On 23 September 2026 the District Court of Gelderland, sitting in Arnhem, gave judgment for Birkenstock in copyright proceedings against Lidl Nederland and Rieg & Niedermayer, named on the sandals' labels as their producer and importer, over sandals resembling five Birkenstock models: the Arizona, Madrid, Gizeh, Boston and Florida (ECLI:NL:RBGEL:2026:7234). Both were ordered to stop offering the sandals in the Netherlands on pain of a penalty of 5,000 euros a day, up to 250,000 euros, to provide information on sales and orders, and to compensate Birkenstock for its loss, in an amount to be assessed in separate proceedings. They were also ordered to pay Birkenstock's legal costs, assessed at about 47,000 euros. The claims against a second Lidl company were dismissed. The judgment is at first instance and can be appealed, though its orders were declared provisionally enforceable.

Birkenstock's published account is that all the challenged sandals infringe, "in particular, the copyright in the footbed design." The judgment itself describes something narrower, and for governance purposes more instructive. The court did not protect the sandal as a whole, or even the footbed as a whole. It found one element of the sole unit original: a height difference that forms a raised wall at the rear and slopes down towards the front. It also protected an element of the upper of the Madrid model. The other footbed features Birkenstock relied on, including the heel support, the raised areas of the footbed and the shape around the toes, were held to be dictated by the anatomy of the human foot. Other features failed because they were functionally determined or, though freely chosen, not creative. None of them was protected.

The commercial facts are interesting. The governance lesson is more important. A court separated a mass-produced, deeply functional product into the parts that reflect free and creative choices, the parts that function dictates, and the parts that were freely chosen but not creative, and it gave protection only to the first. That is exactly the analysis a product business needs to be able to run on its own designs, and on its competitors', before litigation rather than during it.

A sandal can be functional and still raise copyright questions

Businesses often treat copyright as relevant to paintings, photographs, illustrations and decorative surface design, and to little else. Under EU copyright law that assumption does not hold. A functional product can qualify for protection where the relevant design reflects original creative choices. That does not mean every useful product becomes copyright protected. It means functionality and copyright are not mutually exclusive.

The same sandals illustrate how contested the line is. In February 2025 Germany's Federal Court of Justice held that the Arizona, Madrid, Boston and Gizeh models were not protected works of applied art under German copyright law (I ZR 16/24 and related cases). Different courts, different claims and different legal questions: the German decision does not make the Dutch one wrong, or the reverse. But a business that assumes a design is either obviously protected or obviously free to copy is making a judgment two national courts have now reached differently about the same products.

EU law already recognises the distinction

The Court of Justice has addressed this question more than once. In Cofemel v G-Star Raw (C-683/17), a case about clothing designs, the Court confirmed that a design may qualify as a copyright work where it meets the EU-law originality threshold. Design protection and copyright protection are separate regimes, and one does not automatically produce the other.

In Brompton Bicycle (C-833/18), the Court held that copyright protection is not excluded merely because the shape of a product is, at least in part, necessary to achieve a technical result. The question is whether, through the shape adopted, the creator actually expressed free and creative choices; the mere availability of other shapes is not decisive. Where the shape is dictated solely by its technical function, protection is excluded.

In December 2025, in the joined cases Mio and konektra (C-580/23 and C-795/23), the Court confirmed that works of applied art are subject to the same originality standard as any other work, with no higher threshold, and that infringement turns on whether creative elements of the work are recognisably reproduced rather than on an overall impression.

That matters for consumer products. A sandal must accommodate a foot. A chair must support a person. A lamp must produce light. A garment must be wearable. Function constrains design, but it does not necessarily eliminate creative choice, and the Gelderland judgment shows a court drawing that boundary element by element.

The compliance mistake is to ask only whether a product is "functional"

That is too crude a question. A better internal review asks:

  • Which elements of the product are dictated by technical necessity?
  • Which elements reflect discretionary aesthetic or structural choices?
  • Which of those choices may be original enough to attract copyright protection?
  • Which elements derive from established industry conventions or common design vocabulary?
  • What source material influenced the design?
  • What evidence shows how the design developed?
  • Are there earlier products with materially similar features that may be protected?

This is not a legal-team exercise conducted immediately before launch. It is a product-development control.

Why this matters for lookalikes

The commercial attraction of a lookalike is obvious: a recognisable silhouette, construction language or product configuration already has market validation. That is precisely why lookalike development needs disciplined review. A business may believe it has changed enough visual details to avoid a registered-design issue and still meet a copyright claim over another protectable aspect of the product. Under a recognisability test, one reproduced creative element can be enough, as a single element of the sole unit was here.

Conversely, similarity alone does not establish infringement. A competitor may lawfully use the same general product archetype, respond to the same functional constraints or independently make similar choices. The governance question is not simply "does this look similar?" It is: what is similar, why is it similar, and what evidence supports the conclusion that the similarities are permissible?

Copyright governance needs provenance

This is where many product-development processes remain weak. By the time counsel sees a challenged product, the final design may be clear but the decision history behind it is not. Who introduced the contested element? Was it in the initial brief? Did it come from a reference image? Was it introduced by a designer, a supplier, an external studio or a generative system? Was it dictated by manufacture? Was an apparently similar feature actually inherited from a generic product type? Did the team identify a resemblance and consciously redesign around it?

Without provenance, these questions become reconstruction exercises. Good IP governance therefore requires more than storing the final CAD file, sketch or product photograph. It requires a defensible record of how the product became what it is.

Generative AI makes that need more acute

A creator can provide an original brief and still receive a generated design containing features the creator never requested. A later edit may carry those features into a production specification. By launch, the organisation may know what it manufactured but not where every material design choice entered the process.

This is why AI governance and IP governance are beginning to converge. The relevant control is not merely whether generative AI was used. It is whether the organisation can preserve the user's original design intent, the reference materials supplied, material changes introduced during generation, human modifications, review and approval decisions, similarity signals or concerns, technical constraints affecting the design, and the final production state. An AI-generated image is not itself proof of provenance.

Screening is not clearance

There is also a temptation to over-automate this problem. Similarity detection can be useful: it can identify designs that deserve human review. It cannot determine by itself whether copyright subsists, whether a protected element has been reproduced, whether similarities derive from common functional constraints, or whether a legal defence applies. Those questions depend on facts, jurisdiction and legal analysis, and as the German and Dutch sandal decisions show, on the court.

Organisations should therefore keep three things apart:

  • Similarity signal. Something appears close enough to warrant attention.
  • Governance decision. The organisation investigates, documents and decides whether to proceed, modify or escalate.
  • Legal determination. Whether copyright exists and whether infringement has occurred.

Collapsing those three stages into an automated "safe" or "not safe" score creates false assurance.

What organisations should do now

Product-led businesses should treat intellectual-property review as part of design governance rather than a final pre-launch checklist. A practical control framework includes:

  1. Design provenance. Preserve the origin of significant design decisions, references and later modifications.
  2. Functional-versus-creative analysis. Identify, element by element, where technical necessity ends and discretionary design begins.
  3. Reference governance. Record third-party designs supplied to designers, agencies, manufacturers and AI systems.
  4. Similarity escalation. Define when a resemblance requires human or legal review.
  5. Human sign-off. Do not delegate copyright conclusions to automated similarity systems.
  6. Version history. Keep enough design evolution to reconstruct why the final product looks the way it does.
  7. Supplier controls. Require manufacturers and external designers to disclose sources, adaptations and externally supplied design elements.
  8. Rights strategy. Consider copyright alongside registered and unregistered design rights, trademarks and other protection, rather than treating each regime in isolation.

The broader lesson

The Birkenstock-Lidl dispute is interesting because footwear sits precisely where many IP-governance assumptions become uncomfortable: it is commercial, mass-produced and deeply functional. Yet EU copyright doctrine does not ask whether an object is useful and stop there. It asks whether protectable creative expression remains within that useful object, and the answer can be one element rather than the whole.

For companies designing physical products, that turns IP governance into a lifecycle question. Not simply "do we own this design?" but: can we show what was created, by whom, from what inputs, under which constraints, through which changes, and why we were comfortable taking it to market? That is a much stronger compliance posture.

How Priventia approaches it

Priventia's governing principle is that every compliance conclusion should be explainable. Its method links the legal requirements that apply to an organisation to the controls that address them, the evidence behind those controls and the monitoring that keeps them current, with every step traceable to the provision it comes from.

Intellectual property is one of Priventia's regulatory domains, with copyright as its first family, and its coverage will be built on the same discipline. The objective is not to automate legal judgment. It is to make the governance surrounding that judgment visible, repeatable and auditable.

Sources: District Court of Gelderland, 23 September 2026, ECLI:NL:RBGEL:2026:7234 (full text, rechtspraak.nl); Birkenstock Group press release; Bundesgerichtshof, 20 February 2025, I ZR 16/24; CJEU, Cofemel v G-Star Raw, C-683/17; CJEU, Brompton Bicycle v Chedech/Get2Get, C-833/18; CJEU, Mio and konektra, joined cases C-580/23 and C-795/23.