The risk assessment should explain why your controls are appropriate
A SOC 2 risk assessment should identify what could prevent the scoped service from meeting its relevant commitments, evaluate the significance of those scenarios, and guide decisions about controls and remaining exposure. A populated risk register is useful only when management can explain how its contents relate to the real service and what decisions followed.
Start with the service outcome rather than a catalogue of familiar threats. Losing access to an internal convenience tool and losing access to the platform customers use for urgent processing may involve similar technology but very different consequences. The assessment needs enough context to distinguish them and direct attention accordingly.
The AICPA Trust Services Criteria provide criteria for evaluating controls over relevant information and systems. The method below helps management connect service risks with control decisions and evidence. Adapt the scoring and review process to the decisions your organisation needs to make. Confirm detailed criteria alignment and reporting implications with the independent CPA. SOC 2 results in an attestation report, not a blanket certification that every organisational risk has been eliminated.
Establish the decisions the assessment must support
Define the system boundary, customer commitments, relevant requirements and decision owners before scheduling workshops. Identify which products, operating units, infrastructure and dependencies are included. Where the boundary remains uncertain, record the uncertainty and resolve it with the scope owner. Assessing an undefined service tends to produce an attractive but generic register.
Collect information that can ground the discussion: architecture, data flows, operating incidents, supplier dependencies, customer complaints, change plans and existing control evidence. Include service staff who know where workarounds occur. Management presentations often describe the intended process, while operators can explain what happens when demand spikes or a dependency fails.
Agree how the assessment will be used. It may inform control design, resource allocation, exception acceptance, supplier decisions or readiness priorities. The output should be appropriate to those decisions. Excessively granular scoring can consume time without improving a choice; excessively broad categories can hide a risk that needs immediate action.
Record assumptions such as expected transaction volume, support coverage, hosting design and customer responsibilities. An assessment can be reasonable at one stage of growth and misleading after the service changes. Explicit assumptions make reassessment more reliable because teams can see which conclusions depend on facts that no longer hold.
Use our SOC 2 scope guide to identify the service, commitments and responsibility boundaries that the risk assessment needs to address.
Write risk scenarios that describe a credible service failure
A useful scenario links a cause or event, an exposed part of the service and a consequence. For example, a stale integration credential could allow an unauthorised party to retrieve customer documents, leading to a confidentiality failure. This is more actionable than recording cyberattack with a high rating and no explanation of the pathway.
Include non-malicious failures. Incorrect configuration, unclear responsibilities, insufficient capacity, operator error, defective processing logic and incomplete supplier handoffs can all undermine service commitments. A register containing only attackers and malware may miss the everyday operating failures most likely to affect customers.
Explore dependencies and combinations. An unavailable identity provider may prevent both ordinary operations and emergency recovery. A staffing shortage combined with a supplier outage may overwhelm manual fallback. You do not need every imaginable combination, but you should challenge assumptions that several controls will remain available under the same disruptive event.
Use scenario detail in proportion to the decision. Describe affected information, access, service function and likely duration or reach where known. Avoid inventing precision that the organisation cannot support. If evidence is weak, identify the uncertainty and decide whether investigation, testing or a cautious interim measure is needed.
Make impact and likelihood understandable to decision makers
Define impact in terms relevant to the service. Possible dimensions include customer disruption, exposure of confidential information, incorrect processing, contractual consequences and recovery effort. Explain what different ratings mean with practical examples. A score labelled severe is more useful when managers understand the kind of consequence that earns it.
Likelihood should reflect the scenario, operating exposure, available evidence and time horizon. Avoid copying a generic industry rating without considering your environment. A publicly reachable administrative interface, an untested manual process and a frequently failing supplier present different evidence for the likelihood discussion.
Distinguish inherent risk from the effect of current safeguards. State what assumptions are removed or retained when assessing the scenario before controls. Organisations use these terms differently, so consistency and transparency matter. Do not subtract arbitrary numbers from a score merely because a policy exists; consider whether the control operates and addresses the actual pathway.
Use the scoring model to support judgement, not replace it. Two risks with the same numeric total may have very different consequences and treatment urgency. Preserve a short rationale and allow defined escalation for unusually serious scenarios. Reviewers should be able to understand the decision even without reproducing every arithmetic step.
Map safeguards to the scenario they actually address
For each material risk, identify the existing safeguards and explain their role. Some reduce the chance of an event, some detect it, and others limit consequences or support recovery. This distinction helps reveal gaps. A monitoring alert may detect unauthorised activity but does not itself prevent access or restore affected data.
Assess design and operation separately. A well-designed approval process may fail because reviewers do not receive complete populations. A backup process may run successfully but be unable to restore the required service. Ask what evidence shows the control is available, functioning and relevant under the scenario being considered.
Avoid mapping every control to every risk. Broad mapping creates the illusion of coverage and makes failure analysis difficult. Identify the safeguards that materially influence the scenario and any dependencies between them. If several controls rely on the same person, platform or data source, acknowledge the concentration rather than counting them as fully independent protection.
When selecting new measures, compare feasibility, expected benefit, cost and operational side effects. A control that staff cannot perform during ordinary workloads may not remain effective. Include the team that must operate it and identify the evidence it will naturally generate. Control design and evidence design should develop together.
Turn risk decisions into owned actions and acceptance records
Treatment options may include reducing the risk, avoiding the activity, changing the service arrangement or accepting a defined residual exposure. Contractual transfer and insurance may affect some consequences, but do not automatically remove operational responsibility or customer impact. State which part of the risk each decision addresses.
An action should identify the intended outcome, accountable owner, target date, dependencies and completion evidence. Replace vague instructions such as improve security with an observable result. For example, require that a named production access route uses an approved permission model and that the responsible owner verifies the deployed configuration.
Residual-risk acceptance needs an authorised decision maker and a clear scope. Describe the remaining scenario, the basis for acceptance, any temporary safeguards and the conditions for reconsideration. Acceptance should not be an indefinite parking place for actions that have lost sponsorship. Time limits and change triggers can keep the decision connected to reality.
Escalate conflicts explicitly. A product launch may depend on accepting a limitation that exceeds a team’s authority. Security should explain the exposure and alternatives, while management makes the appropriately authorised business decision. The record should preserve that decision without implying that a compliance coordinator can silently approve any risk to meet an audit deadline.
Give review meetings a defined management purpose
Assign a risk owner who can influence the affected service, not simply the person who discovered the issue. Security can coordinate assessment and challenge assumptions. Technology, operations, procurement and other specialists provide evidence. Senior management establishes decision authority, funds treatment and resolves priorities when several teams are involved.
Set a review rhythm appropriate to the environment and relevant commitments. Significant changes should also trigger reassessment: new services, material incidents, acquisitions, major suppliers, deployment regions or altered customer requirements. Do not present one annual workshop as universally sufficient or one fixed cadence as universally mandated by SOC 2.
Use review meetings to decide, not merely read the register aloud. Focus on changed exposure, overdue material actions, exceptions nearing review, uncertain assumptions and controls whose evidence contradicts earlier confidence. Record what changed and why. A short decision log can be more informative than a large slide deck showing the same risk colours every month.
Maintain version history and preserve the evidence behind important conclusions. The organisation should be able to distinguish what management knew at a given time from later information. Updating an assessment after an incident is appropriate; rewriting the old record to imply the risk was always understood is not.
Hypothetical SaaS example: a regional expansion changes the recovery assumption
A fictional scheduling platform plans to serve customers in a second geographic region. Its existing risk register rates an extended infrastructure outage as moderate because management assumes traffic can move to an alternate region. The expansion workshop reveals that the alternate environment has never processed the full production workload and depends on an identity service configured only for the primary region.
The team rewrites the scenario around failure of the primary region while emergency administrators cannot authenticate to the alternate environment. It distinguishes availability impact from the separate confidentiality considerations of moving customer data. Engineering supplies architecture and capacity evidence, while commercial and legal teams clarify customer commitments relevant to the proposed deployment.
Management commissions a targeted recovery exercise and prioritises an independent administrative access route. The exercise demonstrates that a data synchronisation dependency would delay recovery beyond the internal objective. The team revises the treatment plan and the launch decision rather than preserving the old moderate score to avoid inconvenience.
The final register includes the tested limitation, accountable remediation owners and the conditions under which the expansion can proceed. The example shows how risk assessment helps challenge a business assumption before it becomes a customer promise. A long pre-existing control list would not have answered the specific recovery question.
Hypothetical BPO example: peak demand changes the likelihood of processing error
A fictional outsourcing provider processes customer eligibility cases with a second-person quality check for selected categories. During normal volume, supervisors complete the reviews promptly. A new contract doubles the expected peak workload, but the original risk assessment assumes the same staffing and review capacity. The procedure remains unchanged while the operating conditions become materially different.
The process owner models the scenario of analysts rushing complex cases while supervisors accumulate a review backlog. Potential consequences include incorrect decisions, missed service commitments and exposure created by informal spreadsheet workarounds. Operations contributes actual handling times and historical rework data rather than relying solely on optimistic sales forecasts.
Management compares several treatments: staggered intake, trained reserve staff, automated completeness checks and revised escalation for complex cases. It selects a combination and tests it using representative peak-volume work. The exercise identifies a queue that can bypass the second-person review when a supervisor changes shifts, so the handoff design is corrected before the new contract starts.
Residual risk remains because demand can exceed the forecast. The service owner records an escalation threshold and the authority to limit intake or agree revised arrangements with the customer. The assessment therefore informs capacity and workflow decisions, with clear evidence and ownership, rather than existing as a document owned only by the security team.
Prepare an evidence set that shows decisions over time
A useful evidence set includes the assessment method, service context, scenario register, scoring rationales, control references, treatment decisions, acceptance records and review history. It should also identify the people involved and the information they considered. The records need to support management’s account of how risks are evaluated and acted upon.
Keep the register proportionate and readable. Long descriptions can conceal the absence of a decision, while single-word entries can prevent meaningful review. Use concise scenarios and link to supporting analysis where necessary. Restrict sensitive technical detail appropriately, particularly where the record describes exploitable weaknesses or confidential business plans.
Before the examination, select a few risks and trace the full chain from service commitment through scenario, evaluation, control, action and review. Then work backwards from a significant incident or change to see whether the assessment responded. These two directions reveal different weaknesses: unsupported control choices and important events that never entered management’s risk process.
Our SOC 2 evidence guide explains how to connect a decision with its source records, relevant population and operating history.
Remove false confidence from the register
Common weaknesses include generic threat lists, unexplained scores, every risk rated medium, controls credited without evidence and acceptance decisions made by people without authority. Another is closing treatment actions when a product is purchased rather than when the safeguard is operating. Completion should reflect the intended risk outcome, not merely procurement progress.
The strongest challenge is often a simple question: what would make this assessment wrong? Answers may reveal an untested recovery assumption, incomplete inventory, changing customer use or unreliable supplier evidence. Record those uncertainties and decide which deserve investigation. A credible assessment makes the limits of management’s knowledge visible and keeps decisions open to new evidence.
Reference basis
2017 Trust Services Criteria, revised points of focus 2022; AICPA SOC suite of services.
Frequently asked questions
How should we choose a risk-scoring method?
Choose a consistent method that helps management compare service consequences, evaluate control evidence and prioritise treatment. Confirm detailed criteria alignment with the independent CPA using authorised materials.
Who should own the risk register?
A coordinating function can maintain it, but individual risks need owners with influence over the affected service. Management must define who can authorise treatment and accept residual exposure.
Can a risk be accepted instead of remediated?
Management may make an authorised acceptance decision within applicable obligations and commitments. Record the scope, rationale, remaining exposure and conditions for review rather than treating acceptance as automatic closure.
Plan the next decision
Bring your service commitments, current risk register and one major planned change. CyberVantage360 can facilitate a scenario-based assessment that connects control choices, management decisions and evidence for the independent examination.
Discuss risk decisions for your scoped service with CyberVantage360. Our SOC 2 readiness and implementation service can be scoped around credible failure scenarios, supported ratings, treatment ownership and residual-risk decisions for management review.