The digital euro pilot is not a merchant rollout

Europe’s digital euro has reached the point where a shop counter can finally enter the discussion. That does not mean merchants are being asked to install it. The latest milestone is a controlled payments exercise, and confusing that exercise with a rollout is the quickest way for an SME to buy certainty that does not yet exist.

On 14 July, the European Central Bank named 36 payment service providers for a pilot spanning the ECB and 19 national central banks. More than 50 providers had applied. Development begins in the third quarter of 2026; the operational phase is planned for the second half of 2027 and will run for 12 months.

The scale sounds commercial. The perimeter is not. The users will be Eurosystem staff and selected business end users. A participating merchant must have a contractual relationship with the ECB or a national central bank involved in the pilot, as well as an acquiring relationship with a pilot provider. This is closer to a supervised rehearsal than a new button that any web shop can enable.

That distinction gives ordinary merchants a useful planning window. The right task for 2026 is not integration. It is to understand whether the existing checkout, refund and accounting chain can accept one more payment rail without turning every exception into manual work.

What the pilot will actually test

The pilot uses a beta instrument designed to resemble a possible digital euro without carrying legal-tender status. Selected providers will play one or both of two roles. Distributing providers will onboard end users and support accounts or wallets. Acquiring providers will enable selected merchants to receive payments.

The test cases cover person-to-person payments, payments at physical points of sale, software-based points of sale, e-commerce and mobile commerce. Offline functionality is narrower than some headlines imply: the pilot documentation describes offline person-to-person transactions through NFC, while merchant payments are tested through connected acceptance flows.

Pilot question	What is inside the exercise	What is not decided
Can people pay?	Online and offline person-to-person journeys, plus in-store and online merchant payments	A public launch date
Can providers connect?	Onboarding, platform integration, testing and back-end certification	The final commercial interface offered by every bank or acquirer
Can merchants receive funds?	Selected merchants linked to central banks and pilot providers	Open enrolment for ordinary retailers
Can the system work offline?	NFC transfers between nearby devices for the offline beta	Universal offline acceptance at shop terminals
Can users trust the experience?	Usability, operational feedback and communications	A final decision to issue the digital euro

Source: ECB pilot announcement and pilot FAQs. Last verified 2026-08-07.

This controlled scope matters because a merchant payment is not complete when a terminal displays a green tick. The order must close, stock must move, the receipt must be issued, settlement must arrive, fees must be classified and the accounting entry must match. A refund three days later has to find the same transaction. A failed connection must not create two charges or one unpaid shipment.

Those are the ordinary seams the pilot is meant to expose. They are also the seams an SME can document without access to the pilot.

A draft rulebook is not an implementation spec

The ECB published version 0.91 of the draft digital euro scheme rulebook in July. It incorporates market feedback and provides a more concrete view of roles, messages and processes. Its own disclaimer is equally concrete: the draft is preliminary, non-binding and not intended as a basis for implementing systems, processes or policies.

That warning should shape procurement. A supplier can reasonably explain how its payment orchestration layer would add a future tender type. It cannot reasonably promise final digital euro compliance, certification or pricing from a document that may change before legislation and issuance.

The legal sequence is unfinished too. The Council agreed its negotiating position in December 2025, including a framework for provider compensation and merchant service charge caps. The European Parliament and Council still have to settle the legislation. The ECB would decide whether to issue only after that framework exists. Its current target is technical readiness for a possible first issuance during 2029, assuming the legislation is adopted.

For an SME, this creates three separate dates:

1. Now: document the payment estate and remove hard-coded assumptions.
2. Second half of 2027: observe what the limited pilot proves about acceptance and operations.
3. Potentially 2029: prepare for a real service only when legislation, issuance and provider offers are settled.

Collapsing those dates into one project is how a roadmap becomes shelfware.

Start with the acceptance map

Most payment inventories list providers and fees. A readiness map must follow the transaction further. Start at the moment a customer chooses a tender and end only when the books, order and cash position agree.

For each channel, record six elements:

• the terminal, checkout page or app that initiates the payment;
• the acquirer, gateway and contract responsible for acceptance;
• the identifier that links payment, order, invoice and customer service record;
• the settlement file or API used by finance;
• the paths for cancellation, partial refund, chargeback or disputed delivery;
• the behaviour when the network, terminal, gateway or accounting export fails.

Then run one ordinary order and one deliberately awkward order. Split the basket, cancel one line after payment, refund another through customer service and close the day. Count the spreadsheets, copied identifiers and unmatched records. A future payment method will amplify those gaps; it will not repair them.

The exercise also exposes architectural lock-in. If every tender type is embedded directly into order logic, adding a rail becomes a rewrite. If the checkout sends a standard payment request to an orchestration layer and receives a stable status, the operational surface is smaller. The goal is not to predict the digital euro API. It is to keep the business process from depending on one provider’s current response format.

Offline needs a failure policy, not a slogan

Offline digital euro payments are often presented as electronic cash. The privacy proposition is meaningful: the ECB says offline transaction details would be known only to payer and payee, with no internet connection required. The pilot, however, must still test secure elements, device funding and defunding, proximity transfers and later clearing.

For merchants, the useful question is not whether offline sounds resilient. It is what the business does when payment state and order state cannot be checked against the same network at the same moment.

An SME can write that policy now:

• Which values may be accepted during an outage?
• Can the same employee override a payment warning and release goods?
• What evidence is stored locally without collecting unnecessary personal data?
• How are duplicate, delayed or reversed records identified after reconnection?
• Who owns the queue of unmatched transactions at opening time the next day?

The pilot may answer technical questions about secure offline value. It will not decide each merchant’s tolerance for loss, delay or customer friction. That remains an operating decision.

Privacy still leaves merchant data work

The ECB’s privacy design aims to prevent the Eurosystem from identifying users from payment data. Offline payments are designed for cash-like privacy, while online data available to the Eurosystem would be pseudonymised. Payment providers would still process the information necessary to meet legal obligations.

None of this removes a merchant’s own data responsibilities. The shop still creates an order, receipt, delivery record and possibly a customer account. It may link a transaction to loyalty data or fraud controls. The payment rail can minimise what it exposes while the merchant quietly rebuilds a detailed profile around it.

Readiness therefore includes a separation test. Can staff resolve a payment problem without opening the customer’s entire marketing profile? Can finance reconcile settlement with a transaction reference rather than an email address? Can a guest purchase remain a guest purchase? A privacy-preserving rail should not become an excuse to retain more data elsewhere.

Ask providers questions they can answer

The next conversation with a bank, acquirer or commerce platform should be specific and modest. Ask whether new tender types can be added without replacing the terminal estate; whether the refund API is tender-neutral; whether settlement records preserve a stable merchant reference; and whether software point-of-sale, e-commerce and mobile channels share the same reconciliation model.

Do not ask for a final digital euro quote. Fee caps, provider compensation, acceptance duties, exemptions and technical requirements still depend on the legislative and scheme outcomes. A precise price now would be an assumption presented as a product.

Also avoid dedicated hardware bought only for the possibility of a 2029 payment method. The pilot includes software point-of-sale and mobile journeys precisely because acceptance may not map neatly onto today’s terminal replacement cycle. Equipment purchased before the interface settles can become the least reusable part of the stack.

One transaction before one project

The digital euro pilot is significant because payment providers, central banks and merchants will have to make a beta instrument behave like an ordinary payment. Its value lies in the operational evidence it will produce, not in the impression that every merchant should move now.

Take one sale from authorisation to settlement this week. Add a partial refund, a connection failure and a finance mismatch. If the team cannot explain who owns each state and which identifier survives the journey, fix that before naming a digital euro project.

By the time the 2027 pilot reports what worked, the SMEs best placed to respond will not be those with the earliest mock-up. They will be the ones whose checkout can add a payment method without losing the order behind it.