By Marçal Santos, CISO and founder of SecureLeap. CISM and CDPSE (ISACA). Previously security roles at Aircall, Citibank and Talkdesk. 20+ years in cybersecurity and more than 30 SOC 2 audits led over his career.
No, SOC 2 does not require a penetration test. The AICPA Trust Services Criteria behind every SOC 2 report never make one mandatory. Penetration testing appears once, as an example in the points of focus for criterion CC4.1, and points of focus are guidance, not requirements. Most SOC 2 auditors we work with still ask for one, and enterprise buyers ask for the report in their security reviews whatever the criteria say.
"SOC 2 does not require a pentest, and I have completed an audit without one. The audit is rarely the reason to test. The first enterprise security questionnaire after the report usually is," says Marçal Santos, CISO and founder of SecureLeap.
Whether you should skip it depends on three things: what your own control set commits you to, what your customer contracts say, and what your buyers' security teams will ask for next. This guide covers all three, then how to scope and time a SOC 2 pentest so the report counts as evidence the first time.
Key takeaways
- Not required by name. Penetration testing is one item in a list of example evaluations under CC4.1, next to internal audit assessments, vulnerability scans and third-party assessments.
- Your own controls can make it required. If your control list includes an annual penetration test, the auditor tests it like any other control, and a missing test becomes an exception.
- Scanning is closer to a requirement. The CC7.1 points of focus list vulnerability scans on a schedule and after significant changes.
- Scope follows your system description: the production app, its APIs and the cloud environment that holds customer data.
- Timing: for Type 1, a test from the 12 months before the report date is normally accepted; for Type 2, test inside the observation window and leave time to fix and retest before it closes.
- Test once a year and after significant change. Most startups pay $4,000 to $8,000 to test one web app and its API.
Does SOC 2 require a penetration test?
SOC 2 reports are built on the 2017 AICPA Trust Services Criteria, with points of focus revised in 2022. The criteria are the requirements. The points of focus under each criterion are examples, and the AICPA says so directly: "Use of the trust services criteria does not require an assessment of whether each point of focus is addressed."
Penetration testing sits under CC4.1, the monitoring criterion: the organization selects, develops and performs ongoing or separate evaluations to check that its internal controls are present and functioning. Its point of focus Considers Different Types of Ongoing and Separate Evaluations lists the evaluations management might use: "first- and second-line monitoring and control testing, internal audit assessments, compliance assessments, resilience assessments, vulnerability scans, security assessment, penetration testing, and third-party assessments." That list is the only place the criteria mention penetration testing. A pentest is one way to meet CC4.1, not the only one.
When a penetration test becomes required for you
The criteria do not require one, but four situations do:
- Your control set lists it. Most compliance platform templates include a control for an annual third-party penetration test. Once it is in your control matrix, the auditor tests it, and no test means an exception in the report. Remove the control before the audit period starts if you do not plan to test.
- A customer contract requires it. Security addenda in enterprise contracts often commit you to an annual third-party test. Mike DeKock, a CPA writing for the Cloud Security Alliance, puts the test simply: "Do you have any customer contracts requiring a penetration test? If yes, then it's required."
- Your auditor expects it. Some CPA firms treat a missing test as a gap in their CC4.1 testing and ask what replaced it. Ask before you sign the engagement letter, using the questions in our guide to the best SOC 2 auditors.
- Your buyers ask for it. Enterprise security questionnaires ask for the date of your last third-party penetration test and a summary of the findings. A SOC 2 report without one moves the question to the next review; it does not answer it. In DeKock's words, "Penetration testing is the most common requirement we hear as the last component to close a deal with an enterprise customer (along with having a SOC 2 report)."
Audit firms put the rule the same way. "Neither are technically required," writes BARR Advisory, a CPA firm, about vulnerability scans and penetration tests for SOC exams; what is required is that you assess and manage security risks from every source.
Which Trust Services Criteria a penetration test supports
A pentest report is evidence for more than one criterion. These are the four where auditors usually place it:
| Criterion | What it asks | What the pentest report shows |
|---|---|---|
| CC3.2 Risk identification | Identify risks to your objectives and analyze how to manage them | Exploitable weaknesses that feed your risk assessment and risk register |
| CC4.1 Monitoring | Run ongoing and separate evaluations of your controls | An independent separate evaluation, the example the points of focus name |
| CC4.2 Deficiencies | Report control deficiencies, in time, to the people who fix them | Findings with owners, remediation tickets and a retest that closes them |
| CC7.1 Vulnerability detection | Detect configuration changes and new vulnerabilities | Flaws scanners miss, such as broken access control between roles and tenants |
For the full criteria set and which categories you need, see the SOC 2 Trust Services Criteria.
Penetration test or vulnerability scan?
If either of the two is close to a requirement, it is scanning. The CC7.1 points of focus include running vulnerability scans on a periodic basis and after any significant change in the environment, and fixing what they find on time. Auditors sample the scan reports and the remediation tickets against the deadlines in your policy, as covered in SOC 2 vulnerability management.
A scan checks for known issues. A pentest has a person try to break the application: log in as one customer and read another customer's data, escalate a basic role to admin, chain two low findings into a high one. Those are the findings enterprise security reviewers ask about, and scanners rarely produce them. Our vulnerability scan vs penetration test comparison goes through the differences in detail. For SOC 2 you need the scans either way; the pentest is the part you can choose. For the wider picture, including manual vs AI-driven tests, see our guide to penetration testing for startups.
What the auditor checks when there is no penetration test
Skipping the test puts the weight on the rest of your monitoring evidence. Have these ready for the whole audit period:
- Vulnerability scan reports at the cadence your policy states, monthly for most startups
- Remediation tickets showing findings closed within your policy's deadlines
- Another evaluation from the CC4.1 list: an internal audit assessment, a third-party assessment such as an ISO 27001 certification audit, or documented control testing by your own team
- A risk assessment that records the decision not to test this period, and why
- A control list that does not promise a pentest you did not run
The audit can pass on that evidence. The cost shows up in sales: the first enterprise questionnaire asks for your latest pentest summary, and having none starts a longer review. Our guide to answering vendor security questionnaires lists what buyers ask for and what to keep ready.
How to scope a SOC 2 penetration test
The scope boundary is your system description: the services, infrastructure and data the SOC 2 report covers. If you have not set that yet, start with how to scope your SOC 2 audit. For most startups, the pentest scope then looks like this:
| Component | In scope? | Why |
|---|---|---|
| Production web app | Yes | It is the service the SOC 2 report describes |
| Public and internal APIs | Yes | Customers and your own front end reach customer data through them |
| Authentication, roles and tenant separation | Yes | Access between customers is the first thing buyers ask about |
| Production cloud account | Yes, as a configuration review or cloud test | It hosts customer data and sits inside the system description |
| Mobile app | If it is part of the service | Same customer data, different client |
| Staging and development | Only if they hold real customer data | Matches how the audit scope treats them |
| Office network and laptops | Rarely | Endpoint and access controls cover them; test only on-premises systems that are in scope |
| Marketing website | No, unless it handles logins or customer data | Outside the system description |
Grey box testing fits this scope best. Testers get a user account for each role, so they spend their time on authorization between roles and tenants instead of guessing logins. Grey box penetration testing explains why it is the usual choice for startups.
Write down what you excluded and why, and put the same scope in the report. The auditor compares the report's scope with the system description, and any gap between the two turns into questions you answer under deadline during fieldwork.
When to test: SOC 2 Type 1 vs Type 2
Type 1
A Type 1 report covers control design as of one date. In the audits we run, auditors normally accept a pentest from the 12 months before the report date, as long as the critical and high findings are fixed and retested before that date. A recent test also keeps answering buyer questionnaires for longer after the report is issued.
Type 2
A Type 2 report covers how controls operated across an observation window of 3 to 12 months. For the pentest to count as evidence for that period, run it inside the window, early enough to fix and retest before the window closes. Two to three months before the end of the window is the latest start that leaves room for both.
Here is how that looks for a first Type 2 with a 6-month window:
| When | What happens |
|---|---|
| Before the window opens | Readiness done, scope agreed, pentest booked |
| Months 1 to 2 | Pentest runs; each finding goes into your ticketing system with an owner and a deadline |
| Months 2 to 4 | Fixes ship; the retest confirms they hold |
| Months 4 to 6 | Scans, access reviews and the other controls keep running, with no open critical or high findings |
| After the window closes | Fieldwork, usually 4 to 8 weeks, then the report |
The SOC 2 Type 2 guide has the full evidence cadence for the window.
Year two and bridge letters
After the first report, most companies move to a rolling 12-month window, and the annual pentest lands inside each one. If a buyer asks for coverage between two reports, a recent pentest report is part of the supporting evidence that turns a SOC 2 bridge letter into an accepted package.
How often to run a SOC 2 penetration test
Once a year, and after any significant change: a new product line, a rebuilt authentication flow, a move to another cloud provider or a major API release. The annual cadence matches what buyers and most auditors expect, and the after-change trigger mirrors what CC7.1 asks for scanning. The best time for a pentest covers timing around releases in more detail.
What your auditor needs from the pentest report
An auditor reads the report as evidence, not as a security document. These parts decide whether it counts:
| What the auditor checks | Why it matters |
|---|---|
| Scope and dates | They must match the system description and fall inside the audit period; a system or a date outside them is not evidence |
| Tester independence and qualifications | An outside firm, or an internal team independent of the engineers who built the system, with named certifications such as OSCP |
| Methodology | A recognized standard such as the OWASP Web Security Testing Guide, NIST SP 800-115 or PTES shows a repeatable process, not an ad hoc scan |
| Findings with risk ratings and evidence | Each finding can be traced to a ticket and an owner |
| Remediation and retest results | They show the critical and high findings were fixed and the fixes held |
| Executive summary | The part you share with buyers under NDA |
Retest evidence is the part most often missing. A report with open critical findings and no retest tells the auditor you found problems and stopped there. Both methodologies in the table are public: the OWASP Web Security Testing Guide and NIST SP 800-115. Our guide to the pentest report walks through each section.
What a SOC 2 penetration test costs
Most startups pay $4,000 to $8,000 for a SOC 2 penetration test of one web app and its API. Broader scopes that add cloud and network testing run $9,000 to $12,000. The price follows the number of user roles, API endpoints and environments in scope, and whether retests are included or billed as a new test.
| Scope | Typical cost | Notes |
|---|---|---|
| One web app and its API, grey box | $4,000 to $8,000 | Most startups' first SOC 2 report |
| Web app, API, cloud and network | $9,000 to $12,000 | Multi-product or compliance-heavy scopes |
| Retests of fixed findings | Included by some firms, billed as a new test by others | Ask before you sign; your auditor needs the retest result |
These ranges come from the same data as our SOC 2 certification cost breakdown, which shows two real first-year budgets with the pentest line included ($5,000 and $7,000), and our guide to penetration testing cost for startups, which explains what moves the price.
Frequently asked questions
Is penetration testing required for SOC 2?
No, the AICPA Trust Services Criteria do not require a penetration test, and penetration testing appears only as an example in the CC4.1 points of focus. It becomes required in practice when your own control set lists it, a customer contract demands it, or your auditor expects it.
How often should you run a penetration test for SOC 2?
Run one at least once a year and after any significant change, such as a new product, a rebuilt authentication flow or a cloud migration. For a Type 2 report, the annual test should fall inside each observation window.
When should the pentest happen for a SOC 2 audit?
For Type 1, auditors normally accept a test from the 12 months before the report date. For Type 2, run it inside the observation window, no later than two to three months before it closes, so you can fix and retest in time.
Can our own team do the SOC 2 penetration test?
Yes, if the testers are qualified and independent of the engineers who built the system. Auditors and enterprise buyers still give more weight to a third-party report, so most startups use an outside firm.
Will vulnerability scanning replace a penetration test for SOC 2?
No, they answer different criteria: scanning is the closer requirement under CC7.1, while a pentest is one of the separate evaluations under CC4.1. Most auditors expect both, and buyers rarely accept a scan report in place of a pentest summary.
Does an automated or AI pentest count for SOC 2?
In our experience, auditors accept automated and AI-driven pentest reports when they show scope, method, findings and retest evidence. A manual test still finds the business logic and cross-tenant flaws that enterprise security reviewers ask about.
What type of penetration test does SOC 2 need?
SOC 2 does not name a test type. For most startups the right test is a grey box test of the production web app and its APIs, plus a review of the production cloud account.
How much does a SOC 2 penetration test cost?
Most startups pay $4,000 to $8,000 to test one web app and its API. Broad scopes that add cloud and network testing run $9,000 to $12,000.
How SecureLeap runs SOC 2 penetration tests
We offer AI Pentest and manual grey box plans scoped to your SOC 2 system description. AI Pentest and Manual Pentest Basic include the APIs that a web or SaaS app uses as part of that target. Standalone APIs are scoped by endpoints in Manual Pentest Advanced, after a scoping call. Manual plans are tested by certified testers (OSCP, eWPTX, CPENT, CompTIA PenTest+), include business logic abuse testing, and reports are formatted as SOC 2 and ISO 27001 evidence. Manual plans include unlimited retests for 60 days, and AI Pentest includes up to 2 retests within 60 days, and most engagements can start within 2 business days.
If we also run your SOC 2 program, the All-In package puts the pentest and the audit fee on one invoice (SOC 2 consulting). See our public starting prices: AI Pentest from USD 2,895 or EUR 2,895, Manual Pentest Basic from USD 4,000 or EUR 4,000, and Manual Pentest Advanced priced after a scoping call.
Sources
- AICPA: 2017 Trust Services Criteria, with revised points of focus (2022)
- Cloud Security Alliance: SOC 2 Reports and Penetration Tests, by Mike DeKock, CPA
- BARR Advisory: Vulnerability Scans vs. Penetration Tests
- OWASP Web Security Testing Guide
- NIST SP 800-115: Technical Guide to Information Security Testing and Assessment


