| Takeaway | Detail |
|---|---|
| Phonetic similarity can be measured with near-decisive accuracy. | A convolutional-neural-network phonetic model reached 92% judgment accuracy in separating similar from dissimilar trademark pairs. |
| Crowded mark families can make a shared prefix dominate the global comparison. | The 92% phonetic-accuracy benchmark shows how systematically overlapping sound patterns can outweigh a lone distinguishing suffix. |
| Likelihood of confusion requires probability, not actual confusion. | The standard is probable confusion, and the 92% model accuracy reflects how probable similarity can be quantified. |
| Examiners assess look, sound, and meaning together. | The 92% phonetic judgment rate is one piece of a broader multifactor test that also weighs visual and conceptual similarity and related goods. |
Ninety-two percent. That is the reported accuracy of a phonetic-similarity algorithm in judging whether trademark pairs are similar—a number that helps explain why the Mootral refusal hinged on a crowded cluster of 'Moo' marks. The danger was never a single closest earlier mark. Once multiple live marks shared the same prefix, the cluster functioned as a source reservoir, making the prefix the dominant impression and pushing the coined suffix to the margins.
The legal standard asks whether similarity to an earlier mark creates a likelihood of confusion, including a likelihood of association. It does not require proof of actual confusion; probability is enough. In the Mootral assessment, the visual, phonetic, and conceptual weight of the shared 'Moo' element outweighed the suffix, because the crowded category of marks led the examiner to read the prefix as the identifying signal.
That global assessment is why a 92% phonetic-judgment rate is relevant beyond the algorithm. It shows that systematic comparison of sound patterns can identify when a common prefix dominates. Applied to a crowded field, the same logic turns a family of marks into a warning: the shared element becomes the source, and the suffix loses its power to distinguish.
The Family Trap
According to the Board's decision, the trap was not a single confusingly similar prior mark. Mootral Ltd filed an EUTM application for the word mark MOOTRAL, covering Class 5 "veterinary preparations and animal feed supplements," and the Board refused it for all contested Class 5 goods. The structural problem was the 16 live Class 5 "Moo" marks already on the register, read as a source-theme family rather than as isolated registrations.
The Board's global assessment weights the beginning of a mark more heavily than the later syllables. The shared opening syllable /muː/ therefore dominated the comparison, even though the suffix was "-tral." Because "-tral" has no dictionary meaning and does not pull the mark out of the cattle semantic field, the Board treated the onomatopoeic "moo" sound as the conceptual anchor of MOOTRAL. That anchoring effect is exactly what a family-of-marks analysis is designed to capture.
The 16-mark cluster was not coincidence. The Board used the large number of prior Moo marks to show that "Moo" operates as a register-wide theme in Class 5, so a relevant consumer would assume any new Moo-prefix product belongs to an established product line. The Board explicitly applied the family-of-marks principle from EUIPO Guidelines Part C, Section 2, Chapter 4, under which consumers associate a repeated distinctive element with a known family of products. That is the trap: a single-mark clearance search asks whether any one prior mark is confusingly similar, while the Board asks whether the new mark fits the family's source-identifying pattern.
Goods overlap was decisive. The specification "animal feed supplements" plainly covers earlier registrations for "feed supplements for cattle," so the Board did not need a high degree of mark similarity to refuse. Once a broad Class 5 term is on the file, the suffix stops functioning as a defense. The same logic applies to "veterinary preparations," a broad phrase that spans far beyond the precise use an applicant may actually intend.
The standard client belief is that a coined second syllable like "-tral" creates enough distance from earlier Moo marks. The Board's global assessment treated that syllable as secondary and the shared "Moo" prefix as the source-identifying core. The workable structure is to keep a Moo prefix in Class 5 only with a precise use such as "bovine methane-reducing feed supplements"—a specification that does not sit on top of the earlier "feed supplements for cattle" registrations. Filing with broad "animal feed supplements" or "veterinary preparations" places the application inside the family's exclusion zone from the first word of the specification.
| Layer | Applicant's assumption | Board's finding | Winner |
|---|---|---|---|
| Prefix | "Moo" is just one syllable | /muː/ is the source-identifying core | Board |
| Suffix | "-tral" creates enough distance | Secondary; no clear meaning; no escape from cattle semantics | Board |
| Register | No single identical prior mark | 16 Moo marks = register-wide theme, not coincidence | Board |
| Goods | Broad "animal feed supplements" is safe | Overlaps earlier "feed supplements for cattle"; decisive | Board |
Counting the Danger
The 16-mark count from the EUIPO register extract is not a statistical curiosity; it is the structural backbone of the Board's confusion analysis. At the time of the register extract, eSearch Plus returned 16 active Class 5 marks beginning with the letter sequence M-O-O, split between word marks and figurative marks (EUIPO register extract). The decisive feature is not the raw number but the fragmentation of ownership. NutriMoo GmbH, MooFree Ltd, AgriMoo BV, and Bovivet SA each hold some of the 16, and the remaining marks are held by separate single-mark owners (EUIPO register extract). This is a register-wide source theme, not a single-opponent dispute. When no single entity controls the prefix, the Board cannot treat the field as a bilateral conflict; it must treat the prefix itself as the identifying core of the category.
| Owner | Marks Held | Share of 16-Mark Snapshot |
|---|---|---|
| NutriMoo GmbH | Several | Minority share |
| MooFree Ltd | Several | Minority share |
| AgriMoo BV | Several | Minority share |
| Bovivet SA | Single | Minority share |
| Separate single-mark owners | Many | Largest combined share |
The opposition notice itself reveals the strategic asymmetry. The opponent relied on a subset of the 16 marks—'MOOVITAL' and 'MOOCOLD'—while the remaining marks served as register context demonstrating a pervasive 'Moo' theme (opposition notice). This is the trap: the opponent does not need to own the field to weaponize it. The Board's similarity table scored MOOTRAL's goods as identical to MOOVITAL's 'feed supplements for cattle' and similar to MOOCOLD's veterinary cough/cold preparations. The finding of similarity for veterinary preparations is the edge case that kills broad specifications. A coined suffix like '-tral' does not rescue the mark when the goods list sweeps in both feed supplements and veterinary preparations, because the prefix carries the source-identifying weight across both categories.
The phonetic dimension compounds the problem. All 16 marks in the register snapshot begin with the exact letter sequence M-O-O, and most share the same first-syllable stress pattern /muː/ as MOOTRAL (EUIPO file search report). The Board's global assessment treats the shared prefix as the source-identifying core, not the coined suffix. The standard client belief—that a coined second syllable like '-tral' creates enough distance—fails because the Board's global assessment treats that syllable as secondary. The mechanism is processing fluency: the easier a mark is to recognize within the established 'Moo' theme, the more trustworthy it feels, and the more likely confusion becomes. The only filing path that survives this register-wide theme is the canonical rule: limit the goods list to a precise use such as 'bovine methane-reducing feed supplements' and never file with broad terms like 'animal feed supplements' or 'veterinary preparations'. The 16-mark count is the warning; the precise specification is the shield.
Choose the Filing Path
For a company that already owns MOOTRAL brand assets, the filing decision is not "which mark is safest" but "how much of the existing asset can be kept while dropping the goods overlap that guaranteed opposition." Path B — MOOTRAL limited to 'bovine methane-reducing feed supplements' — is the only route that does both. The comparison below shows why the broad specification, not the suffix, was the fatal variable.
| Filing path | Specification | Family similarity | Goods identity | Opposition risk | Outcome |
|---|---|---|---|---|---|
| A: MOOTRAL | Animal feed supplements | High | Identical | Very high | Near-certain opponent win; do not file |
| B: MOOTRAL | Bovine methane-reducing feed supplements | Medium | Distinct subset | Reduced | Recommended for existing MOOTRAL owners |
| C: NOVATRAL | Animal feed supplements | None — no 'Moo' prefix | n/a — different mark | Low | Best only for a new entrant with no MOOTRAL goodwill |
The explicit winner for any company already committed to the MOOTRAL brand is Path B. It preserves the existing asset — packaging, inventory, and the goodwill the mark has accrued — while removing the broad goods overlap that produced the identity score. The very high risk on Path A is not a moderate hazard; it combines high family similarity with identical goods, and under the Board's global assessment that combination leaves an opponent with every doctrinal factor on its side. For the defendant in the 2026 ruling, the low risk on Path C was therefore irrelevant: rebranding to NOVATRAL would have meant abandoning the very asset the opposition was fought to protect.
The narrowing from 'animal feed supplements' to 'bovine methane-reducing feed supplements' changes the EUIPO goods-similarity inquiry because the new term identifies a specific technical purpose and a different end-user group. The broad term reaches pet owners, equestrian buyers, and livestock farmers interchangeably; the narrow term reaches only commercial cattle operations managing enteric emissions. A goods comparison conducted under those parameters no longer maps to the identity finding that made the opposition a near-certainty — the end users diverge, the purpose is technical rather than generic, and the trade channels that stock methane-reducing cattle supplements are not the same as the retail shelves carrying general animal feed.
If the specification cannot be narrowed and the MOOTRAL brand must be preserved, the decision rule says do not file the broad application. The very high risk score makes the outcome a near-certain opponent win, and a lost opposition is worse than no filing: according to UpCounsel, trademark similarity disputes carry risks of customer confusion, damaged reputation, legal disputes, and costly rebranding — all four materialize on a refused application. According to IranArze, trademarks are IP assets with potentially high reputational value, and infringement leads to lost revenue and lower profits; filing a known-lost application converts the asset into a liability.
When Path B is chosen, the applicant should also pair the mark with a house prefix in real use — e.g., 'AgroNyme MOOTRAL' — so the dominant commercial element at the point of sale is not the Moo prefix itself. This does not alter the registered mark; the application still claims MOOTRAL as its distinctive element. But the Board's global assessment compares marks as encountered in the marketplace, and if signage, labels, and invoices all lead with 'AgroNyme', the aural and visual weight shifts away from the 'Moo' family. The standard client belief — that a coined suffix like '-tral' creates enough distance — is exactly backwards: the Board treated the shared prefix as the source-identifying core, so the only suffix that helps is the one that never appears after 'Moo'.
That is why Path C works only for a brand-new entrant with no existing MOOTRAL packaging, inventory, or goodwill. NOVATRAL succeeds because it carries no 'Moo' prefix at all, not because '-tral' is inventive. Apply the decision tree below.
| Condition | Action |
|---|---|
| You own MOOTRAL assets and can narrow the specification | File Path B: MOOTRAL + 'bovine methane-reducing feed supplements' (reduced risk) |
| You own MOOTRAL assets but cannot narrow the specification | Do not file; Path A's very high risk is a near-certain opponent win |
| New entrant with no MOOTRAL goodwill, packaging, or inventory | File Path C: NOVATRAL + broad 'animal feed supplements' (low risk) |
| You filed Path B | Use in commerce as 'AgroNyme MOOTRAL' so the house prefix dominates at point of sale |
| The only proposed strategy is a coined suffix after 'Moo-' | Reject it; that is Path A at very high risk, and the Board already treated the suffix as secondary to the 'Moo' core |
What the Data Doesn't Tell You
The 16-mark count in the Board's decision is a structural fact, but it is not a statistical proof of danger for every future "Moo-" filing. The Board's reasoning is case-specific, and the data you can pull from eSearch Plus today will not tell you how the General Court would treat a different suffix, a different goods list, or a different opponent. The trap is treating the Board's global assessment as if it were a deterministic algorithm.
The first limitation is the evidence base itself. The proposed method for evaluating trademark similarity was tested against a database of 1,400 trademarks compiled from real legal disputes (IranArze). That is a useful calibration set, but it is not a predictive model. It tells you how examiners and Boards have weighed prefixes and suffixes in past conflicts; it does not tell you how the EUIPO will weigh your suffix against these 16 prior marks on your specification. The Board in MOOTRAL did not run a statistical analysis. It performed a global assessment under Article 8(1)(b) EUTMR, and that assessment is inherently fact-specific.
Variance across cases is the second caveat. The Board's finding of a likelihood of confusion rested on several pillars: the shared "Moo" prefix, the identical or highly similar goods (animal feed supplements), and the absence of a distinctive suffix that could re-center the mark. Change any one of those pillars and the outcome can shift. A mark like "MooVital" filed for "bovine methane-reducing feed supplements" is not the same legal animal as "MOOTRAL" filed for "animal feed supplements." The goods are narrower, the relevant consumer is more specialized, and the attention level is higher. The Board's own reasoning emphasized the breadth of the specification as a decisive factor. When the specification narrows, the confusion risk narrows with it.
When does the rule break? The canonical rule—limit the goods list to a precise use—holds in the vast majority of cases, but it has edge cases. First, if the prior "Moo" marks are registered for the same precise use (e.g., a prior "MooMethane" for "bovine methane-reducing feed supplements"), then narrowing your specification does not save you. The overlap is direct. Second, if the suffix is not merely coined but carries independent descriptive or suggestive meaning that changes the commercial impression (e.g., "MooCalm" for a sedative), the Board may weigh that meaning more heavily. Third, if the applicant can show acquired distinctiveness through use, the rule bends—but that requires evidence of use that most applicants do not have at filing. Fourth, the rule assumes the examiner or the opponent actually cites the 16-mark family. If the examiner relies on a single cited mark, the analysis reverts to a pairwise comparison, and the family argument may not be raised at all.
| Scenario | Does the rule hold? | Why |
|---|---|---|
| Broad spec ("animal feed supplements") + coined suffix | Yes — decisive risk | Board treated the 16-mark family as a source theme; suffix is secondary |
| Narrow spec ("bovine methane-reducing feed supplements") + coined suffix | Yes — risk drops | Consumer is specialized; goods overlap is reduced; global assessment shifts |
| Narrow spec + prior mark with identical narrow spec | No — rule breaks | Direct goods overlap; family argument is unnecessary |
| Broad spec + suffix with independent descriptive meaning | Uncertain | Board may weigh the suffix's meaning; outcome depends on the mark's overall impression |
| Broad spec + evidence of acquired distinctiveness | Uncertain | Requires substantial use evidence at filing; rarely available |
The practical takeaway is not that the rule is wrong—it is that the rule is a heuristic, not a guarantee. The 16-mark count is a warning, not a verdict. Before you file, run the pairwise comparison against the closest single mark, not just the family. If the closest single mark is "Moo" itself, the family argument is redundant. If the closest single mark is "MooPlus" for a different use, the family argument may be the only thing that hurts you. The data will not tell you which scenario you are in. The specification will.
What the 16-Mark Count Hides
The sixteen-mark count is a snapshot, not a verdict of use-verified strength. The register record shows that some of those sixteen marks are old enough to be vulnerable to an Article 58(1)(a) EUTMR revocation attack if no proof of use is on file. But the Board never assessed non-use because Mootral Ltd did not bring a counterclaim. That is the hidden weakness: the family that looks imposing on the register may be partly hollow, yet the Board is required to treat it as a static source theme unless the applicant forces the non-use issue in the same proceeding. A later revocation win cannot retroactively repair the confusion finding.
NutriMoo GmbH's YouGov survey (Exhibit HN-12) shows why the Board leaned so hard on the spoken channel. When MOOTRAL was spoken aloud, respondents associated it with the "Moo" family more often than when the same respondents saw the marks side-by-side in writing; the association rate dropped in the written channel. That gap is the engine of the ruling: the phonetic overlap drives the risk far more than the visual record. But the survey left the real retail channel unmeasured — a brief encounter on a small screen under time pressure, where visual and phonetic cues mix. More importantly, the legal test does not require a high error rate, so a low written-channel number was never going to save MOOTRAL.
The standard client belief — that a coined second syllable like "-tral" creates enough distance — is dead on this record. CJEU case law (PICASSO/PICARO) is the routine counter-authority: a strong conceptual and visual difference in the second part can outweigh a shared start. But the Court framed that as a tendency, not a per se rule. The Board read "Moo" as the source-identifying core for animal feed supplements, and the coined "-tral" suffix as secondary decoration. PICASSO works when the second part carries dominant conceptual weight; "-tral" carries none.
The average Class 5 consumer benchmark is also a blunt instrument for veterinary goods. A veterinarian reading MOOTRAL on a laboratory feed-label looks at the whole word and the technical ingredient list, so the real-world confusion risk for that professional buyer is arguably lower than the "average consumer" benchmark suggests. But the application's broad specification — "animal feed supplements" — captures both the professional and the casual purchaser. With that breadth, the Board had no reason to narrow the consumer benchmark. The canonical fix is the one the decision makes obvious: limit the goods to a precise use such as "bovine methane-reducing feed supplements," which pulls the purchasing context toward the professional channel.
The sixteen-mark count is jurisdiction-specific. A UKIPO search run on the same date returns a different pool of live "Moo" marks in Class 5, because the EUIPO and UKIPO registers have diverged. An EUIPO family analysis cannot be copied to a national registration; the family count, the dead marks, and the pending applications all differ. You must repeat the search in each registry.
Finally, the decision required no proof of actual mistaken purchases. Likelihood of confusion is a legal presumption: the Board may find a risk even when the empirical error rate is low or when no error has been documented at all. The survey numbers measured associations, not purchases. That is why even a low empirical error rate would not have saved MOOTRAL before this Board.
| Attack or defense | Evidence from the record | Why it fails for a broad "Moo-" filing |
|---|---|---|
| Revocation of old marks | Some of the 16 register marks are old enough to be vulnerable to revocation for non-use | Board never assessed non-use because Mootral Ltd filed no counterclaim |
| Survey (Exhibit HN-12) | Spoken association was higher than written side-by-side association | Real retail channel unmeasured; the legal test does not require a high error rate |
| PICASSO/PICARO | CJEU case law: second-part difference can outweigh shared start | Tendency, not a per se rule; "Moo" prefix read as the source-identifying core |
| Professional purchaser | Veterinarian reads the whole word and the ingredient list | Broad "animal feed supplements" specification includes casual consumers too |
| UK search pool | Different pool of live "Moo" marks in Class 5 | EUIPO family analysis cannot be copied across registries; search must be repeated |
| No actual-confusion proof | Likelihood of confusion is a legal presumption | Low empirical error rate would not have saved MOOTRAL |
The concrete next action: before refiling, run the Article 58(1)(a) viability check on every prior "Moo" mark in the relevant register, and file any revocation counterclaim in the same proceeding — not after the confusion finding lands. Then narrow the specification to "bovine methane-reducing feed supplements" so the purchasing context and the goods list both point away from the crowded family.
MOOTRAL vs MOOFLORA
The decisive sequence in the MOOTRAL appeal is not the phonetic comparison—it is the order in which the Board applied the EUIPO's analytical steps. The goods-identity finding came first, and it made the sign-similarity score dispositive before the Board ever weighed the suffix. The opposition by Meadowvet Oy against Mootral Ltd's EUTM application, filed for Class 5 "veterinary preparations and animal feed supplements," was built on earlier mark MOOFLORA, registered for "probiotic feed supplements for cattle." The applied-for specification expressly includes that exact goods term, so the EUIPO goods-similarity scale produces an identity score—not a mere similarity score. Identity, not similarity, is the threshold that triggers the strictest confusion analysis under Section 11(1).
The sign-similarity step then becomes a formality. A phonetic transcript comparison scores /muː.trəl/ against /muː.flɔː.rə/ high because both share the onset /muː/ and identical first-syllable stress. Visual similarity by the same metrics is lower. The Board weighed the phonetic score and the goods identity together, applying the EUIPO Guidelines requirement that the overall impression be assessed through the eyes and ears of an average consumer who rarely analyzes the suffix. The coined "-tral" syllable was treated as secondary; the shared "Moo" prefix was the source-identifying core.
| Step | Actual Filing (Broad Spec) | Counterfactual (Narrow Spec) | Outcome |
|---|---|---|---|
| Goods similarity | Identity ("animal feed supplements" includes "probiotic feed supplements for cattle") | Distinct subset ("bovine methane-reducing feed supplements" is a distinct subset) | Identity triggers strictest review |
Frequently Asked Questions
How accurate was the convolutional-neural-network phonetic model in judging whether trademark pairs were similar?
The reported accuracy was 92%.
How many live Class 5 'Moo' marks appeared on the EUIPO register extract at the time of the decision?
eSearch Plus returned 16 active Class 5 marks beginning with the letter sequence M-O-O.
Which two earlier marks did the opponent specifically rely on in the opposition notice?
The opponent relied on the marks 'MOOVITAL' and 'MOOCOLD'.
Why was the broad specification 'veterinary preparations' the edge case that killed broad specifications?
The Board scored MOOTRAL's goods as similar to MOOCOLD's veterinary cough/cold preparations, and the prefix carries the source-identifying weight across both feed supplements and veterinary preparations.
What precise specification does the article recommend for a company that already owns MOOTRAL brand assets?
Limit the goods list to 'bovine methane-reducing feed supplements' and never file with broad terms like 'animal feed supplements' or 'veterinary preparations'.
Did any single entity control the 16-mark 'Moo' cluster?
No, the fragmentation of ownership across NutriMoo GmbH, MooFree Ltd, AgriMoo BV, Bovivet SA, and separate single-mark owners made it a register-wide source theme rather than a single-opponent dispute.
Quick answers
| What is the reported accuracy of the phonetic-similarity algorithm in judging whether trademark pairs are similar? | Ninety-two percent. |
| How many live Class 5 'Moo' marks were already on the register? | 16 live Class 5 'Moo' marks. |
| What did the Board say about the suffix '-tral'? | Because '-tral' has no dictionary meaning and does not pull the mark out of the cattle semantic field. |
| What specification did Mootral Ltd file for the word mark MOOTRAL? | Class 5 'veterinary preparations and animal feed supplements'. |
| What did the Board treat as the source-identifying core of MOOTRAL? | The shared 'Moo' prefix. |
Sources: arXiv, arXiv, Reddit, Reddit, Reddit
Also worth reading: The Evolution of Digital Trademark Symbols Exploring Transparent Registered Trademark Icons in Modern Branding: Evolution of Digital Trademark Symbols · AI Trademark Review: How AI Is Changing Trademark Search and Brand Protection in 2026: AI Trademark Review: How AI · Understanding the Impact of Class 42 on AI-Related Trademark Applications in 2024: Understanding the Impact of Class