Compliance means you met a defined set of rules when someone checked. Security means your systems resist, detect and recover from real attacks every day, and a PCI DSS validation proves only the first. The Titanic is the oldest and clearest illustration, and it still explains why a clean report on compliance can sit next to a breach.

Key takeaways

  • Compliance is evidence that you met a written standard at a point in time, while security is the ongoing ability to stop and detect attacks.
  • PCI DSS is a baseline built from known attack patterns, so it trails new techniques by design.
  • Heartland and Target both held passing PCI results before their breaches, and neither result prevented the intrusion.
  • The biggest gaps are scope that is drawn too narrowly, controls that pass an audit but are never watched, and evidence gathered only before the assessor arrives.
  • PCI DSS 4.0.1 is the only active version and every requirement is now mandatory, so the practical goal is to run it as a floor and add continuous monitoring, risk-based testing and, where justified, the customized approach.

Why does the Titanic keep coming up in the compliance vs. security debate?

Because the Titanic met every safety rule that applied to it, and the rules were still wrong for the ship. When it sailed in 1912, a member of Parliament noted in the Commons that it sailed under regulations that had not been substantially altered since 1894. The lifeboat rules were tied to a ship's tonnage, and the table topped out at vessels of 10,000 tons and up. Titanic was more than four times that size.

The numbers show the gap. The same Hansard debate put the Board of Trade's requirement at lifeboat capacity for about 960 people, while Titanic carried 20 boats with room for 1,178. According to this overview of the ship's lifeboats, it complied with the regulations and even went past them by adding four collapsible boats. About 2,207 people were aboard, and 1,496 died. The inspectors who signed off had done their jobs, but the checklist described a different generation of ship.

Security writers have used this story for years. Anton Chuvakin pointed readers to a Guerilla CISO post making the same argument in January 2009, and the comparison has held up because the mechanism has not changed. A standard is written from past failures, approved through a committee process and applied to systems that have kept changing since it was drafted.

The analogy has limits. Nobody serious argues that lifeboats were pointless, and nobody should argue that PCI DSS controls are pointless either. The failure on the Titanic was treating the legal minimum as the safe amount, and that is the same mistake as treating a passing PCI report as proof of security.

What is the difference between compliance and security?

Compliance is meeting an external set of requirements and being able to prove it, and security is the measured ability of your systems and staff to prevent, detect and contain attacks. The two overlap heavily, since most PCI controls are sound security practice, but they answer different questions. Compliance asks whether you did what the standard says. Security asks whether an attacker can get in, and whether you would notice.

The practical differences show up in who sets the bar, how it is measured and what failure looks like.

ComplianceSecurity
Question it answersDid we meet the stated requirements?Can we withstand and spot real attacks?
Who sets the barA standard body, card brands, regulatorsYour own threat model and risk appetite
When it is measuredAt an assessment, usually annuallyContinuously
ScopeThe defined environment under reviewEverything an attacker can reach
Evidence of successA report, attestation or certificateAttacks blocked, detected and contained quickly
Typical failureFailed audit, fines, lost card acceptanceData loss, downtime, fraud

A business can be strong on one and weak on the other. A small merchant with a tidy, honest self-assessment but no one watching its logs is compliant and exposed. A well-defended company that skipped required paperwork is the reverse. The goal is to be both, and to treat compliance as the part you can be forced to prove to others.

Why can a PCI DSS compliant company still be breached?

PCI DSS validation is a point-in-time judgment about a defined slice of your environment, so it can be true while the rest of your risk is untouched. Three gaps account for most of the distance between a passing report and a safe business.

The first is timing. An assessment checks controls as they stand during the review window. Systems change the week after: a new vendor connection, a server built from an old image, a firewall rule opened for a project and never closed. The standard has long told entities to build PCI DSS into business-as-usual operations. The PCI SSC's own quick reference guide for the prior version says that approach lets an entity monitor its controls between assessments. Many organizations never do that and prepare for the assessor once a year.

The second is scope. PCI DSS applies to the cardholder data environment and to systems that can affect its security, and someone has to draw that boundary. If a database holding old card numbers, a flat network segment or a third-party remote access path is left out of scope, it is also left out of testing. Attackers do not respect the boundary. Large-organization assessments rely on sampling, and small merchants at lower validation levels can attest through a Self-Assessment Questionnaire, so the report is only as good as the scoping and honesty behind it. Finding the card data you forgot about is the first fix, and our guide to finding stored cardholder data covers how.

The third is the control that exists but is not watched. A firewall, an antivirus agent or a logging system can satisfy a requirement on paper while generating alerts that nobody reads. An audit confirms the tool is present and configured. It rarely confirms that a human acts on an alert at 2 a.m. on a holiday weekend, which is when attackers like to work.

Which companies were PCI compliant and breached anyway?

Two of the best-documented cases are Heartland Payment Systems and Target, and both had recent passing PCI results before intruders got in.

Heartland, a major US card processor, reported in its SEC filing that it received confirmation of PCI DSS compliance from a third-party assessor in April 2008. It then disclosed on January 20, 2009 that it had suffered an intrusion in its processing system. A Federal Reserve Bank of Philadelphia case study records that assessors repeatedly rated Heartland compliant without finding the SQL injection weakness that attackers used. Visa removed the company from its list of compliant service providers after the breach, and Heartland was revalidated in April 2009. Our Heartland and Hannaford breach explainer covers what card security changed afterward.

Target followed in 2013. The company told Congress it passed its annual PCI audit in September 2013. Attackers then installed malicious software on point-of-sale devices, and Target said the data was taken between November 27 and December 15, 2013. About 40 million card accounts were compromised, with the entry point traced to credentials of a heating and air-conditioning contractor. Banks later sued Target's security vendor, Trustwave, alleging it should have found the weaknesses. Those are allegations, and the dispute shows the other side of the problem: after a breach, people argue over whether the company was ever truly compliant.

CompanyPassing result before the breachWhat happenedWhat it showed
Heartland Payment SystemsThird-party assessor confirmation, April 2008Intrusion disclosed January 2009; removed from Visa's compliant list, restored April 2009A validated environment can hide a basic web flaw for months
TargetPassed annual audit, September 2013Point-of-sale malware stole data in late 2013Third-party access and unwatched alerts can defeat a passing audit

Both breaches surfaced the way most do, through outside parties. Card networks and issuers noticed fraud patterns that pointed back to a single merchant, a process explained in our piece on common point of purchase analysis. Neither company's internal controls caught the intrusion first, which is the clearest sign that compliance was being measured and security was not.

What has PCI DSS changed to close the gap?

PCI DSS 4.0 pushed the standard toward continuous operation and risk analysis, and every one of its requirements is now mandatory. Version 4.0.1 is the only active version. The PCI Security Standards Council described it as a limited revision with no new or deleted requirements and no change to the March 31, 2025 effective date for the new ones.

DateMilestone
March 2022PCI DSS v4.0 published
March 31, 2024v3.2.1 retired
June 2024v4.0.1 published as a limited revision
December 31, 2024v4.0 retired; v4.0.1 is the sole active version
March 31, 2025The 51 future-dated requirements became mandatory

Version 4.0 added 64 new requirements, 13 of which applied immediately and 51 of which were future-dated best practices until March 31, 2025. Several target the exact weaknesses that old breaches exposed. Multi-factor authentication (requirement 8.4.2) now covers all access into the cardholder data environment, not only remote or administrative logins. Passwords need at least 12 characters where passwords are used. Requirements 6.4.3 and 11.6.1 address scripts on payment pages, aimed at web skimming. Requirement 12.3.1 requires targeted risk analyses wherever the standard lets you choose how often to perform a task.

Even so, PCI DSS is still validated through an assessment or questionnaire at a point in time, and the council cannot watch your network for you. The standard now asks for more evidence of ongoing operation, but evidence of operation is not the same as someone acting on what the tools report. That part remains your job.

How do you treat PCI DSS 4.0 as a minimum instead of a goal?

You build a security program that would exist even if no one audited you, then let the assessment confirm its compliance portion. These steps work for most organizations that handle card data.

  1. Map actual card data flows. Scan for stored card numbers, trace each path from checkout to settlement, and include vendors, remote access and backups. Confirm the scope in writing at least once a year, and again after any major change.
  2. Shrink the scope. Hosted payment pages, tokenization and validated point-to-point encryption keep card data off your systems, which removes whole categories of risk rather than documenting controls around them.
  3. Monitor continuously. Centralize logs, alert on changes to critical files and payment-page scripts, and make sure a named person or managed service responds to alerts around the clock. Test that response with a drill, not only a screenshot.
  4. Test based on risk. Run vulnerability scans on the schedule PCI requires, but add penetration tests built around how an attacker would actually reach your card data, including through suppliers. Retest after significant changes.
  5. Collect evidence monthly. Controls performed right before an assessment do not demonstrate continuous compliance. Keep a calendar of recurring tasks and record who did them and when.
  6. Measure outcomes. Track time to detect and time to contain incidents, patch age on exposed systems, and how many alerts go unreviewed. These measure security in a way an attestation does not.

Risk-based testing is also where the standard now helps you. Requirement 12.3.1 asks for a documented targeted risk analysis whenever a control's frequency is left to you. Done honestly, that analysis forces a decision about how often a system holding card data really needs review, rather than defaulting to the minimum interval.

When does the customized approach make sense?

The customized approach makes sense when an existing technology or design meets a requirement's security objective better than the prescribed method does, and you can prove it. PCI DSS 4.0 introduced it as an alternative to the traditional defined approach. You document the control, perform a targeted risk analysis for it (requirement 12.3.2), and your assessor tests that it achieves the stated objective.

A typical case is a company with modern cloud tooling that cannot do a control the old way, such as a legacy scheduled check, but that gets the same or better protection through automated detection. The approach lets the security design lead and the audit follow. It is not a shortcut. Assessors apply more scrutiny to customized controls, and you must keep the design documentation, monitoring criteria and evidence ready.

It also has a hard boundary for small businesses. The approach is not supported in SAQs, so it is for entities that validate through a Report on Compliance, which in practice means larger merchants and service providers. If you fill out an SAQ, you follow the defined approach and get your flexibility from reducing scope.

Treat it as a way to move beyond the checklist, not around it. If you cannot show with data that your alternative meets the objective, the defined approach is the safer path, and you can still add extra controls on top.

What should a small business do differently?

A small business should shrink what it has to protect, pick payment tools that keep card data off its systems, and confirm that someone is watching what remains. Most small merchants validate with an SAQ, which is an honest self-description rather than an outside test, so the temptation to answer to the minimum is strong. The risk is the same as at Heartland and Target, just at smaller scale.

Start with your payment setup. A processor-hosted checkout or a validated encrypted terminal shifts most PCI obligations to the provider and limits what an intruder can steal from you. Ask your acquirer which SAQ type applies and why, and do not choose a lighter form because it is easier. Then confirm basics that apply even when the questionnaire is short: multi-factor authentication on every account that can touch payments, prompt patching of the point-of-sale system and router, and unique credentials for each vendor and employee.

Last, decide who reviews alerts and what happens when one fires. For a small shop that may be a managed security provider or the processor's monitoring service. The cost is usually modest compared with a forensic investigation, card reissuance charges and weeks of disrupted sales after a breach.

Frequently asked questions

Does being PCI compliant protect me from liability after a breach?

No. Validation is not immunity. Heartland was removed from Visa's list of compliant service providers after its breach, and the company later agreed to pay up to $60 million to fund a settlement program for Visa issuers. Card brands and acquirers can assess costs after an incident regardless of earlier results.

Is PCI DSS 4.0 still the current version?

The current version is 4.0.1. It replaced v4.0 on December 31, 2024, and added no new requirements. All the former future-dated requirements became mandatory on March 31, 2025, so there is no best-practice grace period left.

How often does PCI DSS compliance have to be validated?

Validation is annual, through an assessment or questionnaire depending on your merchant level and your acquirer's rules. Beyond that, quarterly external vulnerability scans by an approved scanning vendor are required, and many other controls have their own recurring tasks. Between those checkpoints, the standard expects the controls to keep operating.

Can a small merchant use the customized approach?

Generally not. The customized approach is not supported in SAQs, so it applies to organizations that validate with a Report on Compliance. SAQ filers follow the defined approach and gain flexibility mainly by reducing scope through hosted checkout, tokenization or validated encryption.

Is PCI DSS enough to secure a business?

No. It is a baseline for protecting card data, and it does not cover every system, threat or business risk. Pair it with a risk-based program such as one built on the NIST Cybersecurity Framework, plus monitoring, incident response and regular testing.