Likelihood of Confusion: Compare the Factors Yourself

Two trademarks do not have to be identical to collide. They only have to be so similar that the public might confuse them. That is exactly what a likelihood of confusion describes, and under § 14 (2) no. 2 MarkenG it never arises from a single factor. Sign similarity, similarity of goods or services and the distinctiveness of the earlier mark work together. The tricky part: the DPMA and EUIPO do not check any of this when you file. In this guide I show you which factors the offices and courts weigh, how to compare sound, spelling and meaning yourself, and why interdependence changes everything. (This is practical know-how, not legal advice.)

Key Takeaways

• A likelihood of confusion needs sign similarity and similarity of goods together (§ 14 MarkenG).
• The offices do not examine relative grounds for refusal on their own initiative, only through an opposition.
• Signs can be similar aurally, visually or conceptually.
• Interdependence balances weak similarity on one axis against strong similarity on the other (CJEU Canon, 1998).
• Strong, distinctive trademarks enjoy broader protection.

Before you work through the factors, it pays to look at concrete signs: you can search your trademark and assess the hits alongside this guide.

What does likelihood of confusion actually mean?

A likelihood of confusion means that the public might attribute two trademarks to the same origin. According to CJEU Sabèl/Puma (1997), it is not judged by arithmetic but in an overall assessment from the perspective of a reasonably well informed consumer. What counts is the overall impression of the signs, not the isolated comparison of individual letters or syllables.

The benchmark is the average consumer of the goods concerned. They usually perceive a trademark as a whole and do not pay attention to every detail. They also rarely see the competing signs side by side, but rely instead on an imperfect recollection. That is exactly why the overall impression carries so much weight.

Important for your planning: whether your desired mark comes too close to an earlier one is not something the office checks when you file. The DPMA and EUIPO only review absolute grounds for refusal, such as a lack of distinctiveness. Relative grounds for refusal, meaning clashes with third-party rights, they do not examine on their own initiative, according to § 9 MarkenG and Art. 8 EUTMR.

The likelihood of confusion with your trademark is only raised once an earlier owner takes action. Until then it stays invisible, but not harmless.

This policing runs through the opposition (a challenge by the owner of an earlier mark). The window is tight: three months from publication of the registration, both at the DPMA and the EUIPO. Within that window it is decided whether your trademark stays unchallenged or gets attacked.

The three dimensions of sign similarity

Signs can be similar on three levels, and even one can be enough. Case law and the EUIPO Trade Mark Guidelines (2025) distinguish aural, visual and conceptual similarity. Each level feeds into the overall impression. How much weight it carries depends on how the goods are usually bought and perceived.

To make the levels tangible, I use three made-up example pairs. They are deliberately fictional and do not assess any real marks, they only show the principle.

Aural similarity

Aural similarity concerns the spoken name. The fictional pair "Velouté" and "Veluta" is spelled differently, but sounds almost the same when spoken aloud. For goods that are ordered or recommended verbally, sound carries a lot of weight. A single letter in the spelling changes little when the ear hears no difference.

Visual similarity

Visual similarity concerns the visual word image. The fictional pair "Sonoro" and "Sonora" differs only in the last letter, yet looks nearly identical on a shelf or a screen. Length, opening letters and syllable structure shape the image. Consumers usually remember the beginning of a word more strongly than the ending.

Conceptual similarity

Conceptual similarity concerns the meaning. The fictional pair "Nordwind" and "Northwind" is spelled and pronounced differently, but carries the same sense. A word of caution here: a purely conceptual similarity is rarely enough on its own to establish a likelihood of confusion, according to CJEU Sabèl/Puma (1997). It does, however, reinforce the weight of the other levels.

How similar are the goods and services?

The second axis is the similarity of goods and services, and it is more than the Nice class. CJEU Canon (C-39/97, 1998) sets out concrete criteria: the nature and characteristics of the goods, their intended purpose, their use, and whether they complement or compete with one another. The distribution channel and the target audience count too.

A common misconception goes like this: same Nice class means a conflict, different class means you are safe. Neither is true. The Nice Classification only sorts goods for administrative purposes. It says little about whether two offerings actually come close in the market. Similar goods can sit in different classes, and dissimilar goods in the same one.

An everyday example: clothing and shoes appear in different contexts, but complement each other in retail and in how they are used. Cleaning services and cleaning products, in turn, can complement each other even though one is a service and the other a product. What matters is the economic connection, not the number.

For your own assessment, that means: compare the real offerings behind the signs, not the class numbers. Which classes even come into question for you is explained in the guide Nice classes explained. And when you want to put concrete signs side by side, you can search your trademark and compare the goods directly.

Why distinctiveness can tip the balance

The distinctiveness of the earlier mark controls how far its protection reaches. The higher it is, the broader the scope of protection, according to CJEU Sabèl/Puma (1997), and the more likely a likelihood of confusion is affirmed. Distinctiveness describes how strongly a sign identifies the origin of the goods, whether inherently or through acquired reputation.

Strong marks are either fanciful and unrelated to the goods, or they have made a name for themselves through intensive use in the market. An invented coined word for a software solution is strong. Such marks cast a wide protective shadow: even clearly different signs can come too close to them.

Weak signs sit at the other end. Anyone who picks a descriptive or suggestive mark, for example a term that almost names the goods, gets only narrow protection. With such signs, even small differences move you out of a likelihood of confusion, because the descriptive core has to stay free for everyone.

This is an often overlooked insight for your own strategy: a particularly obvious, descriptive desired mark protects you worse against copycats later on, even though it feels catchy at first. Anyone who chooses a distinctive sign from the start builds a broader scope of protection.

Interdependence: why you must never assess the factors in isolation

The single most important principle is the interdependence of the factors. A lower degree of similarity between the goods can, according to CJEU Canon (C-39/97, 1998), be offset by a higher degree of sign similarity, and the other way around. That is exactly why no factor on its own produces a verdict. Only the interplay decides.

Picture it like a scale with two pans. On one sits the similarity of the signs, on the other the similarity of the goods. If the signs are almost identical, even a moderate closeness of the goods is enough for a conflict. If the goods are practically level, a lower closeness of the signs is enough in return. The distinctiveness of the earlier mark amplifies or dampens this swing on top.

This principle is not a single opinion. It shapes the settled case law of the CJEU, it appears in the settled case law of the German Federal Court of Justice (BGH), and it structures the EUIPO Trade Mark Guidelines (2025) on the assessment under Art. 8(1)(b) EUTMR. Anyone who wants to assess the likelihood of confusion therefore always thinks in combinations, never in single values.

An aurally identical name in an unrelated industry is often more harmless than a slightly altered name in the same market. It is not one factor that decides, but their interplay.

The most common thinking error in self-assessment sits exactly here: people check the sign first, spot a small difference and sound the all-clear. Or they only look at the class and feel safe. Both are looking at one pan of the scale in isolation. Interdependence forces you to keep both pans in view at the same time.

A simple self-check before you file

A structured self-check mirrors the factors that the offices and courts weigh. It does not replace a legal assessment, because under § 9 MarkenG the likelihood of confusion remains a matter of judgment in each individual case. As an orderly first pass, though, it helps you separate risky hits from harmless ones before you put money into the application.

Work through the factors in this order. Answer each question for every relevant hit from your search.

1. Compare sign similarity: Are the signs similar aurally, visually or conceptually? Say both names out loud, write them one below the other, ask what they mean. Pay particular attention to the start of the word and the overall impression.
2. Compare similarity of goods: Are the goods or services similar in nature, purpose and use, do they complement or compete with one another? Compare the real offering, not the Nice number.
3. Place distinctiveness: Is the earlier mark fanciful and well known, and therefore strong, or descriptive and weak? Strong marks have a wider scope of protection.
4. Build the interdependence: Bring the axes together. High closeness on both axes is a clear warning sign, high closeness on one axis can be offset by the other.

Document every hit with its sign, goods, classes and status. This list is worth its weight in gold when you adjust your naming choice or later talk to a specialized attorney. It turns a vague gut feeling into a basis you can follow and explain.

An honest word about the limits of this check: it organizes, it does not judge. Whether a likelihood of confusion exists in the legal sense is decided in the end by a legal assessment of the individual case. For genuine borderline cases, high investments or a strong hit in the same market, the mark should be checked by an attorney before you file. How to put the hits together technically is shown in the EUIPO trademark search.

Conclusion

The most important takeaway first: a likelihood of confusion never comes from a single factor. According to § 14 MarkenG and CJEU Canon (1998), it requires sign similarity and similarity of goods in interplay, weighted by the distinctiveness of the earlier mark. Signs can be similar aurally, visually or conceptually, though meaning alone is rarely enough. You compare goods by nature and purpose, not by the Nice number. And interdependence forces you to always see both axes together. Think in combinations, document every hit and get legal advice when you have genuine doubts, because this guide is no substitute for it. Anyone who knows the factors files with open eyes. Put your desired mark up against the hits now and have it searched.