal counsel and the privacy lead to agree business objectives, obligations to others, impact levels and the acceptance line. Minute the decisions: this record is the core of your due care evidence.
- Evidence collection (weeks 2 to 4). For each in-scope Safeguard, gather implementation evidence and a maturity rating. Reuse what you already hold from CIS CSAT or an internal audit rather than re-surveying teams.
- Pilot one business unit (week 4). Score one unit end to end, then review the results with its owner. If scores feel wrong, fix the criteria wording now, before scaling.
- Full scoring and treatment (weeks 5 to 8). Score the remaining scope, propose Safeguards for every risk above the line, and run step 5 on each proposal.
- Sign-off. The risk owner signs acceptance for residual risks and approves the treatment plan. Feed treatments into the Risk Treatment Plan Template and risks into your Risk Register Template.
- Rollback and correction. If a later incident or audit shows a criterion was wrong, change the criterion through the same governance route and re-score affected risks; never adjust individual scores to reach a preferred answer.
Small versus large organisations: a small organisation can complete an IG1 assessment with a part-time owner over a few weeks, since that edition automates much of the calculation. A large enterprise should run CIS RAM per business unit with a shared criteria set, and should plan for the workbooks to become a register that lives in a GRC tool once the first cycle is complete. Combining a top-down scenario view with this bottom-up Safeguard view is covered in the Top-Down vs Bottom-Up Cyber Risk Assessment.
Where identity fits
Account and access Safeguards (Controls 5 and 6) are where many CIS RAM assessments find their highest-impact gaps, and where the reasonableness test bites hardest. Service accounts, API keys and automation tokens often sit outside the account inventory that Safeguard 5.1 expects, so their maturity scores are overstated if teams only count human users. When modelling Controls 5 and 6, ask explicitly about non-human identities and include them in the evidence. See the CIS Controls IG1 Rollout Plan for the hygiene baseline.
Measuring whether it is working
- Risks above the line: count and trend per assessment cycle, by Control. The number should fall, or the explanation for a rise should be a new threat or a stricter line.
- Treatment on time: percentage of treatment actions closed by their agreed date.
- Acceptance coverage: percentage of residual risks above zero that carry a signed acceptance with an expiry date.
- Criteria age: months since impact and acceptance criteria were last reviewed by executives and counsel; aim for at least annual review.
- Evidence freshness: percentage of Safeguard maturity ratings backed by evidence less than 12 months old.
Framework mapping
| Framework | Where CIS RAM output fits |
|---|---|
| NIST CSF 2.0 | GV.RM-02 (risk appetite and tolerance statements), ID.RA-04 (impacts and likelihoods recorded), ID.RA-05 (risk used to prioritise response), ID.RA-06 (risk responses chosen and tracked) |
| NIST SP 800-53 Rev. 5 | RA-3 Risk Assessment, RA-7 Risk Response, PM-9 Risk Management Strategy, CA-5 Plan of Action and Milestones |
| ISO/IEC 27001:2022 | Clause 6.1.2 information security risk assessment and clause 6.1.3 risk treatment; CIS RAM's acceptance criteria correspond to the risk acceptance criteria the standard requires |
| CIS Controls v8.1 | Every Safeguard in the chosen Implementation Group; the assessment justifies partial or deferred implementation |
Common failure patterns
- Scoring before criteria are agreed: teams fill the workbook first and back-fit criteria, which leaves no evidence that the acceptance line was set independently of the answer.
- Ignoring harm to others: impact criteria that only count cost to the organisation fail the "appropriate risk" test, which looks at every foreseeably affected party.
- Skipping safeguard evaluation: recommending controls without scoring their burden removes the half of the method that makes it about reasonableness.
- Treating expectancy as probability: reporting CIS RAM expectancy scores to the board as annual likelihoods misstates what the method measures.
- Edition mismatch: using an IG3 workbook in an organisation without the skills to feed it produces confident numbers built on guesses.
- One-off assessment: a register that is not re-scored after major changes or incidents stops reflecting reality within a year.
Practitioner checklist
- Download the CIS RAM v2.2 edition that matches your Implementation Group target, plus the Core document.
- Name a risk owner with authority to accept residual risk.
- Agree impact, expectancy and risk acceptance criteria with business leaders and legal counsel, and minute it.
- Include customers, regulators and partners in impact criteria, not just internal cost.
- Back every Safeguard maturity rating with evidence, including non-human accounts in Controls 5 and 6.
- Pilot one business unit before scoring the full scope.
- Score the burden of every recommended Safeguard before approving it.
- Record signed acceptances with expiry dates for all residual risks.
- Re-score after major change, incident or criteria update, and at least annually.
Standards and references
- CIS: CIS Risk Assessment Method (CIS RAM)
- CIS: CIS Risk Assessment Method (RAM) v2.0 for CIS Controls v8
- CIS: CIS Risk Assessment Method (RAM) v2.1 for CIS Controls v8
- CIS: CIS RAM v2.1 for Implementation Group 2 (IG2)
- CIS: Defining "Reasonable" Security with a Risk Assessment Method
- CIS: 3 Things You'll Learn Conducting a Cyber Risk Assessment with CIS RAM
- CIS: CIS RAM v1.0 Puts the CIS Controls into Action
- CIS: Reasonable Cybersecurity: From Legal Theory to Practice
Related NHI Mgmt Group resources: CIS Controls Guide · Top-Down vs Bottom-Up Cyber Risk Assessment · Risk Register Template · Accept vs Transfer vs Mitigate vs Avoid