PCI DSS v4.0 Explained in Simple Words for Business Owners
Most business owners think PCI DSS is done the day they sign an SAQ or receive a certificate from a vendor. Then a card scheme forwards a breach notification, a forensic firm called a PFI (PCI Forensic Investigator) shows up, and the same owner discovers that the certificate they framed was worth exactly nothing because their card data was flowing through a system nobody had scoped.
PCI DSS v4.0 was written by people who have watched that scene play out hundreds of times. It is not new bureaucracy for its own sake. It is the card industry closing the exact gaps that let breaches happen while everyone was technically compliant. This is a plain-English walk through what changed, what it costs, and what your team actually has to do — from an auditor who signs the report.
What PCI DSS actually is, in one honest paragraph
PCI DSS stands for Payment Card Industry Data Security Standard. It is a private contractual standard owned by the PCI Security Standards Council, which is backed by Visa, Mastercard, American Express, Discover and JCB. It is not a law and no government enforces it. Your acquiring bank enforces it through your merchant agreement. If you store, process or transmit cardholder data — a card number (the PAN, or Primary Account Number), the cardholder name, the expiry, or the service code — you are in scope. In India that means almost every e-commerce merchant, payment aggregator, PA-PG entity regulated by the RBI, and most SaaS platforms that touch card rails.
Version 4.0 was published in March 2022. Version 3.2.1 was formally retired on 31 March 2024. A large batch of the tougher v4.0 requirements were future-dated and became mandatory on 31 March 2025. So if your last assessment was against 3.2.1, you are already behind, and the future-dated controls are no longer optional. In late 2024 the Council issued a minor revision, v4.0.1, which is now the operative version — clarifications, not new burden.
The four numbers that decide your entire compliance journey
Before anything else, you need to know your merchant level. This single fact decides whether you fill in a self-assessment questionnaire over a weekend or endure a three-month on-site audit by a QSA (Qualified Security Assessor). Levels are set per card brand by annual transaction volume, but Visa and Mastercard thresholds are the ones almost everyone follows.
| Level | Annual card transactions | What you must do | Who signs off |
|---|---|---|---|
| Level 1 | Over 6 million (or after any breach) | Full Report on Compliance (ROC) + quarterly ASV scan | QSA on-site, mandatory |
| Level 2 | 1 million to 6 million | SAQ or ROC + quarterly ASV scan | Internal Security Assessor or QSA |
| Level 3 | 20,000 to 1 million (e-commerce) | SAQ + quarterly ASV scan | Self-signed, QSA optional |
| Level 4 | Under 20,000 e-comm / under 1M total | SAQ + ASV scan if applicable | Self-signed |
Two traps here. First, a breach can vault you to Level 1 overnight regardless of volume — the schemes reserve that right. Second, if you are a Level 4 merchant using a compliant payment gateway, you still are not off the hook; you complete the right SAQ (usually SAQ A for fully outsourced e-commerce, or SAQ A-EP if your site touches the payment page). Picking the wrong SAQ is the single most common mistake I see, and it silently invalidates the whole exercise.
The twelve requirements, translated out of auditor-speak
PCI DSS has always had twelve core requirements grouped under six goals. The structure survived into v4.0. Here is what each one really asks of you, without the clause language.
| Requirement | In plain words |
|---|---|
| 1 & 2 | Put a real firewall between the internet and your card systems, and stop shipping kit with default passwords and vendor accounts still enabled. |
| 3 | Do not store card data you do not need. If you must store the PAN, render it unreadable — strong cryptography, and never store the CVV/CVV2 after authorisation. Ever. |
| 4 | Encrypt cardholder data whenever it crosses a public or untrusted network — TLS 1.2 minimum, and you must inventory where it flows. |
| 5 & 6 | Run anti-malware, patch known vulnerabilities, and build software securely with a change-control process you can evidence. |
| 7 & 8 | Give access on need-to-know only, one unique ID per person, and enforce multi-factor authentication (MFA) for all access into the cardholder data environment. |
| 9 | Physically lock down the places card data lives — server rooms, backup media, paper printouts. |
| 10 | Log everything, keep logs for at least a year, and actually review them — v4.0 pushes toward automated log review. |
| 11 | Scan and penetration-test regularly, and monitor for unauthorised change — including on your payment page. |
| 12 | Have a written security policy, run awareness training, manage your third parties, and keep an incident response plan you have tested. |
What v4.0 changed, and why each change exists
Do not treat v4.0 as a cosmetic re-number. Roughly 64 new requirements were added, most of them future-dated to 31 March 2025. Each one maps to a real-world attack the old standard let through. These are the ones that actually change your budget and your work.
MFA everywhere, not just remote admin
Under 3.2.1 you needed MFA only for administrative and remote access into the cardholder data environment (CDE). Requirement 8.4.2 now demands MFA for all access into the CDE, including internal users at a console. If your developers and support staff log in with a password alone, that is a finding. Budget for an identity platform if you do not have one.
The e-skimming clauses — 6.4.3 and 11.6.1
This is the headline change and the one most Indian e-commerce firms are unprepared for. Magecart-style attacks inject malicious JavaScript into a merchant's checkout page and quietly siphon card data as the customer types it, upstream of any gateway. Requirement 6.4.3 says you must inventory and authorise every script that runs on your payment page and confirm its integrity. Requirement 11.6.1 says you must deploy a change-and-tamper detection mechanism that alerts you when your payment page's HTTP headers or scripts are modified. A framed certificate does nothing against this; you need tooling. This alone catches a huge share of real-world card breaches, which is precisely why the Council made it mandatory.
Passwords, encryption and the death of one-size rules
Minimum password length moved from 7 to 12 characters (Requirement 8.3.6). Disk-level encryption on multi-user systems no longer counts as rendering the PAN unreadable (Requirement 3.5.1). Anti-phishing controls are now explicit. And targeted risk analysis, under Requirement 12.3.1, lets you set the frequency of certain activities based on documented risk rather than a fixed calendar — freedom that comes with the burden of writing and defending the analysis.
The customised approach — powerful and dangerous
v4.0 introduces two ways to meet most requirements. The defined approach follows the stated control as written. The customised approach lets a mature organisation meet the security objective using a different control of its own design, provided it documents a full targeted risk analysis and the QSA independently derives testing procedures. This is genuinely useful for cloud-native and container-heavy shops. It is also a trap for the immature: if you cannot produce rigorous risk documentation, do not go near it. Most first-time assessees should stick to the defined approach.
What compliance actually costs in India
Owners always ask for a single number and there isn't one — it depends on your level, your architecture, and how much remediation you need. But here are honest working ranges from the Indian market, separating the assessment fee from the far larger remediation and tooling spend that nobody warns you about.
| Cost element | Level 4 / small e-comm | Level 2 mid-market | Level 1 enterprise |
|---|---|---|---|
| QSA assessment / SAQ support | Rs 1.5L to 4L | Rs 6L to 12L | Rs 15L to 35L+ |
| Quarterly ASV scans (annual) | Rs 40k to 1L | Rs 1L to 2.5L | Rs 2.5L to 6L |
| Annual penetration testing | Rs 1L to 3L | Rs 3L to 8L | Rs 8L to 20L+ |
| Tooling (SIEM, MFA, e-skim detection, FIM) | Rs 2L to 5L | Rs 8L to 25L | Rs 30L to 1Cr+ |
| Remediation / staff effort | Highly variable | Usually the biggest line | Usually the biggest line |
The lesson buried in that table: the audit fee is rarely your main cost. Remediation and ongoing tooling are. A first-time Level 1 assessment for an Indian payment aggregator commonly runs three to six months end to end, and the RBI's own PA-PG framework requires valid PCI DSS compliance (a current Attestation of Compliance) as a precondition for authorisation — so for regulated fintechs this is not optional, it gates your licence.
What actually happens in the room
Let me give you a scene, because the abstract requirements never land otherwise. A mid-size Indian D2C brand, roughly Level 2, is confident. Their payment page is a hosted iframe from a well-known gateway, so they assumed SAQ A and a light touch. On day one I ask a single question: show me the source of your checkout page. It loads an analytics tag, a heat-mapping script, and a chat widget — three third-party JavaScript files running in the same DOM as the payment iframe.
That changes everything. Because those scripts share the page, the merchant is exposed to e-skimming and drops out of SAQ A into SAQ A-EP, with Requirements 6.4.3 and 11.6.1 fully in play. They had no script inventory, no integrity monitoring, and no idea who had added the chat widget. The finance director's face when the scope tripled is the moment PCI DSS stops being paperwork and becomes real. We spent the next six weeks building a script allow-list, deploying tamper detection, and removing two marketing tags nobody could justify. The certificate came later — the security came first, which is the entire point of v4.0.
The risks of getting it wrong — and they are not theoretical
Non-compliance is not a paperwork problem; it is a financial and legal one, and in India it now stacks with the Digital Personal Data Protection Act (DPDP Act, 2023).
- Card scheme fines: acquirers can be fined and they pass it straight to you — typically USD 5,000 to USD 100,000 per month until you are compliant.
- Forensic and reissuance costs after a breach: PFI investigation, card reissuance, and fraud liability can dwarf any fine.
- RBI consequences: for PA-PG entities, losing your PCI DSS status can jeopardise your authorisation to operate.
- DPDP Act exposure: card data is personal data. A breach can trigger separate penalties under DPDP of up to Rs 250 crore per instance, adjudicated by the Data Protection Board.
- Reputational and contractual damage: enterprise customers increasingly demand a current Attestation of Compliance before they sign.
A practical order of operations
If you are starting cold or moving off 3.2.1, do it in this sequence. Skipping straight to buying tools before you have scoped is how firms waste lakhs protecting systems that were never in scope.
- Scope first. Draw an accurate data-flow diagram of where the PAN enters, moves, is stored, and leaves. Everything that touches it, and anything connected to it, is your CDE.
- Segment ruthlessly. Use network segmentation to shrink the CDE. A smaller CDE means a cheaper, faster, less painful assessment — this is the single highest-leverage move.
- Pick the correct SAQ or confirm you need a ROC. Get this wrong and the rest is wasted.
- Run a gap assessment against v4.0.1 specifically, not 3.2.1. Flag every future-dated control now in force.
- Close the e-skimming gap. Build a payment-page script inventory and deploy change-and-tamper detection — 6.4.3 and 11.6.1 are non-negotiable.
- Enforce MFA into the entire CDE and lift minimum passwords to 12 characters.
- Stand up centralised logging with at least twelve months retention and evidence that logs are reviewed.
- Schedule ASV scans quarterly and penetration testing at least annually and after any significant change.
- Write and actually test your incident response plan — the assessor will ask when you last ran it.
- Manage your third parties: maintain a register with each provider's PCI responsibility matrix.
The benefits worth naming out loud
It is easy to treat PCI DSS as a cost centre. Done properly it is not. The segmentation work reduces your attack surface for every threat, not just card fraud. The logging and MFA you build satisfy DPDP, ISO 27001 and SOC 2 controls simultaneously, so one disciplined programme services several obligations. And a current Attestation of Compliance is a commercial asset — it shortens enterprise procurement and unlocks partnerships that gate on it. The firms that suffer are the ones treating it as an annual certificate hunt; the ones that thrive treat the twelve requirements as a baseline operating standard.
Where this leaves you
Come back to the owner with the framed certificate and the flowing card data. The gap between those two things — the paperwork and the reality — is exactly what v4.0 was written to close. MFA into the whole CDE, script integrity on your payment page, honest scoping: none of it is glamorous, all of it is the difference between being compliant and being secure. Aim for the second and the first follows.
At CyberSigma we are senior CERT-In empanelled auditors and PCI QSAs who do this hands-on — scoping, gap assessment, and the awkward payment-page questions — so you find the third-party script before an attacker does, not after. If you want a straight read on where you actually stand against v4.0.1, that is the conversation we are happy to have.
FAQs
Is PCI DSS legally mandatory in India?
Not by statute — it is a contractual obligation enforced through your acquiring bank and card scheme agreements. However, the RBI's PA-PG framework makes valid PCI DSS compliance (a current Attestation of Compliance) a precondition for payment aggregator authorisation, so for regulated fintechs it is effectively mandatory. Separately, card data is personal data under the DPDP Act, which does carry statutory penalties.
We use a hosted payment gateway. Are we automatically compliant?
No. Outsourcing payment processing reduces your scope but does not remove it. If your checkout page loads any of your own scripts alongside the payment element, you likely fall under SAQ A-EP and the e-skimming requirements 6.4.3 and 11.6.1 apply to you. You still complete an SAQ and sign an attestation.
What is the difference between v4.0 and v4.0.1?
v4.0 was the major rewrite published in 2022. v4.0.1, issued in 2024, is a limited revision that corrects errors and clarifies wording — it adds no new requirements. v4.0.1 is now the operative version you should be assessed against, and v3.2.1 is fully retired.
Which future-dated v4.0 requirements matter most now that they are live?
The ones that changed budgets: MFA for all access into the CDE (8.4.2), 12-character minimum passwords (8.3.6), payment-page script inventory and integrity (6.4.3), and change-and-tamper detection on the payment page (11.6.1). All became mandatory on 31 March 2025.
How long does a first-time Level 1 assessment take?
Realistically three to six months end to end. The QSA on-site portion is a fraction of that; most of the time goes into scoping, remediation, gathering a full year of evidence, and completing the required scans and penetration tests before the Report on Compliance can be signed.
Can we use the customised approach to avoid buying new tooling?
Only if you are mature enough to defend it. The customised approach lets you meet a requirement's objective with your own control, but it demands a rigorous targeted risk analysis and independent testing procedures derived by your QSA. For most first-time assessees the defined approach is faster, cheaper and safer.
Liked the post? Share on:




Leave A Comment