Most EMV Level 3 delays do not begin in the lab. They begin weeks earlier, when a team enters the processor queue with gaps it could have found on its own. Paying.co has completed 133 EMV Level 3 certifications, and the same preventable problems show up again and again.

This is not another explanation of what EMV certification is. It is a working assessment you can use to judge whether your project is ready, or to find out why a stalled one is stuck. Score yourself honestly as you go.

How to score

For each of the ten dimensions below, give yourself:

0 = not started or unknown
1 = partly in place
2 = ready and verified

The maximum is 20. Interpretation is at the end.

The ten readiness dimensions

1. Application and transaction-flow readiness

Ready when every transaction type you plan to certify has a complete, tested flow: sale, refund, void, reversal, and any pre-authorization or other flows in scope. Contact, contactless, and fallback paths are implemented and documented, and nothing depends on hardcoded test data.

2. Device, firmware, configuration, and kernel readiness

Ready when you know which kernel approvals apply to your exact device and firmware version, and you hold the documentation. Terminal configuration, including AIDs, public keys, risk parameters, floor limits, and cardholder verification settings, matches what the processor requires. The firmware version is frozen for testing, because a mid-test change can put earlier results in doubt.

3. Processor specification and test-plan alignment

Ready when you have the current processor specification and an agreed test plan that states which card brands, entry modes, and test cases are in scope. Questions about the spec have written answers, not verbal ones.

4. EMV tags, receipts, and transaction records

Ready when the required tags are captured and sent in the authorization message in the format the host expects, receipts carry every required field, and stored transaction records can reproduce what happened in each test.

5. Reversals, timeouts, and exception behavior

Ready when the application behaves correctly when things go wrong: host timeouts, reversals after a timeout, a chip decline after an online approval, communication failure, and card removal mid-transaction. These are tested, not assumed.

6. Debit, fallback, and cardholder-verification requirements

Ready when debit routing, PIN entry, mobile-wallet verification, and fallback rules follow the processor and brand requirements for your market. Debit is where many otherwise solid projects discover gaps late.

7. Logging and certification-evidence collection

Ready when every test run produces timestamped logs, complete tag data, host messages, and receipts, stored in a format the processor accepts. Evidence collection is designed before testing, not improvised during it.

8. Test-tool and environment availability

Ready when you have confirmed access to the required test tools, test cards, host test environment, and certificates for the whole test window. Teams without their own tools often lose weeks waiting for rental or scheduling.

9. Processor communication and issue ownership

Ready when there is a named contact at the processor, an agreed way to raise and track issues, and a named owner on your side for each open item. Projects stall when everyone assumes someone else is following up.

10. Retest planning, release control, and regression management

Ready when builds are numbered and controlled, there is a regression suite, and you have a plan for retesting after a defect fix. A fix that breaks a previously passed case is one of the most common causes of lost time.

Warning signs your project is not ready

✦ The processor specification is still in draft or changing.

✦ Firmware or configuration keeps changing between builds.

✦ Nobody clearly owns the relationship with the processor.

✦ The plan says logs will be collected once testing begins.

✦ The test cases have never been run internally.

✦ Nobody can say which kernel approvals apply to your device.

Issues that commonly appear only during Level 3 testing

✦ Tag format or length mismatches that the host rejects.

✦ Reversal behavior under host timeouts.

✦ Receipt content that differs from the requirement.

✦ Debit routing and cardholder-verification edge cases.

✦ Regressions introduced by a fix for a different defect.

These are rarely found in ordinary functional testing, which is why a pre-certification review pays off.

What you can validate before entering the processor queue

Much of the risk can be removed early. Run the test cases internally against a simulator or test host. Compare every host message with the specification field by field. Force the exception cases, such as timeouts and card removal, on purpose. Do a dry run of evidence packaging so you know the logs hold everything the processor will ask for. Each defect found here costs far less than the same defect found mid-queue.

When managed certification makes sense

Managed certification fits when your team can build the application but does not have certification experience, the test tools, or an established processor relationship. In that model your developers write the code, while a certification team runs the test cases and owns communication with the acquirer. Other teams prefer an embedded model, where specialists work inside their own process only where needed, or full ownership, where the certification partner builds and certifies the application. The right model depends on your internal skills and schedule pressure.

How external dependencies shape the schedule

Even a well-prepared team does not control the whole calendar. Processor queue times, brand test windows, test-tool availability, device vendor firmware releases, and retest cycles after defects all sit outside your sprint plan. Build the schedule as a range, name an owner for each dependency, and track them as carefully as your own tasks.

Your score

17 to 20: You are ready to enter the queue. Confirm the schedule and begin.

12 to 16: You are close. Fix the dimensions where you scored 0 or 1 before you start, because they will cost more time later.

0 to 11: You are not ready yet. Start with your lowest scores, and consider an outside review before spending test time.

Next step

If your score is lower than you hoped, or a project is already stalled, a short review can show where the time is going. Request a pre-certification review or discuss Paying.co's managed certification program: paying.co/emv-certification

EMV #EMVCertification #Level3Certification #PaymentEngineering #PointOfSale #Fintech