Digital product passport: the hard part isn't the QR code, but reading it in the factory.

The European ecodesign framework requires a growing share of products placed on the market to carry an identifier that links to a structured set of information: composition, material origin, reparability, and end-of-life. Companies that have begun preparing quickly discovered that the conceptual part—deciding which data to publish—is the least costly. The bottleneck is physical and lies in production: marking each individual unit with a unique code, verifying that the code is readable, and associating it with the correct record requires an industrial camera for code reading , installed at the correct point on the line, and a logic that holds up even when something goes wrong.

What does a digital product passport contain?

The content varies by product group, as it is defined on a case-by-case basis and not by a single, applicable list. Some elements recur in almost all the settings discussed so far:

• unique identification of the unit or batch, with reference to the manufacturer
• information on materials, with particular attention to substances of interest for recycling
• indications on durability, repairability and availability of spare parts
• instructions for correct end-of-life treatment

The difference compared to a traditional label is that this data isn't on the product itself; it's in an archive accessible through the product. The physical medium serves only to establish the connection, and for this reason, it must be extremely reliable: if the code can't be read, the information exists but is unreachable.

The QR code is the easy part

Generating a two-dimensional code is a trivial operation, handled by any information system. Even printing it on an adhesive label has been a problem solved for decades. The difficulties begin when the label isn't approved, and this often happens in industrial practice: components that go into the oven, metal parts exposed to solvents, products that must remain identifiable for their entire useful life.

In these cases, the code is engraved, laser-marked, or punched directly into the material. The result isn't black on white: it's a relief or micro-alteration of the surface, the contrast of which depends entirely on how it's illuminated. A code engraved on polished steel is practically invisible under frontal lighting and becomes perfectly legible with a grazing ring illuminator that casts sharp shadows on the edges of the cells.

Reading a code on a moving line

The second constraint is time. In the lab, a stationary code is read, with all the time necessary. On the production line, the part moves, often without a guaranteed position and with variable orientation. The reading system has a window of a few milliseconds and must produce a result, not a maybe.

The variables to be determined during the design phase are few but binding: transit speed at the reading point, part position tolerance, code cell size, and working distance. These factors determine exposure time, illumination intensity, and field of view. These are decisions made once, but they affect reliability for years to come, which explains why it's better to measure in the field rather than estimate at a desk.

Marking, verification, association: three different operations

Many projects fail because they confuse three distinct steps that need to be managed separately:

1. marking, that is, physically applying the code on the piece, with the correct process parameters for that material
2. code quality verification, which is not reading but rather evaluating its readability according to standardized criteria of contrast, uniformity and cell damage
3. the association, that is the connection between the read code and the record containing the data of that specific unit

Skipping verification is the most common and costly mistake. A code that today reads at ninety percent of the quality margin will continue to read in the factory for months, and then will stop working downstream, perhaps at the customer's site, when wear has consumed that margin. Verification exists precisely to intercept any deviations in the marking process before they become a field problem.

What happens when the code is not readable

No line reads 100% of codes on the first try, and designing as if it does guarantees downtime. Explicit exception management is required: the unread part is diverted, resubmitted, or taken to a manual station where an operator forces the association with a portable reader.

The rule worth writing into the specifications is that no unit is allowed to proceed without association. A part that leaves the plant with an unreadable code, or worse, associated with the wrong record, is exactly the case the digital passport should eliminate. The first-read rate is therefore the metric to monitor daily, because a drop in its rate signals a tagging problem long before anyone else notices.

Frequently Asked Questions about the Digital Product Passport

Since when is it mandatory?

There's no one-size-fits-all deadline, because the requirement applies to individual product groups, not the entire market. However, the most useful question in the factory is: how long does it take to be ready? Between choosing the marking method, testing actual materials, calibrating the reading system, and connecting to the management system, these projects are measured in months, not weeks, and the long part isn't the purchase, but the finalization.

Who should store the data, the manufacturer or an external platform?

Responsibility for the data remains with the economic operator placing the product on the market, regardless of where the data is physically hosted. Many companies will rely on service providers, but this does not shift responsibility and makes it crucial to define portability and retention periods in the contract.

Do you need a different code for each individual piece?

It depends on the required granularity, and in production the difference is clear. With batch identification, the marker repeats the same data for thousands of units and can operate autonomously. With serialization, each piece receives different data, which requires the marker, reader, and information system to remain synchronized in real time and a rule to be defined in advance regarding what to do when the sequence is interrupted.

Where is the best place to start?

The lowest-risk investment isn't in the information content, which will be defined elsewhere, but in the ability to mark and read reliably, which is still needed. The current status can be measured in a week and without purchasing anything: take a hundred or so pieces fresh off the line, evaluate their marking quality using standardized criteria, count how many are read on the first try at the current reader location, and how many associations end up in manual management. If the first-read rate isn't already high on new pieces, it won't be on pieces that have been through washes, ovens, and months in storage. In many plants, that effort pays off in internal traceability, regardless of any deadline.