Required, not optional
HIPAA Security Rule, PCI DSS v4.0, SOC 2, ISO 27001, NIST 800-171, GDPR, NYDFS Part 500, and DORA all mandate a documented risk assessment. Many require annual or post-change reassessment.
A risk assessment is the structured process of identifying what could go wrong, scoring how likely and how damaging each risk is, and deciding what to do about it. The output is a current register of named risks with owners, scores, controls, and treatment plans. Every credible framework (ISO 31000, NIST SP 800-30, ISO 27005, COSO ERM) describes the same five activities: identify, analyse, evaluate, treat, and monitor.
A risk assessment is the structured process of identifying hazards or threats, analysing how likely each one is and how damaging it would be, and deciding what to do about it. The output is a register of named risks with scores, owners, controls, and a treatment plan. Frameworks differ in labels; the underlying activity does not.
The discipline goes back decades in health-and-safety law (OSHA in the US, the HSE in the UK), and was generalised by the International Organization for Standardization in ISO 31000 in 2009 (revised 2018). Today the term applies across domains: cyber, operational, financial, strategic, compliance, physical security. The mechanics are the same; the risks, controls, and stakeholders differ.
The thing that could cause harm. A vulnerability in a system, a chemical in a workplace, a market shift, a vendor failure, an adversary.
How probable the harm is, over a defined time window. Expressed on an ordinal scale (Low/Medium/High) or a probability distribution for quantitative work.
The consequence if it occurs. Dollar loss, downtime, regulatory penalty, harm to people, reputational damage, lost opportunity.
“Risk assessment is the overall process of risk identification, risk analysis and risk evaluation.”
Boards, regulators, insurance carriers, and customers now expect a defensible risk story, not a control checklist. A working risk assessment produces that story.
HIPAA Security Rule, PCI DSS v4.0, SOC 2, ISO 27001, NIST 800-171, GDPR, NYDFS Part 500, and DORA all mandate a documented risk assessment. Many require annual or post-change reassessment.
The 2023 SEC cyber disclosure rules and the EU's Digital Operational Resilience Act require boards to demonstrate that material risks are identified, assessed, and managed. A current register is the answer to both.
Cyber insurance applications now treat the existence of a documented risk assessment as a pre-condition for quoting. Carriers want to see the methodology, the register, and the treatment plan.
ISO 31000 calls them risk identification, analysis, evaluation, treatment, and review. NIST SP 800-30 splits the same activities across prepare, conduct, communicate, and maintain. The labels drift; the work does not.
List the things that could go wrong, against the assets, processes, or objectives that matter. Pull from interviews, audit findings, past incidents, threat intelligence, and the risk register from the last cycle. Each risk gets a one-sentence statement: a threat, a vulnerability, and a consequence.
For each risk, work out how likely it is to occur and what the consequences would be if it did. Use the same scale across the whole register so the numbers are comparable. Most teams start with a 5x5 ordinal matrix; a smaller subset of risks gets a dollar-denominated analysis later.
Compare the inherent score to the organisation's documented risk appetite. Anything above the line needs a treatment decision. Anything below the line gets accepted, with the rationale captured in the register so the next reviewer understands why.
For risks that need treatment, design the controls. Some are preventative (training, access reviews, segmentation), some are detective (monitoring, audits), some are corrective (incident response, backups). Map each control to the risk it reduces, and set a target residual score that the treatment is expected to deliver.
Risks change. Controls drift. New threats appear. A working assessment is reviewed quarterly for the top quartile, annually for the full register, and immediately when material events occur (new system, new vendor, new regulation, post-incident). Key risk indicators (KRIs) carry the watch between reviews.
The fifth step is where most programmes go wrong. Running steps one through four once a year and producing a deck is not a working risk assessment, it is a snapshot that goes stale within a quarter. Programmes that hold up under audit and under board scrutiny run step five continuously: KRIs breach thresholds, control tests fail, incidents land, and the residual scores on the affected risks update without waiting for the next cycle.
The 5x5 risk matrix is the workhorse of step two. You rate likelihood from 1 to 5, rate impact from 1 to 5, and multiply the two for a risk score from 1 to 25. The colour bands turn that number into a decision: green is accept, amber is treat where it pays, orange is escalate, red is act now.
| Likelihood ↓ / Impact → | 1Insignificant | 2Minor | 3Moderate | 4Major | 5Severe |
|---|---|---|---|---|---|
| 5Almost certain | 5Medium | 10High | 15Extreme | 20Extreme | 25Extreme |
| 4Likely | 4Low | 8Medium | 12High | 16Extreme | 20Extreme |
| 3Possible | 3Low | 6Medium | 9Medium | 12High | 15Extreme |
| 2Unlikely | 2Low | 4Low | 6Medium | 8Medium | 10High |
| 1Rare | 1Low | 2Low | 3Low | 4Low | 5Medium |
Accept and monitor. Manage by routine procedures.
Treat where cost-effective. Assign an owner and a review date.
Treatment required. Escalate to management; track to a target.
Immediate action. Senior leadership attention; stop or remediate.
The matrix is fast and easy to read, which is exactly why boards like it. The catch is calibration: "likely" and "major" have to be defined in writing so two assessors score the same risk the same way. Document the scale once, and a 3x3 matrix works for a small register while the 5x5 gives you the resolution a larger programme needs.
Want the grid ready to use? Read how the risk assessment matrix works in depth, or download the free risk matrix template.
The honest answer is that mature programmes run both. Qualitative for breadth across the whole register; quantitative for depth on the risks the board will ask about by name.
Hard to defend with precision. 'High' means different things to different people unless the criteria are tightly documented and assessors are calibrated.
Slow and data-hungry. Easy to anchor on bad assumptions; the maths hide the inputs. Overkill for tail risks where qualitative will do.
Score the whole register qualitatively on a 5x5. Identify the top 10 to 20 risks the board cares about. Run FAIR on those. Reconcile both views on the dashboard so the heat map shows the qualitative picture and the loss-exceedance curve shows the quantitative one. Two views, one register, no double-counting.
Every risk carries two scores, and the gap between them is the point of the whole exercise. Score the matrix twice: once before controls, once after.
The exposure before any controls are credited: the raw threat against the asset, scored on likelihood and impact as if nothing were in place to reduce it. It tells you how big the problem is at its source.
The exposure that remains after the controls currently in place are credited. It is the number that decides whether a risk sits inside your appetite or needs more treatment. Residual, not inherent, is what you manage day to day.
The gap between inherent and residual is the work your existing controls are doing. The gap between residual and your target residual is the work the planned treatment still owes. Boards want all three numbers on one line: where we were, where we are, and where we are heading. Read the full inherent vs residual risk guide.
The two terms get used as if they mean the same thing. They do not. Risk analysis is one step inside the broader risk assessment.
ISO 31000 is explicit about the nesting. Risk assessment is the overall process, and it contains three activities in order: risk identification (finding what could go wrong), risk analysis (working out the level of each risk by scoring likelihood and impact), and risk evaluation (comparing that level against your appetite to decide what to act on).
So risk analysis is the middle stage, the part where you actually put numbers on a risk. Risk assessment is the full identify-analyse-evaluate cycle that wraps around it and produces the register and the treatment decisions. Every risk analysis happens inside a risk assessment; not every part of a risk assessment is risk analysis.
You do not have to pick one and live with it forever. Most mature programmes run a stack: an enterprise wrapper (ISO 31000 or COSO ERM), a methodology (ISO 27005, NIST 800-30, or OCTAVE), and a quantitative engine (FAIR) for the top risks.
The international standard for risk management principles, framework, and process. Sits above any specific domain (cyber, operational, financial) and gives the common vocabulary boards and regulators recognise.
The US National Institute of Standards and Technology guide for conducting information system risk assessments. Pairs with NIST 800-39 (organisational) and NIST 800-37 (the RMF). Default for US federal contractors.
The risk methodology that lives inside ISO 27001. Tells you how to identify, analyse, evaluate, and treat information security risks against the Annex A control set. Required reading if you are pursuing or maintaining ISO 27001 certification.
Operationally Critical Threat, Asset, and Vulnerability Evaluation. Developed at Carnegie Mellon SEI, focuses on the information assets themselves rather than on systems. Compact and self-directed, popular with mid-size teams.
Factor Analysis of Information Risk. The de facto standard for expressing cyber risk in dollar terms. Pairs with a Monte Carlo simulation to produce loss-exceedance curves the CFO and board will recognise.
The Committee of Sponsoring Organizations framework. Widely used by US public companies and is the reference cited by SOX internal controls work. Twenty principles organised around governance, strategy, performance, review, and information.
Six categories cover the vast majority of enterprise risks worth tracking. Use them as a starter taxonomy, then split or merge as the programme matures.
Risks from people, processes, and systems. Process failures, key-person dependencies, change-management gaps, capacity issues, vendor outages. The category that touches the most day-to-day work.
Threats to confidentiality, integrity, and availability of information systems. Ransomware, phishing, exploitation, insider misuse, cloud misconfiguration. The fastest-growing category on most registers since 2020.
Liquidity, credit, market, currency, and counterparty exposures. Owned by the finance and treasury functions, but the controls (segregation of duties, reconciliations, approvals) overlap with operational and compliance work.
Risks to the organisation's ability to meet its objectives. New competitors, technology shifts, customer-base changes, M&A integration. Long-horizon, hard to quantify, but the category boards ask about first.
Risks of non-compliance with laws, standards, and contractual obligations. HIPAA, PCI DSS, GDPR, SOC 2, SOX, sector-specific rules. Often the trigger for the first formal risk assessment a programme runs.
Site security, workplace safety, fire, flood, severe weather, infrastructure failure. Usually owned by facilities or EHS teams, but the residual score feeds the enterprise register alongside cyber and operational.
No category is independent of the others. A ransomware event is a cyber risk that lands as an operational outage, a financial loss, a compliance breach (HIPAA, GDPR), and a strategic reputational hit. The register should record the primary category, but the controls and the impact analysis touch the rest.
The three terms get used interchangeably. They are not the same thing. Knowing the difference is the first sign of a mature programme.
| Term | What it is | Frequency |
|---|---|---|
| Risk assessment | A bounded activity that identifies, analyses, and evaluates risks against a defined scope (an asset, a system, a process, the enterprise). Produces a register and a treatment plan. | Periodic (typically annual, with quarterly updates and event-triggered reassessments) |
| Risk management | The ongoing discipline that includes governance, appetite-setting, assessments, treatment execution, monitoring, incident response, and reporting. The programme that contains the assessments. | Continuous (every working day) |
| Risk register | The artefact. A structured record of every risk identified, with scores, owners, controls, treatment decisions, target scores, review dates, and an audit trail. The source of truth that every dashboard and board pack reconciles to. | Live (updated as evidence arrives) |
The simplest way to remember it: the assessment is what you do, the register is what you keep, and risk management is the programme that runs both. A programme that does an assessment without maintaining the register has a report. A programme with a register but no scheduled reassessments has a museum.
A spreadsheet works for a first assessment. It breaks the moment a programme adds a second framework, a third assessor, or a fourth business unit. Below are the free templates we publish, plus what the platform adds when the spreadsheet gives out.
Use these as starting points. Each one maps to a recognised framework and is structured so the work transfers cleanly into any GRC platform later.
A platform centralises the register, the control library, and the treatment workflow. Assessments run against any framework; controls cross-map so one piece of evidence pays off across HIPAA, ISO 27001, SOC 2, PCI DSS, and NIST.
The frameworks referenced on this page are published by international standards bodies, US federal agencies, and academic research centres. Direct links below.
Everything on this page, ready to use. Each download maps to a recognised framework and is structured so the work transfers cleanly into a GRC platform when the spreadsheet gives out.
The 5x5 matrix as a ready-to-use grid with the bands and scoring built in.
A structured worksheet for the full identify, analyse, evaluate, treat flow.
The living artefact: risks, scores, owners, controls, and review dates.
How the matrix works, when to use 3x3 versus 5x5, and how to calibrate it.
Apply the same method to third parties before and after onboarding.
The questions that come up on the way to a working programme, with practitioner answers.
The Global Register, Risk Templates for ISO 31000 / ISO 27005 / NIST 800-30 / FAIR, KRI library, and Risk-vs-Compliance mapping, all on one platform. 30-day free trial, no credit card.
No credit card required · 30-day free trial · Cancel anytime