Which policies SOC 2 requires and why the auditor reads them first
SOC 2 does not prescribe a fixed list of named policies, but it does require management to deploy its controls through written policies and procedures. That requirement sits in CC5.3, which restates COSO Principle 12. In practice, the independent CPA firm will expect written policies covering code of conduct, information security, access, change management, incident response, vendor management, and retention and disposal, plus privacy policies if privacy is in scope. If your customer or contract now requires SOC 2, the first decision is not which template to download. It is which criteria your system must satisfy, because that choice determines which policies the examiner will test.
The examination depends on scope. Common criteria CC1 to CC9 apply to every SOC 2, so every report needs policies that support the control environment, communication, risk assessment, monitoring, control activities, access, operations, change and risk mitigation. Adding availability, confidentiality, processing integrity or privacy brings in additional policy areas, such as backup and recovery, data classification, data retention and disposal, or privacy notice and consent. Ask your CPA firm to confirm the category selection in writing before you draft anything, because a policy written for a narrower scope will leave gaps the auditor will find in fieldwork.
The organization owns these policies. Leadership must approve them, assign owners, communicate them to staff and contractors, and review them on a schedule the organization sets and then follows. An implementation partner can draft the policy set, align each document to the systems you actually run, and map it to the criteria, but it cannot approve policies or attest that they operate. Only the independent CPA firm tests them against the system and issues the report. Confirm with your auditor how it expects policy review dates and version history to be evidenced, since a policy that has not been reviewed on its stated cycle is an easy exception to raise.
Which policies to write and the criteria each one supports
A SOC 2 auditor tests whether each policy exists, is approved, is communicated and is followed, so every policy should trace to a specific criterion. The cards below pair each policy with the criteria it supports, and the criteria, not a vendor template, decide what the text must cover.
Code of conduct
Sets the standards of conduct employees and contractors must meet and the expectation that management and the board model them. Under CC1.1, the points of focus ask you to set the tone at the top, evaluate adherence and address deviations in a timely manner. Check that the policy covers contractors and vendor staff, not only employees, and that each person acknowledges it on joining and at set intervals.
Information security policy
The top-level policy that states management's security objectives and assigns responsibility for the control environment. It is the document an auditor reads first and anchors the rest of the set through CC5.3, which asks for policies that set expectations and procedures that carry them out. Keep it short enough that the team follows it, and confirm that a named owner reviews it on a defined cycle.
Access control policy
Defines how users are registered, authorized, given roles and removed from logical and physical systems, which the CC6 criteria cover in detail. The policy should name who approves access, how least privilege is applied and how privileged accounts are handled. Check that the procedures it describes match the joiner, mover and leaver records and the periodic access reviews you actually run.
Change management policy
Governs how changes to infrastructure, software, data and procedures are authorized, built, tested, approved and deployed, as the CC8 criteria require. Write it so that production changes need a recorded approval and a testing step that the team can show from its ticketing and pipeline records. Confirm that emergency changes have a defined path and a after-the-fact review.
Incident response policy
Establishes the defined incident response program that CC7.4 requires for understanding, containing, remediating and communicating security incidents. Name the roles, severity levels, escalation paths and the people authorized to notify customers. Check that the plan is tested, that post-incident reviews are recorded, and that the incident tickets an auditor samples follow the steps the policy describes.
Vendor management policy
Sets how vendors and business partners are assessed, assigned an owner, monitored and exited, which the CC9.2 points of focus describe in detail. The policy should require a risk assessment before engagement, record who is responsible for each vendor and state how assessment findings and performance problems are handled. Confirm that subservice organizations, including your cloud provider, appear in the inventory with their own SOC reports on file.
Communication and awareness policy
Covers how responsibilities, incident reporting methods, system changes and security awareness information reach employees, the board and external parties, as CC2.2 describes. Specify the reporting channels, including any separate line for concerns, and the training cadence. Check that training completion records and the communication to customers about system changes can be produced for the audit period.
How CC5.3 turns criteria into policies and procedures
CC5.3 asks the organization to deploy its control activities through policies that set expectations and procedures that carry those expectations out. It restates COSO Principle 12 inside the 2017 Trust Services Criteria, and it is the criterion that explains why a SOC 2 auditor reads your policies before testing anything else. A policy states what management expects, such as who may approve production access or how a vulnerability is triaged. A procedure states the steps people follow to make that happen, and the evidence those steps leave behind. An auditor testing CC5.3 will look for both halves: a written expectation for each control area in scope, and proof that someone actually performs the procedure on the dates the examination period covers.
The points of focus under CC5.3 give you a practical checklist for each policy. They ask whether policies and procedures support management's directives, whether a named role is responsible and accountable for carrying out each one, whether tasks are performed in a timely way, whether corrective action follows when something fails, whether the people doing the work are competent, and whether policies are reassessed over time. Points of focus are characteristics that help you design and evaluate controls. They are not separate requirements, and not every one has to be addressed. Decide, for each policy, which points of focus your design needs and document that choice, so the reasoning is visible when the auditor asks.
Confirm with your CPA firm how it will treat a minimum set of common controls and whether your policy set meets the criteria as you have drafted it. The AICPA staff FAQs published in December 2020, reported by the Journal of Accountancy, address that question, but how a particular examination applies them is the auditor's judgment. Ask the firm early which policies it expects to see, how it wants the review cycle recorded, and whether it needs a separate document for areas like vendor management that CC9.2 also touches. Settling those points before fieldwork is far cheaper than rewriting a policy set mid-period.
Policies for the board, staff and third parties
Leadership policies set the tone at the top and define how the board oversees the control environment. Under CC1.1, the points of focus ask the organization to demonstrate commitment to integrity and ethical values, establish standards of conduct, evaluate adherence to them, and address deviations in a timely manner. Under CC2.2, the points of focus include communicating with the board about internal control and system objectives. An auditor will want to see that management's expectations reached the board in a form the board could act on, and that the board's review was recorded.
- Board charter or governance statement: Sets the board's oversight of security and risk, with the frequency of reporting on controls.
- Standards of conduct for officers: Signed by directors and executives, covering conflicts of interest and reporting of deviations.
- Security reporting to the board: Minutes or a dated briefing pack showing incidents, open risks and audit findings were presented.
- Separate reporting line: A documented route for raising concerns to the board or a committee without going through management.
- Annual policy approval: Board or executive sign-off on the information security policy and the policy set, with the version and date.
Check that each acknowledgment is dated, names the policy version, and that the minutes or briefing record exist for every period under review.
Staff and contractor policies translate the standards of conduct into daily behavior. The CC1.1 points of focus extend to contractors and vendor employees, so the acceptable use and conduct policies must reach people who are not on payroll. The CC2.2 points of focus add security awareness information and incident reporting methods, so each person needs to know what to report and where to report it. Auditors usually sample individual acknowledgments, so the record has to tie a named person to a named version.
- Code of conduct: Expected behavior, confidentiality duties, and the consequence of breaches, acknowledged at hire and when the policy changes.
- Acceptable use policy: Rules for company devices, accounts, personal devices and data handling, with the prohibited activities listed.
- Incident reporting instructions: The channel, the time frame for reporting, and who to contact after hours, written so a new hire can follow it.
- Awareness completion: Training completion linked to the acknowledgment of the policy, so the two records can be matched.
- Contractor coverage: The same acknowledgments applied to contractors through a contract clause or a signed agreement.
Check that the acknowledgment is captured in your identity or HR system with a timestamp, that leavers are removed from the list of active acknowledgers, and that contractors are not missing from the population an auditor samples.
Third-party policies address CC9.2, which covers requirements for vendor and partner engagements, risk assessment of those parties, assignment of responsibility, communication and exception handling, performance assessment, remediation of issues found, and termination. Regulators have published guidance on third-party risk management, such as the 2023 interagency guidance in the Federal Register, which many customers read alongside SOC 2 when they assess a supplier. The vendor policy must say how you decide which suppliers matter and how you act on what they tell you.
- Vendor risk assessment: A tiering method that ranks suppliers by data access and criticality, with a reassessment schedule.
- Contract requirements: Security, confidentiality and incident notification terms that every material vendor must accept.
- Exception handling: A documented route for approving, time-limiting and reviewing any vendor that cannot meet a requirement.
- Performance review: Periodic review of vendor reports, such as SOC reports, with issues logged and followed up.
- Termination steps: Data return or deletion, access revocation and records kept when a relationship ends.
For acknowledgment, vendors sign the contract or a supplier code of conduct, and your team records the risk tier and the date of the last review. Confirm that the reassessment cadence in your policy matches what you actually performed.
How often policies must be reviewed and what proves it
SOC 2 does not hand you a fixed review interval for policies. The relevant point of focus sits under CC5.3, which restates COSO Principle 12: controls operate through policies that set expectations and procedures that carry them out, and management reassesses those policies and procedures. The auditor is therefore looking for a credible rhythm of reassessment that you define and then follow, not a number printed in a template. Set the interval in your own information security policy, choose one that your team can actually keep (an annual review with earlier reviews after material changes is a common pattern), and write down who owns each review. Confirm with your CPA firm, before fieldwork, what cadence and what triggers they expect you to document.
A Type 2 examination tests operating effectiveness over an observation period, so the auditor samples evidence from across that window rather than checking a single moment. For each policy, the evidence that matters is the record of approval by an accountable owner, the review date, the version history showing what changed, and the change tickets or meeting notes that link a revision to a real event such as a new vendor, a platform migration or an incident. A policy last approved before the period began, with no reviews inside it, is an exception waiting to be written up, even if the text is sound. Keep these records in a system that preserves history, such as a version-controlled document repository, so the approval trail cannot be rewritten after the fact.
Before the period starts, produce a short review calendar for every policy in scope and run one full cycle of it on purpose, including the approval signature and the change log. This is cheaper than discovering gaps during sampling. Name the triggers that force an out-of-cycle review, such as a material change to hosting, a significant incident, a new subservice organization or a change to customer commitments, and confirm that your change management records show when those triggers were acted on. Ask your auditor how they plan to sample policy reviews and whether they expect a specific evidence format. Their answer, not a generic template, should set the standard you build to.
How to write the policy set in order
This sequence turns the criteria your auditor will test into a short set of approved, owned and acknowledged policies, so each policy traces back to a control and each control has evidence behind it.
Confirm scope and categories
Start by fixing the system boundary, the services and locations in scope, and which Trust Services categories you will report on. Security and the common criteria CC1 to CC9 apply to every SOC 2, while availability, confidentiality, processing integrity and privacy are added only when they are in scope. Leadership and the engagement owner sign off on this scope statement. The output is a written scope record that names the categories, so you know which criteria the policy set must cover.
Map each in-scope criterion to a control
Work through every in-scope criterion and name the control that operates against it, the person who performs it and the system or record that proves it ran. Use the points of focus as prompts for designing and checking controls, not as a checklist you must fully address. Where one control serves several criteria, record each link so the auditor can follow it. The output is a control matrix that lists criterion, control, owner, frequency and evidence source for every in-scope criterion.
Draft policies from the control matrix
Write each policy from the control matrix rather than from a generic template, so every policy states the expectations and procedures that a named control carries out. Cover the areas the control matrix requires, such as code of conduct, information security, access, change, incident response, vendor management and retention and disposal. Keep each policy specific enough that a reviewer can test it against the system you actually run. The output is a set of draft policies, each tagged with the criteria it supports.
Assign owners and accountability
Give every policy and every control a named owner who is accountable for executing it and for raising problems. Confirm that each owner accepts the role and understands the procedure they must perform. Where a control depends on a third party, such as a cloud provider or an identity platform, record who inside your organization oversees that dependency. The output is an ownership table that matches each policy and control to a person, which the auditor will use to ask who performs each activity.
Obtain management approval
Send the draft policies and the control matrix to the people with authority over the organization, typically the executive who is accountable for security, and record their approval with a date and version number. Approval should reflect that management has reviewed the controls and accepted the responsibilities they imply. Do not approve a policy that describes controls nobody operates, because the auditor will test whatever the policy states. The output is an approved policy set with version history and recorded approval.
Publish and collect acknowledgments
Publish the approved policies where staff and contractors can find them, then ask each person in scope to acknowledge that they have read and will follow the policies that apply to their role. Track acknowledgments against your current roster and follow up on gaps, including contractors and any vendor staff with access to in-scope systems. The output is acknowledgment records tied to policy versions, which auditors commonly sample as evidence that expectations were communicated.
Schedule review and keep a review log
Set a review cycle for each policy and record each review in a log that shows the date, the reviewer, what changed and why. Trigger an early review when the system, the scope or the control matrix changes, or when an incident or audit finding shows a gap. Confirm the review cadence with your auditor, because the report period and the frequency of each control will shape what evidence is needed. The output is a review log that shows policies stayed current throughout the examination period.
Policy gaps that cause auditor exceptions
Most SOC 2 policy exceptions come from documents that do not match how the company actually operates. The cards below cover the gaps auditors find most often when they compare the policy set against the system description and the evidence in the observation period.
Copied vendor templates
A template often names a password manager, SIEM, or ticketing system the company never adopted. Auditors compare each policy statement with the tools in use, so a named product that does not exist in the environment becomes an exception. Before sign-off, read every policy line by line and delete or replace any tool, role, or meeting that does not match your environment. Check that each control named in the policy has a matching owner and record in the control matrix.
Policies with no named owner
A policy that says the security team will review it gives an auditor nothing to test. Each policy should name one accountable role, such as the CTO or head of security, along with the people who maintain the procedure beneath it. Record the owner in the policy header and in the document register. Confirm that the owner has the authority and time to approve exceptions and changes, because an unowned policy usually stops being maintained within a few months.
Policies describing controls no one performs
Some policies promise quarterly access reviews, monthly vulnerability scans, or annual tabletop exercises that never happen. An auditor tests the stated frequency against tickets, logs, and sign-offs, and a missing run is an exception even when the control is sound. Only write down activities you can perform at the stated frequency and that produce a dated record. Walk each control with its operator and confirm that the artifact it creates is stored where the auditor can request it.
Missing acknowledgment records
Policies that staff never formally accept leave a gap in the CC1 and CC2 evidence on communicating responsibilities. Auditors commonly sample new hires and current employees and ask for proof that each person received and acknowledged the policies. Use an identity or HR system that records the version each person accepted and the date. Build the process so that new joiners acknowledge before access is granted, and existing staff re-acknowledge after each material revision.
Policies not updated after platform changes
A policy written for a single cloud account often stays unchanged after the company adds a second region, a container platform, or a new identity provider. Auditors look for evidence that policies were reassessed when the system changed, which ties to the CC5.3 point of focus on reassessing policies and procedures. Keep a change trigger list, such as new hosting, a new subservice organization, or a new data type, and review affected policies whenever one of those triggers occurs. Record each review, even when the outcome is no change.
Vendor management without assessment records
A vendor management policy that lists criteria but has no completed assessments fails the CC9.2 expectations on assessing vendor risk and performance. Auditors ask for the vendor inventory, the risk tier assigned to each vendor, the due diligence performed, and any issues raised and resolved. Keep a dated assessment file for each critical vendor, including the SOC report reviewed, its period, and the exceptions noted. Confirm which subservice organizations carry their own SOC reports, because the company will need to address them in its own review.
Questions buyers ask about SOC 2 policies
Is one SOC 2 policy enough for an examination?
No. A single document rarely covers what the criteria test, although the information security policy is the top-level document an auditor reads first. The criteria expect a set of policies that set expectations for access, change, incident response, vendor management, retention and disposal, and code of conduct, and each policy should connect to the controls that carry it out. Auditors look for a clear link from each criterion to a policy, a procedure and a record. A thin top-level policy with no supporting procedures usually produces exceptions. Confirm the expected set with your auditor early, because the scope of your system determines which policies apply.
Can we use a policy template from a vendor?
A template can be a starting point, but it is not enough on its own. Auditors test whether the policy describes what your organization actually does, and a generic document that names tools you do not run or roles you do not have creates findings. The criteria also ask that policies reflect your system objectives, boundaries and risks. The point of focus on communicating system objectives and boundaries reinforces this. Adapt each template to your architecture, owners and tooling, then confirm that the staff who follow it can describe how it applies to their work.
How much effort does writing the policy set take, and what drives it?
Effort depends mainly on how much of your environment is already documented and how many systems fall within scope. Organizations with an existing risk assessment, asset inventory and defined roles usually adapt policies faster than those starting from nothing. Drivers include the number of products and environments, whether privacy or availability categories are added, how many third parties you rely on, and how many people must review and approve each document. Policy writing is largely one-time work, but reviews, updates after architecture changes and annual approvals recur. Ask any consultant to show the proposed policy list, owners and review cadence in a written scope before agreeing to a price.
Can our own team write the policies without outside help?
Often yes, if someone owns the work and has time to tie each policy to real controls. The main risk is not writing quality but consistency: policies that contradict each other, or that promise reviews nobody performs. Internal teams usually struggle to judge what an auditor will accept, so an outside reviewer can be useful even when staff do the drafting. An implementation partner can help prepare systems, configuration and evidence for the controls the policies describe. Only the independent CPA firm performing the examination can issue the report, and the organization remains responsible for approving and operating its policies.
What will the auditor ask to see for the policies?
The auditor will typically request the current approved versions, the approval history, the review dates and evidence that staff received and acknowledged the policies. They then test whether the policies are followed. For example, if a policy requires access reviews, the auditor will ask for completed review records during the period. Expect questions on who owns each policy, how exceptions are approved, and how the policies are communicated to contractors and vendor personnel. Keep a single index that maps each policy to its criteria, owner, procedure and records so that sampling requests can be answered quickly.
How do policies change when privacy or confidentiality is in scope?
When confidentiality or privacy is in scope, the policy set has to address how confidential information is identified, protected and disposed of, and for privacy, how personal information is noticed, collected, used, retained, disclosed and accessed. Retention and disposal rules become a stand-alone concern, and vendor management must cover processors that handle personal information. Privacy adds criteria for notice and consent, which need documented sources and owners outside the security team. Confirm with your auditor which categories are in scope before drafting, because adding categories later often means reopening policies and extending testing.
What the organization owns and what an outside partner can prepare
Management owns the policy set. The organization's leadership decides which policies exist, who approves them, and what each one commits the company to do. A SOC 2 auditor expects those policies to be approved by someone with authority over the business, dated, and reviewed on a stated schedule. An implementation partner can draft the policies, map each statement to the criteria it supports (for example CC5.3 on deploying control activities through policies and procedures), and check that the wording matches the tools and processes you actually run. The partner cannot approve the policies on your behalf, and a policy that describes controls nobody operates will be found as an exception in testing.
Alignment work is where an outside partner usually adds the most value. Confirm that each policy names an accountable owner, that the access, change, incident response and vendor management rules point to systems that exist in your environment, and that the review dates in the policies match the review you will actually perform. Ask the partner to show how each policy statement becomes evidence, such as an exported access list, a ticket or a pipeline record. Those artifacts are what the examination samples, so the partner's work product should include an evidence source for every policy commitment.
The CPA firm examines and reports; management asserts. Only the independent CPA firm performs the SOC 2 examination and issues the report. Under AT-C section 205 of the AICPA attestation standards, the service auditor must obtain written representations from management, which is why management's assertion and the representation letter are management's responsibility. Management signs them, and no consultant or platform partner signs them for you. Before fieldwork, confirm with your auditor which policies and evidence they expect, and keep a written record of who approved each policy and when.
How DeployOpen prepares policies and evidence for your SOC 2
DeployOpen works on the policy set a SOC 2 examination expects by scoping which criteria your system touches, then preparing the policies, procedures and records that show those controls operate. Your organization owns its policies and approves them, and an independent CPA firm performs the examination and issues the report.
How we scope
- We confirm the trust services categories in scope and the system boundary the policies must cover.
- We map each policy to the common criteria it supports, starting from your actual systems and teams.
- Engagements are fixed-scope, with deliverables and timelines agreed in writing before work starts.
How we build
- We draft policies in plain language that match how your team really works and what your tools enforce.
- We set up the self-hosted systems that carry out each procedure, such as identity, ticketing and logging.
- We define review cadence, owners and approval records so each policy stays current between audits.
How we hand over
- You receive approved policy documents, a control-to-policy cross-reference and the procedures behind them.
- We document where each policy's evidence is generated, such as system reports, tickets or pipeline logs.
- We walk your team through evidence collection so you can answer the auditor's requests without us.
How we support
- Ongoing support and policy upkeep are available as a retainer, covering annual reviews and scope changes.
- We can help your team respond to auditor questions about policies, but the auditor decides what is sufficient.
- Confirm with us during scoping which support level, review cycle and response expectations apply to your program.

