Expedition. Free, on-demand.

Technical tracks for practitioners, outcomes for leaders.

PCI DSS Compliance: Shared Responsibility Explained

A platform provider’s PCI DSS status is only part of the compliance picture. See how to trace cardholder data through your environment, understand where provider responsibility ends and keep your own controls aligned as workloads change.

PCI DSS COMPLIANCE DEFINED

The process of meeting the Payment Card Industry Data Security Standard requirements that apply to an organization’s environment, including understanding which controls are handled by service providers and which remain the organization’s responsibility, then validating that those requirements are being met.

A company moves payment data from its transaction systems into a data environment for fraud analysis. The platform it uses has its own Payment Card Industry Data Security Standard (PCI DSS) attestation. So is that analytics workload covered by the platform?

That question gets to a distinction that’s easy to lose in data security compliance discussions. A service provider’s PCI DSS attestation covers the controls and services included in that provider’s assessment — it doesn’t automatically establish compliance for every workload a customer builds on top of them. PCI SSC explicitly requires customers and third-party service providers to understand where various responsibilities sit, including which PCI DSS requirements the provider handles and which remain with the customer.

For data and security teams, PCI DSS compliance starts with two views of the environment: where cardholder data flows and who owns each applicable control. Together, those views define the systems that need protection, the platform capabilities your organization needs, and the configuration, monitoring and operational responsibilities your teams own.

What is PCI DSS compliance?

PCI DSS is a baseline of technical and operational requirements for protecting payment account data. It applies to entities that store, process or transmit cardholder data (CHD) or sensitive authentication data (SAD), as well as entities whose systems or services could affect the security of the cardholder data environment (CDE). The current version is PCI DSS v4.0.1, published in June 2024.

The standard, developed by PCI Security Standards Council, covers areas including network security, protection of account data, vulnerability management, access control, monitoring and security policy. Organizations also have to validate compliance through the method required for their particular environment and compliance program. Depending on the organization and circumstances, that evidence might take the form of a Self-Assessment Questionnaire (SAQ) or a Qualified Security Assessor-led Report on Compliance (ROC). Payment brands, acquirers and other compliance-accepting entities determine the applicable validation and reporting requirements.

Unlike statutory requirements such as GDPR or HIPAA, PCI DSS is an industry standard enforced through payment-brand and acquiring-bank relationships rather than government law.

It’s important to note that outsourcing infrastructure or payment-related services doesn’t outsource the customer’s responsibility to understand how cardholder data is protected. Even a merchant that outsources all payment processing remains responsible for confirming its provider’s applicable PCI DSS compliance, documenting shared responsibilities and monitoring the provider’s compliance status.

See how Snowflake Trust Center helps teams identify and monitor security risks, review account configurations and address findings within Snowflake’s shared-responsibility model:

What counts as cardholder data — and where does it sit in your data stack?

To determine PCI DSS scope, teams first identify where cardholder data is stored, processed or transmitted, then trace the systems, people and processes involved in handling or protecting it. At a minimum, cardholder data consists of the full primary account number (PAN). When present with the PAN, it also includes cardholder name, expiration date and service code. PAN determines whether PCI DSS requirements apply. Sensitive authentication data includes information used to authenticate a transaction, such as card verification codes and PIN data — PCI DSS prohibits storing this data after authorization, even in encrypted form.

From there, the scoping exercise follows that data through the environment. The CDE includes the people, processes and system components that store, process or transmit CHD or SAD, along with components that have unrestricted connectivity to those systems. An authentication service, for example, might fall into scope even though PAN never resides there.

Any system that stores, processes or transmits cardholder data — along with systems that can affect the security of that environment — needs to be considered during PCI DSS scoping. This includes downstream analytics and reporting environments when they receive cardholder data, whether through a production pipeline, a copied table, a data science workflow or a local extract.

How to reduce PAN exposure in your data environment

Once teams know where primary account numbers move through the environment, the next question is whether every downstream system actually needs them. Keeping PAN out of analytics, reporting and data science workflows where it isn’t required reduces unnecessary exposure and, depending on the architecture, can also reduce PCI DSS scope.

Tokenize PAN before downstream use

Tokenization replaces PAN with a non-sensitive substitute value before the data reaches systems that don’t need the original account number. The mechanism that allows the token to be associated with or exchanged for PAN remains protected within the tokenization solution, while downstream applications work with the token instead.

Implemented appropriately, tokenization can reduce the number of systems exposed to PAN and therefore the number that need to be evaluated as part of the cardholder data environment. The tokenization system itself, along with systems that can reverse or otherwise affect the protection of the tokenized data, still needs to be considered during scoping.

Use masking to limit who sees sensitive values

Dynamic masking serves a different purpose. Rather than removing PAN from the underlying data, it controls what particular users or roles see when they query it.

Masking is useful when some users need access to a table but don’t need to see the full account number. Because the PAN still exists in the underlying system, however, masking alone doesn’t remove that system from PCI DSS scope.

Segment systems that handle cardholder data

Segmentation limits how far the cardholder data environment extends by separating systems that handle PAN from systems that don’t need access to it. Network boundaries, access controls and other isolation mechanisms can help prevent unrelated workloads from becoming part of the same PCI DSS environment.

However, the effectiveness of that separation must be validated. If a supposedly separate system can connect to, administer or otherwise affect the security of the CDE, it may still need to be considered in scope.

Classify sensitive data so controls reach the right places

Tokenization, masking and segmentation depend on knowing where sensitive data resides in the first place. Data classification helps teams identify and tag PAN and other regulated data so the appropriate controls can be applied consistently as data moves into new tables, pipelines and downstream workloads.

Classification is an especially important data security control as the environment changes. A newly created analytics table or copied data set can introduce another instance of PAN even when the original source remains properly protected, so classification and monitoring need to extend beyond the initial scoping exercise.

QUICK TIP

Before sending payment data into an analytics or reporting workflow, check whether the workload actually needs the full PAN. Tokenizing or otherwise removing PAN upstream can reduce unnecessary exposure and, depending on the architecture, may reduce PCI DSS scope.

The shared responsibility model: What your platform provides vs. what you own

A platform provider’s PCI DSS compliance documentation shows which services and controls were assessed. Your team still owns the responsibilities tied to how you configure the platform, whom you give access to and how you operate the workload.

To understand that split, review the provider’s Attestation of Compliance (AoC) and responsibility matrix. The AoC documents the provider’s assessment, while the responsibility matrix shows which PCI DSS requirements the provider handles, which remain with your organization and which are shared.

In practice, that split shows up in how platform capabilities are configured and operated. The platform might supply encryption, access-control mechanisms, network protections and audit logging. Your team decides which roles receive access, where masking or row-level policies are applied, how logs are reviewed and what happens when the environment changes.

For each applicable control, the responsibility matrix should help answer three questions: What does the provider handle? What do we still own? What evidence shows that both sides are operating as intended?

When a PCI DSS control gap exposes cardholder data, the issue moves beyond compliance and into incident. Responding to a data breach is another responsibility that stays with your organization. PCI DSS requires a documented and tested response plan for suspected or confirmed compromise of cardholder data, including defined roles and escalation procedures.

COMMON PITFALL

Don’t treat a provider’s PCI DSS documentation as evidence that every workload you run on the platform is compliant. Use the responsibility matrix to identify which controls the provider handles and which still depend on your configuration and processes.

What to look for in a data platform for a PCI DSS program

When evaluating a platform that will store or process cardholder data, map the capabilities you’re considering back to the controls and evidence your PCI DSS program requires. The feature name is only the beginning — configuration and operating evidence determine how it fits into the program.

Encryption and key management

Stored PAN has to be protected according to applicable PCI DSS requirements, while data moving between systems requires appropriate protection in transit. Evaluate what encryption the platform applies by default, how cryptographic keys are managed and whether your architecture requires customer-controlled key material.

Access control at the appropriate grain

PCI DSS access controls should reflect legitimate business need. Role-based access control (RBAC), row-level controls and centrally managed policies help limit access to the specific cardholder data each user needs. Reusable policies also make those decisions easier to inspect than permissions scattered across individual views and applications.

Protection of sensitive values in downstream use

Analytics, business intelligence (BI) and AI workloads frequently need attributes from a payment transaction without needing the underlying PAN. Tokenization, masking and related controls help limit how much sensitive information those users and workloads receive.

When assessing masking in particular, look at where the policy is enforced. A centrally applied policy attached to the data object provides a different control model from manually creating a separate sanitized view for every downstream use.

Network and authentication controls

Private connectivity, network policies and strong authentication controls help limit how users and systems access cardholder-data workloads. That gives teams a clearer boundary around the systems handling cardholder data and reduces unnecessary exposure through broadly accessible endpoints or overly permissive access.

Queryable audit evidence

For data workloads, “who accessed what” often requires query-level evidence rather than infrastructure logs alone. Teams should be able to determine which user or role queried a sensitive object, when the access occurred and what policies applied. That evidence must also be retained and reviewed in a form that supports the organization’s assessment process.

The provider’s compliance evidence

Ask for the provider’s applicable AoC and documentation of the responsibility split. PCI SSC specifically expects assessed third-party service providers to provide sufficient evidence that the services used by a customer were covered by the assessment, and Requirement 12.9 addresses communication of provider and customer responsibilities.

Risks when cardholder data reaches your analytics stack

PCI DSS scope can expand through ordinary data work: a team copies raw transaction records into a new table, an analyst exports data for ad hoc analysis or a pipeline retains fields that no longer serve a business need. Once cardholder data leaves the original payment workflow, those downstream copies and access paths need the same scoping discipline as the source.

Raw PAN in downstream analytics

If analytics, reporting or data science workflows receive a full PAN where a token or less-sensitive value would serve the same purpose, the organization has unnecessarily increased the number of systems that need to be evaluated and protected under PCI DSS.

This is especially easy to miss when the data arrives through a copied table or pipeline rather than a payment application. The technical path is different, but the scoping question is the same: does the system store, process or transmit cardholder data, or affect the security of the environment that does?

Data retained longer than needed

PCI DSS requires organizations to limit stored cardholder data to what is necessary for legal, regulatory or business purposes, then securely delete or render it unrecoverable when that need ends.

For analytics environments, retention policies should account for copied tables, historical extracts and derived data sets as well as the original source. A retention rule applied only to the production system leaves downstream copies behind.

Exports outside the governed environment

A CSV downloaded for ad hoc analysis, a locally cached extract or a file moved into another tool creates another copy of cardholder data to secure, monitor and eventually dispose of.

Where exports are necessary, teams need to account for who can create them, where they can be stored and how long they remain available. Otherwise, those copies can sit outside the controls applied to the primary data environment. 

Data loss prevention (DLP) controls can help detect or restrict sensitive data moving into unauthorized destinations, including local files and unmanaged applications.

Configuration drift

Controls that are adequate today may no longer cover the environment after new tables, roles, pipelines or integrations are added. A masking policy might protect the original column while a copied table remains unprotected, for example, or a new role might receive broader access than intended. As data flows, integrations and supporting infrastructure change, teams must revisit PCI DSS scope and control coverage.

Gaps between assessment cycles

An annual assessment provides evidence at a point in time, while the data environment continues to change between reviews. Ongoing classification, policy monitoring and audit history help teams identify new copies of sensitive data, access changes and protection gaps before the next formal assessment.

How Snowflake can support PCI DSS workloads

Snowflake customers can request Snowflake’s Attestation of Compliance and PCI Shared Responsibility Matrix to see which requirements are covered by Snowflake’s assessment and which depend on the customer’s own configuration and processes.

Within the Snowflake environment, security and governance capabilities help teams apply and monitor the controls they’re responsible for. The Snowflake Trust Center evaluates Snowflake accounts for potential security risks and surfaces findings that teams can investigate and remediate. Its Data Security experience also uses sensitive-data classification to identify regulated data, including PCI data, show where it resides and surface whether masking policies are protecting it. 

How to keep PCI DSS scope current as your environment changes

PCI DSS compliance has to keep pace with changes in the environment: where cardholder data is copied, which systems can reach it, who has access and how responsibility is divided between the platform provider and your team. A responsibility matrix and an initial scoping exercise give you a baseline, but they need to stay aligned with the environment as new pipelines, tables, roles and integrations are introduced.

For data and security teams, the practical goal is to keep that picture current. If you can trace where cardholder data moves, identify who owns each applicable control and show that those controls are still operating as intended, you’re in a much stronger position when the next architecture change — or assessment — comes around.

KEY TAKEAWAY

A provider’s PCI DSS assessment covers the services and controls included in its scope. Your organization still needs to understand where cardholder data flows, which controls remain your team’s responsibility and how those controls are configured and monitored.

Frequently Asked Questions

Your common questions about PCI DSS compliance, answered by Snowflake experts.

Organizations that store, process or transmit cardholder data — and certain organizations whose services could affect the security of that data — fall within the intended scope of PCI DSS. The specific validation method varies according to the organization and the requirements of its payment brand, acquirer or other compliance-accepting entity.

Organizations that fail to maintain compliance may face monthly fines from card brands, higher transaction processing fees, and mandatory on-site audits. In severe cases or following a cardholder data breach, payment processors can revoke the organization’s ability to accept credit cards, alongside steep liabilities for forensic investigations, fraudulent charges, and replacement card costs.

No. A provider’s PCI DSS assessment covers the applicable services and controls included within that provider’s assessment. Customers remain responsible for their own applicable requirements, including understanding shared responsibilities, configuring their environment appropriately and validating compliance through the method required by their compliance program.

Explore Data Security Resources

Explore Data Security Topics

Deep dives into every aspect of data security