Noru

SOC 2 Policies: The List Most Reports Rely On, and What Each Must Say

The policies most SOC 2 reports rely on, the Trust Services Criteria each one supports, what it has to say, and how approval and acknowledgement are tested.

Bip Thelin2 min readPublished

The AICPA does not publish a list of required SOC 2 policies. The Trust Services Criteria expect management to set out its expectations in policies and procedures and to put them into practice, and in practice most reports rely on the same core set: information security, access control, change management, incident response, risk assessment, vendor management, business continuity and disaster recovery, data classification and retention, acceptable use, human resources security, logging and monitoring, vulnerability management and secure development. Each one needs an owner, management approval, a review date and, where it applies to staff, an acknowledgement, because the auditor tests those as well as the content.

Policies are the part of SOC 2 that looks easiest and causes surprising amounts of rework. A downloaded template can be approved in an afternoon, and then fieldwork finds that it promises a quarterly review nobody runs or a tool the company does not use. The auditor tests what the policy says, so the job is to write down what you will actually do.

The core set and the criteria they support

PolicySupports (2017 criteria)What it has to cover
Information securityCC1, CC2, CC5Scope, roles and responsibilities, management oversight, how the other policies fit together
Access controlCC6.1–CC6.3Provisioning and removal, least privilege, authentication and MFA, periodic access reviews
Change managementCC8.1How changes are proposed, reviewed, tested, approved and deployed, including emergency changes
Secure developmentCC8.1Code review, testing, dependency and secret handling, separation between writing and deploying
Incident responseCC7.3–CC7.5Detection, classification, escalation, customer communication, recovery and post-incident review
Logging and monitoringCC7.2What is logged, how long logs are kept, which events alert and who responds
Vulnerability managementCC7.1Scanning, penetration testing, severity ratings and remediation timelines
Risk assessmentCC3How risks are identified, rated, treated and reviewed, and how often
Vendor managementCC9.2How suppliers are assessed before use and monitored after, including their own reports
Business continuity and disaster recoveryCC9.1, A1.2–A1.3Backups, restore testing, recovery objectives and the recovery plan and its test
Data classification and retentionC1.1–C1.2Data categories, handling rules per category, retention periods and disposal
Acceptable useCC1.1, CC2.2Expected behaviour with company systems, devices and data
Human resources securityCC1.1, CC1.4Background checks, onboarding, training, disciplinary process and offboarding

Availability, Confidentiality, Processing Integrity and Privacy policies are only needed when those categories are in scope, and some companies fold several of the policies above into one document. Either is fine, as long as each criterion is covered somewhere and a control puts it into practice.

What the auditor tests about policies

  • That the policy exists and covers the criteria the description says it does.
  • That management approved it, with a date, and reviews it on the schedule it states.
  • That staff it applies to acknowledged it, usually at onboarding and again after significant changes or annually.
  • That the controls described in the report match what the policy requires, such as the access review frequency.

Writing policies that survive fieldwork

  • Describe what you do, not what a larger company would do. A monthly review you skip is an exception; a quarterly one you run is not.
  • Name roles, not people, so the policy survives someone leaving.
  • Keep procedures, such as step-by-step runbooks, out of the policy, so a small process change does not need a new approval.
  • State frequencies and timelines you can meet, because each one becomes something to test.
  • Leave criterion numbers out of the policy text and keep the mapping separately; the 2027 draft renumbers the common criteria.

Approval and acknowledgement

Record who approved each policy and when, set a review date, and collect acknowledgements from everyone the policy applies to. These are evidence items like any other, and the SOC 2 evidence list covers how the auditor samples them. The SOC 2 checklist shows where policies fit in the readiness work.

Primary sources

Framework overviews

Related reading

FAQ

SOC 2 questions answered

Talk to us

How many policies does SOC 2 require?

There is no fixed number. Most reports rely on roughly a dozen policies, but what matters is that every criterion in scope is covered by a policy and a control, whether that takes one document or twenty.

Can we use SOC 2 policy templates?

As a starting point, yes. Edit every frequency, tool and role to match what you actually do before approving it, because the auditor tests the policy as written.

Who has to approve SOC 2 policies?

Management with authority over the area, often the CEO or CTO for the information security policy and the relevant owner for the others. The approval and its date are evidence the auditor will ask for.

How often should SOC 2 policies be reviewed?

At least annually is the common expectation, and whenever something significant changes. Whatever frequency the policy states is the one the auditor tests.