Office Consumer is reader-supported. We may earn an affiliate commission from qualified links on our site.

Does SOC 2 Require Penetration Testing? (w/Examples) + FAQs

No. SOC 2 does not require penetration testing in writing, and the AICPA Trust Services Criteria name no mandatory test. In practice, though, CPA audit firms in 2026 expect an annual independent penetration test as audit evidence for an unqualified SOC 2 Type II report, and auditors treat a missing test report as a red flag.

The written rule and the real-world rule point in different directions. The AICPA's criteria are principles-based, so they tell you what a control must achieve, not which tool proves it. Yet the points of focus under CC4.1 name penetration testing as one way management checks that controls work, and auditors treat a missing test report as a red flag rather than a neutral gap.

That gap matters because SOC 2 has become the price of admission for selling software and services to larger customers. A first-year SOC 2 program costs at least $25,000 for a small startup in 2026, so a failed or delayed audit wastes real money. The companies most affected are SaaS vendors, cloud platforms, fintechs and managed service providers that hold customer data and face a procurement team asking for a Type II report.

This article reflects AICPA professional standards and audit practice as of October 2026 and covers the 2026 audit year. Rules change, so confirm current figures before you act.

  • 📜 What the Trust Services Criteria say about penetration testing, and which criteria auditors lean on
  • ⚖️ Why "not required" and "expected" are both true, and what that means for your audit
  • 🗓️ How Type I and Type II reports change when and how often you test
  • 💵 What a SOC 2 penetration test costs in 2026
  • 🚫 The mistakes that get a test report rejected, and how to avoid each one

What SOC 2 Says About Penetration Testing

SOC 2 is an attestation, not a certification, and its text never orders a penetration test. The framework rests on the AICPA Trust Services Criteria, and a CPA firm attests that your controls meet those criteria. The AICPA promulgates the professional standards for SOC engagements and describes SOC as a suite of services CPAs provide over a service organization's system-level controls. The consequence shows up in the auditor's opinion instead.

The Five Trust Services Criteria

The framework is built on five Trust Services Criteria defined by the AICPA: Security, Availability, Processing Integrity, Confidentiality and Privacy. Security is the only criterion required for every SOC 2 audit, and the other four are added to scope only when your service commitments call for them. Security is also called the Common Criteria, and it breaks into nine categories numbered CC1 through CC9. The penetration testing question lives inside Security, so it touches every SOC 2 report no matter how narrow the scope.

For the reader, this means you cannot avoid the question by leaving a category out. Even a report scoped to Security alone will reach CC4.1 and CC7.1, the two criteria where testing evidence is judged. A common misconception is that a small company with a "Security-only" report faces lighter technical scrutiny.

CC4.1 and Its Points of Focus

CC4.1 is the criterion that names penetration testing. It corresponds to COSO Principle 12. The AICPA attaches points of focus to each criterion, which are examples of how management might meet it. The CC4.1 points of focus state that management conducts penetration testing to evaluate whether controls function as designed.

The plain-English meaning is simple: you must have some way to prove your controls work, and the AICPA's own example of that proof is a penetration test. The next step for most readers is to accept that CC4.1 is the hook and plan a test that produces evidence mapped to it.

CC7.1 and Vulnerability Scanning

CC7 covers System Operations. A scan is automated and broad; a penetration test is manual and deep. Auditors expect both, because scanning alone does not satisfy the expectation for manual testing by an independent third party that can find business logic flaws.

The misconception here costs companies real money: many teams present a scanner export as their "pen test." Auditors increasingly know the difference, and a scanner report presented as a penetration test is a common reason evidence is kicked back. CC6 and CC7 are also where auditors typically find the most gaps in first-time engagements, so weak evidence under CC7.1 rarely goes unnoticed.

Why Auditors Treat It as Required

The expectation is a de facto mandate, not a written one. CPA audit firms overwhelmingly expect an annual independent third-party penetration test report as audit evidence for an unqualified SOC 2 Type II report, according to a 2026 SOC 2 testing guide. An unqualified opinion is the clean result customers want. A qualified opinion notes exceptions, and procurement teams read it as a warning.

The practical consequence is that a company without a recent test report faces hard scrutiny during the audit, and the test is practically required though not technically mandatory. The right next step is to ask your auditor, before the observation period starts, what test evidence they will accept.

Which Situation Applies to You?

Your answer depends on your report type, your audit boundary and your customers' demands. SOC 2 flexibility means two companies with similar headcounts can face different expectations, so start by locating yourself in the list below.

  • You are pursuing a Type I report for the first time. One penetration test near your report date usually serves as design evidence, and one test often takes 10 to 12 working days to run.
  • You are pursuing a Type II report. The test must fall inside your observation period of 3 to 12 months, and several tests across that window make a stronger case than one.
  • You already have a Type II report and are renewing. Expect an annual test timed inside the new window, plus extra testing if your audit boundary changed in a material way.
  • You added a new product, API or admin portal since the last audit. Every system inside the boundary needs coverage, so the new surface must be tested or the auditor will compare scope against your system description and flag the gap.
  • You run only automated scanning today. Scans satisfy the CC7.1 detection baseline, but you still need a manual, independent test for CC4.1 evidence.
  • Your customers ask for the pen test report directly. Full reports hold exploitation steps, so most providers issue an attestation letter summarizing scope, method and retest results for procurement teams instead.

The thread running through every situation is the same: the auditor wants proof that someone tried to break your controls, found what they could, and that you fixed it. If you cannot show all three parts, the type of report you chose will not save you. Companies that fit two situations at once, such as a renewal with a new API, should plan the larger scope and budget for it.

Type I Versus Type II Testing Timing

The report type changes when you test, how often, and what evidence you keep. A Type I report evaluates whether controls are designed well enough at a single point in time. A Type II report evaluates whether controls operated effectively across a window that is typically 3 to 12 months, and enterprise buyers usually demand Type II. The timing rules below come from auditor practice, not the AICPA text, so confirm them with your own firm.

QuestionType IType II
What the auditor judgesControl design on a set dateDesign and operating effectiveness over the period
Observation periodA single point in timeTypically 3 to 12 months
Penetration tests expectedOne test near the report date is often enoughAnnual test inside the window; several tests make a stronger case
Remediation evidenceFindings and a fix planCritical and High fixes closed within policy SLAs before the period closes, with retest proof

The timing trap in Type II is testing too early. A test in month two of a twelve-month window may not satisfy an auditor who reviews evidence in month eleven, because findings and fixes are no longer current. The opposite trap is testing too late: a test requested three weeks before the auditor needs the report gets compressed scoping and often a rush premium on the base quote.

Remediation deadlines are the other timing rule. Critical and High vulnerabilities found during testing must be remediated within strict policy SLAs before the audit period closes, and independent retesting must verify the fix. The policy SLAs are yours to set, but one practitioner guide suggests fixing critical issues within 24 hours, high ones within 7 days and medium ones within 30 days. Missing your own SLA becomes an exception in the report.

What a SOC 2 Penetration Test Costs

A SOC 2 penetration test for a typical seed to Series A SaaS company runs $8,000 to $20,000 in 2026, and complex multi-service environments cost more, according to a September 2026 pricing guide. A compliance platform's 2026 cost breakdown puts the same line item at $5,000 to $15,000 and notes that many customers demand an annual test as part of due diligence. Scope, timing and evidence format drive the spread, not company size alone.

The cheap end carries its own risk. Penetration tests priced under $4,000 tend to lean on automated tools, miss real issues and give false comfort, and an auditor who rejects the report forces you to pay twice. A quote that excludes retesting also looks cheaper up front and costs more once the auditor asks for proof that findings were fixed.

Reader Scenarios From Real Situations

Most audit trouble around penetration testing comes from scope, timing or evidence format, not from the test itself. The three tables below map common situations to what usually happens next.

Scope situationWhat the auditor typically does
Test covers the production app but not a new customer-facing APICompares test scope to the system description and flags the untested API as a gap
Test covers staging rather than productionQuestions whether the evidence represents the in-scope system
Test covers all systems in the boundary and maps findings to CC4.1 and CC7.1Accepts the report as evidence and moves to remediation review
Timing situationWhat usually happens
Test run in month two of a twelve-month Type II windowAuditor may ask for newer evidence when fieldwork runs in month eleven
Test booked three weeks before the window closesVendor compresses scoping and often adds a rush premium
Test booked six to eight weeks before the window closesNormal turnaround and normal pricing
Evidence situationResult
Automated scanner export labeled as a penetration testEvidence kicked back; a manual independent test is still needed
Findings listed with no retest proofAuditor asks for remediation and retest evidence before closing the point
Full report plus attestation letter and retest resultsReport accepted and the letter can go to customers without exposing exploit details

The pattern across all three tables is that a penetration test is only as useful to the audit as its paperwork. A clean technical result with a mismatched scope or a stale date can fail where a messier result with full remediation records passes.

Three Named Examples

Each person below faces a different version of the same expectation. The people are illustrative; the rules they run into are the ones described above.

Maya runs a seed-stage analytics startup with eight employees and a single web app. Her first customer with a security questionnaire asks for a Type I report, and her auditor tells her a design-stage review will need some proof that controls work. She books one penetration test that takes about two working weeks, pays near the $8,000 low end for a single-app scope, and files the report as CC4.1 evidence.

Daniel leads engineering at a Series A fintech starting a twelve-month Type II window in January 2026. He runs a penetration test in February, closes the findings, and assumes he is done. When fieldwork begins in November, the auditor asks whether the February test still represents the environment after two major releases and a new partner API. Daniel ends up ordering a second test on a rush timeline, paying a premium he could have avoided by scheduling the test in the back half of the window.

Priya owns a managed service provider whose largest client demands a Type II report. Her team already runs weekly vulnerability scans, so she submits the scanner export as her penetration test. The auditor rejects it because scans satisfy the CC7.1 detection baseline but not the expectation for manual, independent testing. She hires a third party, maps the report to CC4.1 and CC7.1, and keeps the scans running as the ongoing control the auditor also wants to see.

Mistakes to Avoid

The seven mistakes below can lead to rejected test evidence. Each one has a direct cost in audit delay, extra fees or a qualified opinion.

  • Treating "not required" as "not needed." The AICPA text does not mandate a test, but auditors expect one, and showing up without it invites hard scrutiny and possible exceptions.
  • Presenting a vulnerability scan as a penetration test. Auditors know the difference, and the evidence gets kicked back, leaving you to pay for a real test on a shorter timeline.
  • Testing outside the Type II observation period. A test before the window opens or after it closes provides limited value, and the auditor may not accept it.
  • Letting scope drift from the system description. Any system inside the boundary that was not tested becomes a visible gap when the auditor compares the two documents.
  • Skipping retesting. Critical and High findings must be fixed within policy SLAs and verified by independent retest before the period closes; findings with no closure become exceptions.
  • Buying the cheapest quote. Tests priced under $4,000 tend to be tool-driven, and a report the auditor rejects means paying twice.
  • Booking late. A test requested three weeks before the deadline gets compressed scoping and often a rush premium.

The common thread is that the test is a compliance deliverable with a calendar, not just a security exercise. Companies that treat the audit timeline as part of the test's scope avoid nearly all of these problems.

Do's and Don'ts

Plan the test as audit evidence from the first scoping call. The lists below turn the rules above into actions you can take this quarter.

Do:

  • Ask your auditor, before the observation period starts, what test evidence and report format they will accept.
  • Match the test scope to your SOC 2 system boundary, including every API, admin portal and cloud account inside it.
  • Schedule the test inside the Type II window with enough runway, ideally six to eight weeks before the window closes.
  • Keep vulnerability scanning running as the ongoing CC7.1 control alongside the periodic manual test.
  • Request retesting and an attestation letter as part of the quote so remediation proof and customer-facing summaries are covered.

Don't:

  • Don't submit a scanner export or an AI-only platform report as a penetration test.
  • Don't test staging when production is the in-scope system.
  • Don't leave Critical or High findings open past your own policy SLA; the auditor will read the policy and check.
  • Don't share the full exploit-level report with customers; use the attestation letter instead.
  • Don't assume one test satisfies both SOC 2 and another framework such as HIPAA without checking scope and timing with both leads.

Following the do's costs little beyond planning time, while each don't maps to a rejection or exception that has already happened to someone else. Treat the lists as a checklist for the scoping call with your vendor.

Pros and Cons of Testing for SOC 2

Running an independent penetration test for SOC 2 is almost always worth it, but it carries real costs and limits. Weigh both sides before you sign a quote.

Pros:

  • It is the AICPA's own example of CC4.1 evidence, so it answers the auditor's core question directly.
  • It finds business logic flaws and chained exploits that automated scans miss, so the security gain is real and not just paperwork.
  • Retest records show steady improvement over time, which is exactly the kind of evidence a Type II report needs.
  • An attestation letter gives sales and procurement a document they can share without exposing exploit details.
  • It removes the red flag an auditor sees when a company arrives with no recent test report.

Cons:

  • It costs $8,000 to $20,000 for a typical early-stage SaaS scope in 2026, before fixes and engineering time.
  • Timing is unforgiving: a test that is too early, too late or outside the window may have to be repeated.
  • Findings create remediation work with deadlines, and missing your own SLA becomes an exception in the report.
  • Scope must mirror the system boundary, so a growing product adds testing hours every cycle.
  • A cheap or tool-driven test can fail auditor review and force a second engagement.

For most companies that need a SOC 2 report to close deals, the pros win because the alternative is a qualified opinion or a stalled sale. The cons are mostly avoidable with scheduling and a clear scope, which is why they appear again in the next-steps list.

What to Do Next

Start with your auditor and your calendar, then buy the test. The order below follows the way a first-time company moves through the process in 2026.

  1. Confirm your report type and observation period with your CPA firm, and ask what penetration test evidence they will accept.
  2. Write down every system inside your SOC 2 boundary, including APIs, admin tools and cloud accounts, so the test scope can mirror it.
  3. Pick a test date inside the window, six to eight weeks before fieldwork, and leave time for fixes and retesting.
  4. Request quotes from independent vendors who can map findings to CC4.1 and CC7.1, include retesting and issue an attestation letter; budget $8,000 to $20,000 for a typical SaaS scope.
  5. Set your remediation SLAs in policy before the test so Critical and High findings have deadlines the auditor can check.
  6. Run the test, fix findings within the SLAs, and obtain retest proof.
  7. File the full report, the retest evidence and the attestation letter with your other audit evidence, and keep vulnerability scans running as the ongoing control.

If your environment is complex or you have never been through an audit, a compliance consultant or virtual CISO can scope the test and the evidence package for you; the same 2026 cost guide puts consultant support at $5,000 to $25,000 or more.

FAQs

Is penetration testing mandatory for SOC 2?

No. The AICPA Trust Services Criteria do not mandate a penetration test. In 2026, though, CPA audit firms expect an annual independent test as CC4.1 evidence, and a Type II report without one faces hard scrutiny or exceptions.

Which SOC 2 criterion mentions penetration testing?

CC4.1 (COSO Principle 12). Its points of focus state that management uses evaluations, including penetration testing, to check whether controls function as designed. CC7.1 covers the related expectation for periodic vulnerability scanning.

How often do I need a penetration test for SOC 2?

At least once a year is the common baseline in 2026. For a Type II report, the test should fall inside the observation period, and several tests across the window make a stronger case than one.

Can a vulnerability scan replace a penetration test for SOC 2?

No. Scans satisfy the CC7.1 detection baseline, but auditors expect manual testing by an independent third party for CC4.1 evidence. A scanner export labeled as a penetration test is a common reason evidence gets rejected.

How much does a SOC 2 penetration test cost in 2026?

Between $8,000 and $20,000 for a typical seed to Series A SaaS scope, with multi-service environments running higher. A compliance platform's 2026 breakdown puts the same line at $5,000 to $15,000.

Does a SOC 2 Type I report need a penetration test?

Usually one test is enough. Type I judges control design at a point in time, so a single test near the report date typically serves as evidence. The test often takes 10 to 12 working days.

When should I run the penetration test in a Type II window?

Inside the observation period, with six to eight weeks of runway before it closes. Testing too early can leave evidence stale by fieldwork; testing too late brings compressed scoping and a rush premium.

Do I have to fix every finding before the audit ends?

Critical and High findings, yes. Auditors expect those remediated within your policy SLAs before the audit period closes, with independent retest proof. Lower-severity items follow your documented timeline.

Who should perform the SOC 2 penetration test?

An independent third party.

Can I give customers my penetration test report?

Usually not the full report. It contains exploitation steps and architecture details. Providers issue an attestation letter summarizing scope, method and retest results for auditors and procurement teams instead.

Is Security the only criterion I need in a SOC 2 report?

Yes. Security is the only Trust Services Criteria required for every SOC 2 audit. Availability, Processing Integrity, Confidentiality and Privacy are added only when your service commitments call for them.

What happens if my auditor rejects the penetration test report?

You need a follow-up engagement or revised report before the auditor accepts it as evidence. Scope mismatch, stale timing and scanner-only methods are the usual causes, which is why confirming format up front matters.