SOC 2 CC6 evidence, without the spreadsheet scramble

What evidence SOC 2 CC6 access controls require, by control number. How to produce it continuously instead of scrambling the week before the audit.

7 min read · Last updated June 2026

Your Type II observation window started in January. The audit kicks off in six weeks. The CC6 controls (logical access security) are usually where teams either coast through or get found. The difference is what evidence you assembled while the observation window was running.

This is the practitioner's checklist. What CC6 requires by control number, what evidence the auditor wants for each, and how to produce that evidence without a wall-of-Excel scramble the week before the audit.

What CC6 is

The CC6 series sits inside the SOC 2 Trust Services Criteria under the Common Criteria. It covers logical and physical access controls. The Type II audit looks for evidence that controls operated effectively over the observation window (typically three to twelve months).

For access governance specifically, five sub-controls matter most.

  • CC6.1. Logical access security implementation.
  • CC6.2. Provisioning new users and credentials.
  • CC6.3. Modifying and removing access.
  • CC6.6. Protection from external threats.
  • CC6.7. Information transmission and removal restrictions.

The remaining sub-controls (CC6.4, CC6.5, CC6.8) cover physical access, asset disposal, and malicious software. Important, but outside the scope of access governance specifically.

CC6.1: Logical access security

The control: the entity implements logical access security software, infrastructure, and architectures over protected information assets to protect them from security events.

Evidence the auditor wants:

  • Policy documents describing the access control framework: role definitions, approval requirements, password policy, MFA enforcement.
  • Logical access security architecture diagram showing IdP, IGA layer, downstream systems.
  • A list of all production systems that fall under access governance scope.
  • Configuration evidence: MFA enforcement settings, password policy snapshots, session timeout settings.
  • Identity records for sample populations: dated role assignments, source-of-truth attribution (HRIS link).

Common failure: the policy document exists but doesn't match what's actually configured in the target systems. The auditor checks both.

CC6.2: User registration and authorization

The control: prior to issuing system credentials and granting system access, the entity registers and authorizes new internal and external users.

Evidence the auditor wants:

  • For each new hire in the observation window: HRIS record showing start date and role, plus the access grant log showing what was provisioned, when, and by whom (or by what policy).
  • For each contractor or external party with system access: registration record, access grant log, expected end date.
  • Sample population: the auditor typically pulls 25 to 40 access grants and asks for the full provisioning trail for each.
  • Authorization evidence: for above-baseline access, the request and approval records.
  • Birthright policy documentation: the role-to-access mapping that fires for new hires.

Common failure: the access grant exists in the target system but no record links it back to the HRIS event that triggered it. The auditor sees a user with access and no provisioning audit trail.

CC6.3: Access modification and removal

The control: the entity authorizes, modifies, or removes access to data, software, functions, and other protected information assets based on roles, responsibilities, or the system design and changes, giving consideration to least privilege and segregation of duties.

Evidence the auditor wants:

  • Joiner-mover-leaver (JML) workflow documentation.
  • For role changes in the observation window: HRIS record + access changes (additions and removals).
  • For each departure in the observation window: HRIS termination record + access revocation log showing every system access was removed, with timestamps.
  • Orphaned account scans: evidence that the team periodically checks for accounts belonging to departed users and remediates them.
  • Access certification campaign records: completion dates, reviewer decisions per entitlement, remediation timestamps for "remove" decisions.
  • Segregation of duties: examples of SoD conflicts detected and remediated, or attestation that conflicts are reviewed.

Common failure: the offboarding evidence covers SSO and the major SaaS apps but misses the long tail of non-SCIM apps where IT handled offboarding manually. The auditor finds an ex-employee with active access on an app that wasn't in the offboarding checklist.

CC6.6: External boundaries

The control: the entity implements logical access security measures to protect against threats from sources outside its system boundaries.

Evidence the auditor wants:

  • External user inventory: vendors, partners, contractors with system access.
  • Boundary controls: VPN configuration, network segmentation, public-facing system identification.
  • External authentication controls: SSO federation for external parties, MFA requirements.
  • Vendor access logs and periodic reviews.

Lighter touch for most mid-market companies, but the access boundary between internal employees and external parties is one the auditor will examine specifically.

CC6.7: Information transmission and removal

The control: the entity restricts the transmission, movement, and removal of information to authorized internal and external users and processes, and protects it during transmission, movement, or removal.

Evidence the auditor wants:

  • Data classification policy.
  • Encryption-in-transit evidence: TLS configuration snapshots, cipher suite audit.
  • Encryption-at-rest evidence for stored data.
  • Removal authorization: who can export data, who can delete records, who can transfer ownership during offboarding.

Adjacent to CC6.3: the offboarding flow handles data ownership transfer alongside access revocation.

What the auditor actually looks at

The Type II audit isn't a document review. The auditor pulls samples and traces individual events end to end.

For provisioning: "show me Sarah's onboarding. She joined March 15. What systems was she provisioned to? When? By whom? What policy authorized it?" The auditor wants a continuous trail from HRIS event to access grant in target system.

For offboarding: "show me everyone who left in April. For each one, what systems did they have access to? What was revoked? When?"

For access reviews: "show me the Q2 certification. Who was the reviewer for each application? What decisions did they make? Were 'remove' decisions actually executed? When?"

If your evidence requires assembling artifacts from five different systems (HRIS exports, SSO logs, app admin panels, ticketing system, manual spreadsheet) per sample, the auditor gets the picture. The control might still pass; the operational maturity rating won't.

How to produce this evidence without the scramble

Three things have to be true.

One audit trail per access event. Every grant, modification, and revocation captured in a single structured log, with the policy or human authorization that triggered it. Not five separate logs that have to be correlated by hand.

Trail covers the full stack. Every system under SOC 2 scope produces audit events in the same format. Including the non-SCIM apps. Including the contractor identities. Including the non-human identities.

Evidence exports as artifacts the auditor will accept. CSV or JSON, with timestamps, identity attributions, policy references, mapped to control numbers. The audit response is "for CC6.3, here's the artifact," not "let me query our system and pull something together."

Most GRC platforms (Drata, Vanta, Secureframe) handle the policy-document side of SOC 2 well. The access-event side is where the work happens. The IGA layer is what produces the events; the GRC platform is what packages them for the auditor.

Where Iden fits

Iden produces SOC 2 CC6 evidence continuously as the lifecycle runs. Every provisioning, modification, and revocation lands in the audit log mapped to the relevant control. The evidence pushes into Drata, Vanta, or Secureframe automatically. When the auditor asks for the Sarah-joined-March-15 trail, the export is a click.

Teams that have done one SOC 2 cycle on Iden tend to describe the second one as uneventful. That's the highest praise the audit experience can attract.

Frequently asked questions

Do we need a separate tool for SOC 2 CC6 evidence beyond our GRC platform?

The GRC platform handles the policy and framework side. It needs evidence to populate. For CC6 specifically, that evidence comes from your IGA layer (or wherever access-control activity happens). Drata, Vanta, and Secureframe all integrate with IGA tools to ingest access events. Without an IGA layer producing structured events, your GRC platform has to be fed manually.

How often should access certifications run for SOC 2?

Most teams run quarterly for SOC 2 Type II. Some run continuously for the certifications that touch sensitive data. The audit checks that the cadence matches your policy and that 'remove' decisions actually executed in the target systems.

What's the difference between an access grant and an access change for CC6 purposes?

Functionally similar; the auditor treats them slightly differently. A grant creates net-new access (CC6.2). A change modifies existing access (CC6.3). Both need the same evidence shape: who, what, when, under which policy.

Are non-human identities in scope for SOC 2 CC6?

Yes, under CC6.1, CC6.2, and CC6.3. Service accounts, API keys, OAuth grants, AI agents, anything that can perform an action on a system within audit scope. The auditor expects the same provisioning and offboarding evidence for them as for human users.

How long should we keep access audit logs?

SOC 2 doesn't prescribe a specific retention window, but most auditors expect 12 months minimum, covering the Type II observation window plus prior period. Some industries layered on top (HIPAA, ITAR, NERC CIP) require longer.

We had a near-miss during the observation window. How do we handle it?

Document it. SOC 2 auditors aren't looking for zero incidents; they're looking for incidents that were detected, responded to, and remediated according to documented procedures. The near-miss with a clear remediation trail is materially better than no detection at all. Include it in your incident log and reference the corrective action.

How do orphaned accounts affect SOC 2?

Orphaned accounts (ex-employees with active access) are a CC6.3 finding waiting to happen. If the auditor samples a departure and finds the access wasn't fully removed, that's a control gap. Most teams either run quarterly orphaned-account scans or implement continuous offboarding to make orphans impossible.

Can we use spreadsheets for the access review trail?

Yes, but auditors increasingly press on whether spreadsheet-based reviews demonstrate operating effectiveness. If the reviewer clicked 'approve' on a 200-row spreadsheet in three minutes, did the review actually happen? Evidence that captures reviewer activity (time spent per entitlement, context shown to reviewer, follow-up questions raised) is harder to challenge than a flat spreadsheet response.

What's the relationship between SOC 2 CC6 and ISO 27001 access controls?

They cover similar ground with different naming. SOC 2 CC6.2 maps to ISO 27001 A.5.16 (identity management) and A.5.17 (authentication information). CC6.3 maps to A.5.18 (access rights) and A.5.19 (supplier relationships). The same evidence usually satisfies both. The 2022 ISO 27001 revision renamed several controls; policy documents should reflect the current numbering.

We're targeting both SOC 2 and HIPAA. Do the evidence requirements overlap?

Substantially yes. HIPAA §164.308(a)(3) (workforce security) maps closely to CC6.2 + CC6.3. HIPAA §164.308(a)(4) (information access management) maps to CC6.3. The audit trail format that satisfies SOC 2 typically satisfies HIPAA with minor additions for PHI-specific access logging.