Position Marks Must Be a Single, Precisely Delineated Sign: Non‑limiting Descriptions Creating Indeterminate Variations Fail Registrability

Case: Thom Browne Inc & Anor v Adidas AG [2025] EWCA Civ 1340

Court: Court of Appeal (Civil Division), England and Wales

Judgment date: 23 October 2025

Panel: Arnold LJ (leading), Lady Justice Falk, Lord Justice Peter Jackson

Introduction

This appeal concerns the registrability of a suite of “position marks” owned by Adidas featuring three stripes placed on apparel. Thom Browne (TB) had successfully challenged various Adidas registrations at first instance (Joanna Smith J: [2024] EWHC 2990 (Ch)) on the ground of lack of registrability. Adidas appealed as regards six UK registrations (three pairs of “Tracksuit Top”, “Tracksuit Bottom”, and “Vest” marks: 588/095, 661/093, 612/092), all depicting three stripes positioned on garments and accompanied by written descriptions containing flexible wording such as “running along one third or more of the length” of the relevant garment part.

The key issues were: (i) whether the subject-matter of each registration was “a sign” at all (the first condition under Article 2 of Directive 89/104 and its analogues), and (ii) whether the sign was “capable of being represented graphically” with the requisite clarity and precision (the second condition: the Sieckmann criteria). The case turned on the relationship between the pictorial representations and the written descriptions for position marks, and the extent to which non-limiting descriptions that go beyond the picture (creating “variations” not shown) create a multiplicity of signs or an unclear scope, rendering the registrations invalid.

There were parallel EUIPO and Dutch proceedings, but the Court of Appeal proceeded under the applicable UK framework and assimilated CJEU jurisprudence, with no material divergence in the governing principles.

Summary of the Judgment

• The Court of Appeal dismissed Adidas’ appeal, upholding invalidity for lack of registrability of the six “position mark” registrations.
• The Court affirmed that the subject-matter failed both:
  • the first condition (there was not a single sign but a multiplicity of signs due to open-textured descriptions like “one third or more”, leading to multiple possible forms), and
  • the second condition (the mark, read as image + description, did not provide a clear, precise, self-contained and objective representation of the protected sign).
• Grounds of appeal:
  • Ground 1 (Dyson/unrepresented signs): Rejected. The judge did not treat the mere existence of “unrepresented signs” as per se fatal, but properly treated the breadth of the description beyond the depiction as a relevant factor to both conditions.
  • Ground 2 (interpretation of scope/position): Rejected. Reading the description as allowing variable start/finish points, lengths, and position on the sleeve (or equivalent garment part) was legitimate; any arguable overreach (e.g. underside of sleeve) was immaterial to the outcome.
  • Ground 3 (“one third or more” clarity): Rejected. Words can be clear in isolation yet render the subject-matter unclear when combined with a depiction; such ambiguity produces an unfair competitive advantage and fails the Sieckmann clarity/precision requirements.
  • Ground 4 (origin message): Rejected. The judge did consider the origin-indicative effect and concluded the relevant public would not always “see the same mark.”
• Standard of review: the Court applied the limited appellate standard for evaluative assessments (Lifestyle Equities v Amazon; Iconix v Dream Pairs), finding no legal or principled error in the judge’s reasoning.
• Concurrences: LJ Falk emphasised that for position marks, distinctiveness depends at least in part on position; if position is not clearly specified, registrability will likely fail. Jackson LJ agreed.

Analysis

1) Legal Framework and the Nature of Position Marks

The governing provisions derive from Article 2 of Directive 89/104 (and the corresponding provisions of the 1994 Act and the Community Trade Mark Regulation), which require that a trade mark must be:

• a sign;
• capable of being represented graphically; and
• capable of distinguishing goods/services (the third condition, not in issue for the appeal).

These conditions are independent and cumulative (Libertel; Heidelberger; Dyson). The Court noted that “graphical representation” encompasses both pictorial representations and written descriptions, and that for “position marks” (marks whose distinctiveness derives from the conjunction of a visual element and its positioning on the goods), the interplay between image and description is central.

2) Precedents and Their Influence

• Sieckmann (C-273/00): Established the canonical representational criteria: a representation must be clear, precise, self-contained, easily accessible, intelligible, durable and objective (the “Sieckmann criteria”). The Court applied these as the second condition’s yardstick and as part of the fairness rationale for a precise, public register.
• Libertel (C-104/01) and Heidelberger (C-49/02): Colour marks are possible, but abstractions “in every conceivable form” without systematic arrangement fail for lack of precision and uniformity. The Court drew on this to reject open-ended descriptions that permit numerous different combinations, leaving the public and authorities uncertain of scope.
• Dyson (C-321/03): A claim to “all conceivable shapes” of a transparent bin is not a sign but a mere product property, conferring unfair advantage. The Court used Dyson to illustrate the first condition: if the subject-matter takes on a multitude of appearances, it is not a single sign. Crucially, the Court below did not misapply Dyson as a per se ban on descriptions that extend beyond the depiction; rather, breadth can show multiplicity and lack of specificity.
• Spear v Zynga [2013] EWCA Civ 1175 and Nestlé v Cadbury [2013] EWCA Civ 1174: The Court of Appeal has previously invalidated marks where descriptions embraced an unknown number of signs (Nestlé: “predominant colour” on packaging) or a multitude of permutations (Spear: letter-and-number ivory tile). These were central to affirming that where a registration wraps up multiple candidates for the comparison exercise, it fails the first and/or second conditions.
• Glaxo v Sandoz [2017] EWCA Civ 335: Confirms that inconsistencies between image, INID code, and description that leave the public “scratching their heads” about the protected sign fail Sieckmann. The Court quoted Kitchin LJ’s rationale: identity/confusion assessments require a precise knowledge of the sign as it will be seen; a registration cannot comprise multiple candidates for “identity.”
• Hartwall (C-578/17) and Red Bull (C-124/18 P): Written descriptions must clarify the subject matter and cannot be inconsistent with the pictorial representation; inconsistencies and plural non-uniform reproductions are incompatible with Article 2/4. These cases powerfully support the Court’s approach to the Adidas descriptions as extending beyond the images in ways that spawn indeterminate variations.
• Stihl (T-193/18): Merely persuasive and “not easy to reconcile” with Hartwall/Red Bull. The Court declined to be guided by Stihl in the face of binding CJEU authority.
• Standard of Review: Lifestyle Equities v Amazon [2024] UKSC 8; Iconix v Dream Pairs [2025] UKSC 25 — evaluative findings are interfered with only for error of law/principle. The Court found none.

3) The Court’s Legal Reasoning

a) Reading image and description together, but avoiding multiplicity: The Court accepted that optional written descriptions can and must be read together with images (as the judge did: [181]). However, if the description goes beyond the depiction to allow a wide variety of lengths, start/end points, and positions (e.g., “one third or more of the length of the sleeve”), the subject-matter can cease to be a single sign and become a multiplicity of signs (first condition), and the representation can cease to be clear and precise (second condition).

b) “Unrepresented signs” are a relevant factor, not an automatic knockout: The Court rejected the submission that the judge had treated the mere possibility of variations not shown in the picture as per se fatal under Dyson. Instead, breadth beyond the depiction was treated as evidence of multiplicity and uncertainty. The question remains one of fact and degree: is there still one sign with permissible variation, or has the description crossed into an indefinite set of signs? The Court agreed the latter applied here.

c) The “one third or more” problem: Although “one third or more” is textually clear, its effect in practice was to generate a large and indeterminate range of appearances not coherently captured by the depiction. Words can be clear in themselves yet render the subject-matter unclear when combined with a depiction (cf. Red Bull; Hartwall; Glaxo). This uncertainty confers an unfair competitive advantage on the proprietor by obscuring the scope of exclusivity.

d) Identity comparisons demand a single candidate: Echoing Glaxo and Nestlé v Cadbury, the Court emphasised that trade mark law requires meaningful “identity” and “likelihood of confusion” assessments. An earlier mark may not present multiple “candidates” for comparison that could yield different answers to “identical or not?” ([44]–[46]).

e) Evidence of use cannot cure registrability defects: Adidas’ “cross-check” argument (that use of variations showed a consistent origin message) was rejected. Whether the subject-matter qualifies as a “trade mark” is logically anterior to distinctiveness (Spear v Zynga). Use evidence does not resolve representational uncertainty or multiplicity, and, in any event, had not shown the public perceived all variants as “the same mark.”

4) Application to the Adidas Marks

The Court substantially adopted the judge’s analysis of the top exemplar (Tracksuit Top Mark 588), holding that:

• read with the written description, the pictorial depiction was only an example;
• the description’s “one third or more” wording encompassed multiple unrepresented variants of length, starting/finishing points, and positions along the sleeve, creating a “wide variety of different forms” ([195]);
• as a result, the marks failed:
  • the first condition: the subject-matter was not a single sign but a multiplicity; and
  • the second condition: the representation (image + text) was not clear or precise as to the nature and scope of the protected sign, leaving the public and authorities confused and conferring an unfair competitive advantage.

Similar reasoning applied mutatis mutandis to the companion marks for tracksuit bottoms and vests. Any arguable overreach (e.g. underside of sleeve, [193]) was immaterial; the invalidity was driven by the breadth and indeterminacy inherent in the “one third or more” architecture.

5) What This Judgment Adds

While grounded in established CJEU and Court of Appeal authority, this judgment makes a pointed statement for position marks in the apparel context:

• Position matters for distinctiveness; if the position is not clearly specified, the registration will likely fail.
• Optional written descriptions are double‑edged: if they extend beyond the depiction with open‑textured wording (e.g., “one third or more”), they may transform a single sign into a plurality and render the representation unclear.
• Permissible variation vs multiplicity remains a fact-and-degree inquiry; however, non‑limiting descriptions that claim variable extents and placements across a garment are at serious risk.
• Evidence of use does not salvage registrability where the subject‑matter and representation requirements are not met.

Impact and Practical Implications

A. For Position Marks Generally

• Applicants must delineate one sign. Open-ended descriptors that allow an indeterminate set of placements, lengths, or orientations will tend to fail both the first and second conditions.
• Where visual identity derives from placement (e.g., stripes on sleeves/pant legs), the position must be fixed or constrained in an objectively verifiable way. Flexible formulas (“at least x% of length,” “predominant portion”) are hazardous.
• Descriptions should clarify the subject matter and not broaden the image into a cloud of variants. Avoid the trap of declaring the depiction “only an example” where the text claims wider territory.

B. Drafting Guidance for Trade Mark Attorneys

• Align image and text: Do not let the description create inconsistencies or scope beyond the depiction (Hartwall; Red Bull).
• Specify position precisely: Identify the garment part and orientation (e.g., “outer lateral surface of the sleeve from shoulder seam to cuff along full length”), avoiding indeterminate ranges.
• Avoid open‑textured quantifiers: Phrases like “one third or more,” “in every conceivable form,” “predominant,” “equal proportions” without fixed arrangement, or “applied to a significant proportion” are red flags.
• Use multiple filings where necessary: If protection is sought for materially different locations or extents, file separate applications (potentially a series, where available) keyed to distinct images/specifications.
• Consider monochrome vs colour claims strategically: If colour is not the essence, register in monochrome to fix form while avoiding unnecessary variability.
• Do not rely on “mark type” labels or boilerplate: As the Court noted, entries like “Three-dimensional mark” or “as illustrated below” did not affect validity; substance controls.

C. Enforcement and Portfolio Strategy

• Identity and confusion assessments require a single, determinate earlier sign; ambiguous registrations weaken enforcement potential and risk invalidation at the outset.
• Acquired distinctiveness cannot cure defects at the “what is the sign?” stage; plan portfolios to avoid defective subject-matter/representation from the start.
• Expect closer scrutiny by registries and courts of position mark descriptions in apparel, footwear, and accessories; precision in placement and extent will be demanded.

Complex Concepts Simplified

• Position mark: A mark where distinctiveness comes from the visual element plus where it is placed on the product (e.g., a tab on a pocket corner). Its protectable essence is inseparable from position.
• First condition (“is it a sign?”): The subject-matter must be a single sign, not an open category that can take countless forms. If the registration would monopolise a mere product property, it fails (Dyson).
• Second condition (“how is it represented?”): The register entry (image + any description) must tell the world clearly and precisely what the sign is—passing the Sieckmann criteria. Inconsistencies or indeterminate groupings fail.
• Permissible variation vs multiplicity: Some small, predictable variation may be tolerable, but once variation becomes indeterminate and material (length, start/finish points, orientation), the sign ceases to be singular.
• Unfair competitive advantage: If ambiguity lets a proprietor claim broader exclusivity than clearly disclosed, that offends the fairness embedded in Article 2/4 and Sieckmann.
• Comparable UK trade marks: Marks cloned from EU registrations post‑Brexit. Here, differences between “comparable” and “pure UK” registrations (e.g., “3D mark” label, “as illustrated below”) were immaterial; the core problem was subject-matter and representation.

Discussion of Grounds of Appeal

• Ground 1 (Dyson): The Court clarifies that Dyson is not a per se rule against any unrepresented variants; rather, breadth in a description that generates multiple unrepresented forms is a strong indicator that the subject-matter is not a single, specific sign and that the representation is unclear. The judge applied the correct, nuanced approach (paras [69]–[72]).
• Ground 2 (Scope/position): The Court upheld the reading that the description encompassed different start/end points and lengths on the sleeve, and that variability in position followed from the text (para [75]). Even if one interpretive step went too far, it was immaterial (para [76]).
• Ground 3 (“one third or more”): Clear words can still produce an unclear subject-matter when combined with a depiction, especially where images are treated as non‑limiting exemplars (paras [77]–[79]).
• Ground 4 (Origin message): The Court confirmed that the judge considered whether the variants would convey the same origin message and rejected Adidas’ contention (paras [80]–[83]).

Relationship with EU and Foreign Decisions

• EUIPO Cancellation Division vs Board of Appeal: The Court noted divergences below and pending appeals; the Board had treated the marks as figurative and limited to the picture, which would render them valid but contradict Adidas’ broader reading. The UK Court was not persuaded to follow the Cancellation Division’s approach.
• District Court of The Hague: Treated with respect but not determinative; it did not reveal any legal error in the UK judge’s approach (para [84]).
• Assimilated CJEU law remains central: The Court applied Sieckmann, Libertel, Heidelberger, Dyson, Shield Mark, Hartwall, and Red Bull; Stihl was deemed difficult to reconcile with the CJEU’s core line and, in any event, only persuasive (paras [47], [84]).

Key Takeaways and Significance

• Core holding: A position mark that is depicted by an image but accompanied by an open-textured description (e.g., “one third or more”) that materially widens possible appearances fails both the “single sign” requirement and the Sieckmann representation criteria.
• Descriptive overreach is risky: Optional descriptions must be carefully drafted to avoid creating multiplicity or uncertainty; if needed, file multiple, tailored applications.
• Fairness and functionality of the register: The public and competitors must be able to determine the scope of protection without ambiguity; the system will not tolerate registrations that create puzzles or afford unfair competitive advantage.
• Position marks under the microscope: Because their distinctiveness depends at least in part on position, courts and registries will demand precise specification of position and extent.
• No curing by use: Evidence of use and acquired distinctiveness does not rescue an application that fails at the definitional/representational stage.

Practical Checklist for Filing Position Marks

• Define a single sign with minimal, clearly bounded variation.
• Ensure the image and description align; the description should clarify, not expand, the image.
• Specify exact position and extent (start/end points, orientation) in objective terms.
• Avoid wording like “one third or more,” “predominant,” “in every conceivable form,” “significant proportion,” unless the image and description tightly constrain how that operates in practice.
• Where necessary, file multiple marks for materially different placements/lengths.
• Consider whether colour is claimed or incidental; if incidental, register in monochrome.
• Do not assume that portfolio use or consumer recognition can fix representational ambiguity.

Conclusion

The Court of Appeal has delivered a clear, position‑mark‑specific restatement of foundational registrability principles. Optional written descriptions cannot be used as a vehicle to capture a broad range of unrepresented positions and extents that materially depart from the depiction. Doing so risks failure of both the first condition (no single sign) and the second condition (no clear, precise representation), thereby undermining the ability of the register to function fairly and predictably.

For brand owners, particularly in apparel and footwear, the message is stark: If the distinctiveness of your mark depends on position, the application must pin down that position and its visual manifestation with precision. Where coverage of different placements or lengths is needed, use multiple filings rather than a single open‑textured description. Anything less will court invalidity and weaken enforcement from the outset.