PCI DSS v4.x lets entities set their own frequencies and even design their own controls, but only if they can prove the decision with a documented targeted risk analysis, and many assessments now stall on exactly that evidence. This guide goes deeper than our overview of the standard on two linked mechanisms: the targeted risk analysis (TRA) for activity frequency under Requirement 12.3.1, and the customised approach under Requirement 12.3.2 and Appendices D and E. It explains which requirements depend on a TRA, what an assessor expects to see, when the customised approach is worth the effort, and how to run both as a repeatable programme. It is written for PCI compliance leads, risk managers, security engineers and internal assessors at merchants and service providers. Requirement numbers follow PCI DSS v4.0.1 (June 2024).
Key takeaways
- There are two different TRAs. Requirement 12.3.1 covers analyses that set how often a flexible activity is performed; Requirement 12.3.2 covers the analysis that supports each customised control. They share a name but answer different questions.
- Frequency TRAs are mandatory for nine named requirements. If you rely on 5.2.3.1, 5.3.2.1, 7.2.5.1, 8.6.3, 9.5.1.2.1, 10.4.2.1, 11.3.1.1, 11.6.1 (when not weekly) or 12.10.4.1, the frequency must come from a documented 12.3.1 analysis that is reviewed at least once every 12 months.
- The customised approach is an assessment route, not a shortcut. It needs a controls matrix, a 12.3.2 TRA approved by senior management, the entity's own effectiveness testing and ongoing monitoring, and it can only be validated in a Report on Compliance by a QSA or ISA.
- Some requirements cannot be customised. Requirements with no Customized Approach Objective, such as 3.3.1 and its sub-requirements, 3.3.2, 11.3.2 and all of Appendix A2, must be met as written.
- Compensating controls and the customised approach do not mix. Compensating controls belong to the defined approach; Appendix D states they are not an option for a customised requirement.
Two mechanisms that share one name
PCI DSS v4.x introduced "targeted risk analysis" as a deliberately narrow tool. The guidance for 12.3.1 contrasts it with a traditional enterprise-wide risk assessment: it looks at one requirement at a time and asks whether the chosen frequency or process is adequate for that entity. An enterprise-wide assessment using a method such as ISO 27005 or NIST SP 800-30 is recommended in the Good Practice text, but it is not required and it does not replace the targeted analyses. See the NIST SP 800-30 Risk Assessment Guide.
| Aspect | Frequency TRA (12.3.1) | Customised approach TRA (12.3.2) |
|---|---|---|
| Question answered | How often should this flexible activity run, and why is that enough? | Does this entity-designed control meet the requirement's Customized Approach Objective with acceptable residual risk? |
| Who must do it | Any entity using one of the requirements that call for a 12.3.1 analysis | Only entities using the customised approach, once per customised requirement |
| Minimum content | Assets, threats, likelihood and impact factors, and a justified resulting frequency or process | Every element in Appendix D, at minimum a controls matrix and a risk analysis, using the information set out in the PCI SSC sample templates |
| Approval | Supported by the entity's security policy; no explicit senior management sign-off in 12.3.1 | Documented evidence approved by senior management |
| Review cycle | At least once every 12 months, with an updated analysis when the review shows it is needed | Analysis performed at least once every 12 months |
| Validation route | Any assessment type, including SAQs where the requirement applies | ROC only, documented by a QSA or ISA |
| Status | Best practice until 31 March 2025, now required | Applies whenever the customised approach is used |
Requirements that depend on a frequency TRA
The standard's section on frequencies explains the principle: where a requirement says "periodically" instead of fixing a minimum, the entity defines the frequency, supports it with its security policy and a 12.3.1 analysis, and must be able to show the frequency is effective. The defined requirements below name the 12.3.1 analysis explicitly. Each one has its own testing procedure that asks the assessor to examine the TRA, so a missing analysis produces a finding against the base requirement as well as against 12.3.1.
| Requirement | Activity whose frequency or approach the TRA sets | Typical inputs for the analysis |
|---|---|---|
| 5.2.3.1 | Periodic evaluation of system components judged not at risk from malware | Platform types, threat intelligence on malware for those platforms, exposure to untrusted networks |
| 5.3.2.1 | Periodic malware scans, where used to meet 5.3.2 | Whether real-time scanning is also present, file system change rate, criticality |
| 7.2.5.1 | Review of access by application and system accounts and their privileges | Number of system accounts, privilege level, change rate of integrations |
| 8.6.3 | Change of passwords and passphrases for application and system accounts (plus on suspected or confirmed compromise) | Password length and complexity, where the secret is stored, who can read it, whether rotation is automated |
| 9.5.1.2.1 | Frequency and type of POI device inspections | Device location, staffed or unattended, public access, history of tampering in the estate |
| 10.4.2.1 | Log review for system components not covered by 10.4.1 | Component role, whether logs are already correlated automatically, volume |
| 11.3.1.1 | Addressing vulnerabilities ranked below high-risk and critical | Risk ranking from 6.3.1, chaining potential, exposure of the affected systems |
| 11.6.1 | Payment page change and tamper detection, when run less often than weekly | Page change rate, script inventory from 6.4.3, e-commerce volume |
| 12.10.4.1 | Training for incident response personnel | Team turnover, complexity of the plan, how often the plan changes |
The guidance for 6.3.3 also recommends, without requiring, a 12.3.1 analysis to document how quickly patches that are not critical or high are installed. Doing it costs little and gives the assessor a clear rationale. For the identity-heavy items, 7.2.5.1 and 8.6.3, see the PCI DSS Requirements for Service Accounts and NHIs Guide (8.6); for 11.6.1 see the PCI DSS for E-commerce Payment Page Scripts Guide.
Writing a frequency TRA an assessor will accept
Requirement 12.3.1 lists what each analysis must contain. In practice a two-page document per requirement, written to a common structure, is enough. PCI SSC publishes an optional sample template for TRA for activity frequency and an information supplement on TRA guidance in its document library.
- Assets: Name the specific system components and data the requirement protects. The guidance gives examples such as log files and credentials. "The CDE" is too broad; "the 14 Linux hosts in the payment API tier and their local logs" is not.
- Threats: State what the requirement protects against, for example malware, an undetected intruder or misuse of credentials. Tie each threat to the purpose text of the requirement.
- Likelihood and impact factors: Record the factors that raise or lower risk. The standard's own examples include exposure to untrusted networks, environment complexity, high staff turnover, criticality of the components and the volume and sensitivity of the data.
- Resulting analysis: State the chosen frequency or process and justify how it minimises the likelihood or impact of the threat. Show the comparison you made, for example why monthly is adequate and quarterly is not.
- Annual review: Review each TRA at least once every 12 months, record whether the result is still valid, and perform an updated analysis where it is not. The guidance also expects review when changes could affect the risk.
- Use a scoring scale but keep the narrative: A simple likelihood and impact scale helps consistency across nine or more analyses, but the assessor reads the justification, not the score. See the Risk Register Template for a compatible format.
- Link the TRA to evidence: Reference the tickets, scan reports or logs that show the activity actually happens at the stated frequency. A sound analysis with no execution evidence still fails the base requirement.
- Different frequencies for different populations: One requirement can carry several frequencies, for example unattended fuel-pump terminals inspected more often than staffed counter devices under 9.5.1.2.1. Document each population separately.
How the customised approach works
Section 8 of PCI DSS v4.0.1 describes two validation approaches. The defined approach implements each requirement as written and the assessor uses the published testing procedures. The customised approach focuses on each requirement's Customized Approach Objective: the entity implements a control that meets the objective without strictly following the requirement text, and because every implementation differs there are no published testing procedures. The standard describes it as intended for risk-mature entities with a robust risk management approach, for example a dedicated risk management function, and says customised controls are expected to meet or exceed the security of the defined requirement.
Two design points are often missed. First, the approaches can be mixed at a fine grain: an entity can use the defined approach for a requirement in one environment and the customised approach for the same requirement on different system components. Second, the objective is outcome language. For example, the objective for 8.3.6 is, in substance, that a guessed password cannot be verified by an online or offline brute-force attack; the objective for 10.4.1 is that suspicious or anomalous activity is identified quickly. A customised control must demonstrably achieve that outcome.
Entity obligations under Appendix D
- Controls matrix: Document and maintain evidence for each customised control, covering all information in the controls matrix template in the PCI SSC publication on sample templates to support the customised approach (Appendix E). The format is optional, the content is not.
- Targeted risk analysis: Perform and document a 12.3.2 analysis for each customised control, covering everything in the sample TRA template.
- Own testing: Test each customised control to prove effectiveness and record methods, scope, dates and results in the controls matrix.
- Monitoring: Monitor and keep evidence of each control's ongoing effectiveness, not only at assessment time.
- Hand-over: Provide the completed matrices, TRAs, testing evidence and effectiveness evidence to the assessor.
Assessor obligations
- Understand the control: Review the controls matrix, TRA and evidence and confirm all documentation requirements are met.
- Derive the tests: Design and document testing procedures appropriate to the specific implementation.
- Test against two outcomes: Confirm the control meets the objective and results in an "in place" finding.
- Stay independent: A QSA involved in designing or implementing a customised control must not derive tests for, assess or assist in assessing that control. If your QSA firm helped build it, a different assessor must test it.
Appendix D also says entities that complete a Self-Assessment Questionnaire are not eligible for the customised approach, though they may choose a QSA or ISA assessment documented in a ROC instead. Payment brands and acquirers may regulate its use, including whether an ISA may perform the assessment, so confirm with them before you design anything.
Choosing between defined, compensating and customised
| Situation | Best route | Why |
|---|---|---|
| You can meet the requirement as written | Defined approach | Published tests, lowest documentation cost, predictable outcome |
| A legitimate, documented technical or business constraint stops you meeting the requirement, temporarily or permanently | Defined approach with a compensating control (Appendices B and C) | Designed for constraints; reviewed and validated annually by the assessor |
| You meet the intent with a stronger or different design, for example continuous control monitoring instead of a periodic manual review | Customised approach | Rewards innovation, but needs mature risk management and extensive evidence |
| The requirement has no Customized Approach Objective | Defined approach only | The customised route is not available |
| You validate by SAQ | Defined approach, or move to a ROC | SAQ entities are not eligible for the customised approach |
- Small organisations: Use frequency TRAs well and stay on the defined approach. The customised approach rarely pays back without a risk function, control monitoring tooling and a ROC-level assessment budget.
- Large service providers: Pick a small number of requirements where your architecture genuinely differs from the defined text, such as authentication or logging designs built around automation, and run them as customised controls with dedicated owners. Use the defined approach elsewhere.
- Software built under the Software Security Framework: Appendix F of PCI DSS says software developed and maintained under the Secure SLC Standard may support the customised approach for Requirements 6.3 and 6.5 objectives, subject to listing and assessment currency conditions. See the PCI Secure Software Standard and Secure SLC Guide.
Where identity fits
Three of the nine frequency TRAs concern accounts and credentials. Requirement 7.2.5.1 (review of system account access) and 8.6.3 (rotation of system account passwords) are where machine identities most often drift, because integrations are added without an owner and secrets are rotated only when something breaks. A strong 8.6.3 analysis states where each secret is stored, who and what can read it, the password length and complexity, and whether rotation is automated through a vault. If rotation is automated and the secret is never human-readable, a longer interval may be justified; a hard-coded secret in a deployment script justifies a short one and a remediation plan. The same accounts are a common candidate for a customised control, for example replacing static passwords with short-lived, workload-issued credentials that meet the 8.6.3 objective that system account passwords cannot be used indefinitely. See the Secrets Management Guide.
Implementation: a phased programme
- Inventory flexible requirements (weeks 1 to 2): List every requirement in scope that calls for a 12.3.1 analysis, plus any "periodically" wording you rely on. Map each to a control owner.
- Agree a method (weeks 2 to 3): Write one TRA procedure covering the five 12.3.1 elements, a likelihood and impact scale, approval and the annual review trigger. Get it into the security policy so 12.3.1.a can be satisfied.
- Draft frequency TRAs (weeks 3 to 6): Start with 8.6.3, 10.4.2.1 and 11.3.1.1, which most often attract findings. Pilot one analysis with your assessor before writing the rest.
- Prove execution (ongoing): Configure tickets, scheduled jobs or GRC tasks at the TRA frequency so evidence accumulates automatically.
- Screen customised candidates (quarter 2): For each candidate, confirm the requirement has an objective, the acquirer or brand permits the approach, and your assessment is a ROC. Drop any that fail.
- Build the controls matrix and 12.3.2 TRA: Describe the control, map it to the objective, record your own test design and results, and obtain senior management approval.
- Run in parallel before cut-over: Keep the defined control operating while the customised one gathers effectiveness evidence for at least one full review cycle. This is the rollback: if the assessor cannot reach "in place", you still have a compliant defined control.
- Review annually and on change: Tie TRA review to the 12.5.2 scope confirmation and to change management so architectural changes trigger reanalysis.
Detection and measurement
- TRA coverage: Percentage of in-scope flexible requirements with a current, approved TRA. Target 100 percent before fieldwork.
- Review currency: Number of TRAs older than 12 months without a recorded review. Target zero.
- Execution adherence: Percentage of scheduled activities (reviews, scans, inspections, rotations) completed within the TRA-defined frequency over the last 12 months.
- Customised control effectiveness: For each customised control, the monitored indicator from the controls matrix, for example time to detect an anomalous log event against the 10.4.1 objective.
- Assessor rework: Number of TRAs or matrices returned by the assessor for missing elements. A falling trend shows the method is working.
- Trigger response: Days between a significant change and the related TRA update.
Framework mapping
| PCI DSS element | NIST CSF 2.0 | NIST SP 800-53 Rev 5 | ISO/IEC 27001:2022 |
|---|---|---|---|
| 12.3.1 frequency TRA method and content | GV.RM-06, ID.RA-05 | RA-3 Risk Assessment | Clause 6.1.2 information security risk assessment |
| Risk the entity is willing to accept | GV.RM-02 | PM-9 Risk Management Strategy | Clause 6.1.3 risk treatment |
| 12.3.2 customised control testing by the entity | ID.IM-01 | CA-2 Control Assessments | Clause 9.1 monitoring, measurement, analysis and evaluation |
| Ongoing effectiveness evidence for customised controls | ID.RA-06 | CA-7 Continuous Monitoring | Clause 9.1 |
Common failure patterns
- One enterprise risk register offered as the TRA: A generic register does not identify the assets, threats and justified frequency for each requirement, so 12.3.1 and the dependent requirements both fail.
- Frequency chosen first, analysis written afterwards: The justification simply restates the existing schedule. Assessors look for a comparison of options and factors that actually drove the decision.
- No annual review record: The analysis was sound in 2025 but nobody recorded a review since, which fails the 12-month element.
- Customised approach attempted on an SAQ: SAQ entities are not eligible; the work has to move to a ROC or back to the defined approach.
- Compensating control dressed as customised: A control that exists because the entity cannot meet a requirement is a compensating control. Appendix D excludes compensating controls from the customised approach.
- Assessor independence overlooked: The QSA firm that designed the customised control is then asked to test it, breaching the independence condition in Appendix D.
Practitioner checklist
- List every in-scope requirement that calls for a 12.3.1 analysis and assign an owner to each.
- Publish one TRA procedure in the security policy covering assets, threats, likelihood and impact factors, justified result and annual review.
- Write a separate TRA for each requirement and each distinct device or system population.
- Link every TRA to evidence that the activity runs at the stated frequency.
- Record a dated review of every TRA at least once every 12 months and after significant change.
- Before using the customised approach, confirm the requirement has an objective, the assessment is a ROC and the acquirer or brand allows it.
- For each customised control, complete the controls matrix and 12.3.2 TRA with the information in the PCI SSC sample templates, and obtain senior management approval.
- Test and monitor customised controls yourself and keep the defined control running until effectiveness is proven.
- Confirm the assessor who tests a customised control had no part in designing or implementing it.
Standards and references
- PCI Security Standards Council: Payment Card Industry Data Security Standard, Requirements and Testing Procedures, v4.0.1
- PCI Security Standards Council: PCI DSS v4.x Sample Templates to Support Customized Approach
- PCI Security Standards Council: Information Supplement, Targeted Risk Analysis Guidance
- PCI Security Standards Council: Sample Template, TRA for Activity Frequency
- NIST: SP 800-30 Rev. 1, Guide for Conducting Risk Assessments
- NIST: Cybersecurity Framework 2.0
- ISO: ISO/IEC 27005:2022 Guidance on managing information security risks
Related NHI Mgmt Group resources: PCI DSS 4.0 Guide · PCI DSS Scoping and Network Segmentation Guide · PCI DSS Readiness Checklist · PCI ISA and QSA Certification Guide · FAIR Model and Open FAIR Standard Guide