Avoiding Confusion: How KIPO Evaluates Similarity under Article 34(1)7

One of the most frequent grounds for trade mark refusal in South Korea is Article 34(1)(7) of the Korean Trademark Act (KTA), which prohibits the registration of a mark that is identical or similar to an earlier pending application or registered mark for the same or similar goods or services. According to KIPO’s statistics, over 40% of all trade mark office actions in 2022 were based on this ground. Understanding how KIPO evaluates similarity — and particularly the weight given to phonetic similarity — is essential for avoiding costly refusals and building a successful filing strategy in Korea.

The Three-Dimensional Similarity Test

KIPO assesses trade mark similarity through a well-established three-dimensional analysis, examining marks across three independent dimensions:

• Visual similarity (외관): The overall appearance of the marks when seen side by side, including their structure, layout, design elements, and spatial arrangement.
• Phonetic similarity (칭호): How the marks sound when spoken aloud, including syllable structure, vowel and consonant patterns, intonation, and rhythm.
• Conceptual similarity (관념): The meaning or idea conveyed by each mark, including direct translations, connotations, and the mental images they evoke.

A critical principle in Korean practice is that similarity in any one of these three dimensions is sufficient to establish overall similarity. This means a mark that is visually different from a prior mark can still be refused if it is phonetically or conceptually similar. This distinguishes Korea from some jurisdictions that require a more balanced, holistic assessment.

The Dominant Element Principle

When comparing composite marks (marks consisting of multiple elements), KIPO applies the dominant element principle. The comparison focuses on the element or elements that consumers are most likely to use to identify and refer to the mark. If the dominant element of the applied mark is similar to the dominant element of the prior mark, overall similarity may be found even if the non-dominant elements differ significantly.

For word-and-design marks, the word element is typically treated as dominant because consumers tend to recall and refer to marks by their verbal component. This means that a distinctive logo may not prevent a similarity finding if the word elements of the two marks are similar.

Phonetic Similarity: The Critical Dimension for Foreign Brands

Phonetic similarity receives exceptionally high weight in Korean trade mark practice, and this is where international brands face the greatest risk. The reason lies in the mechanics of the Korean writing system, Hangul, and its interaction with foreign-language marks.

When Korean consumers encounter a foreign word mark, they naturally transliterate it into Hangul for pronunciation. KIPO’s examiners do the same when assessing similarity. The comparison is therefore made not between the original Roman-script marks but between their Hangul transliterations. Two marks that look quite different in Roman script can produce Hangul transliterations that are phonetically very close, leading to a similarity finding.

For example, marks like “CLEAN” (클린) and “KLEAN” (클린) may produce identical Hangul renderings, resulting in a finding of phonetic identity. Similarly, marks that are phonetically distinct in English may converge in Hangul because the Korean alphabet does not distinguish certain sounds (such as “r” and “l,” or “f” and “p”) in the same way English does.

This Hangul convergence effect is one of the most significant practical challenges for foreign brands filing in Korea, and one that is frequently underestimated by applicants unfamiliar with the Korean language system.

Goods and Services Similarity

Similarity under Article 34(1)(7) requires not only mark similarity but also similarity of goods or services. KIPO uses the Nice Classification system and maintains internal similarity groupings that determine which goods and services are considered similar. Goods within the same similarity group are presumed to be similar, while goods in different groups are presumed to be dissimilar (though this presumption can be rebutted).

Korea’s similarity groupings do not always match international expectations. Applicants should review KIPO’s specific groupings rather than relying on similarity assessments from other jurisdictions.

The Role of Distinctiveness in the Comparison

The degree of distinctiveness of the prior mark affects the scope of comparison. Highly distinctive marks (coined words, arbitrary marks) receive a broader scope of protection, meaning that even marks with more distant similarity may be caught. Weakly distinctive marks (descriptive or common elements) receive a narrower scope, and closer similarity is required for a refusal.

This means that applicants facing a citation of a weak prior mark have stronger arguments for distinguishing their application, while applicants facing a citation of a strong, coined prior mark face a more challenging task.

The Opposition Process

If KIPO finds no grounds for refusal, the application is published in KIPO’s official Gazette. Third parties have two months from publication to file an opposition, and may amend the grounds and evidence for up to two months and 30 days from filing. The opposition period cannot be extended, but the amendment period can be extended once or twice for up to 30 additional days each. If KIPO accepts the opposition, it simultaneously issues a refusal decision, which can be appealed to the Intellectual Property Trial and Appeal Board (IPTAB).

The New Consent Agreement Option

Since May 1, 2024, South Korea accepts letters of consent as a mechanism to overcome similarity refusals under Article 34(1)(7). This major reform, discussed in detail in a separate article, provides a powerful new tool for applicants facing prior mark citations. However, the consent system does not apply when the marks and goods are both identical.

Strategic Recommendations

• Conduct a comprehensive clearance search before filing: Use the KIPRIS database to search not only Roman-script marks but also Hangul transliterations of your proposed mark. This captures phonetic conflicts that a Roman-only search would miss.
• Assess the Hangul rendering of your mark: Before filing, work with a Korean trade mark professional to determine how your mark will be transliterated into Hangul and whether the rendering creates conflicts with existing marks.
• Focus on phonetic differentiation: Given the high weight KIPO assigns to phonetic similarity, prioritise creating phonetic distance from prior marks. Visual and conceptual differences alone may not be sufficient.
• Use the consent system where available: If a similarity refusal is issued and the prior mark owner is willing to coexist, a letter of consent can now overcome the refusal (except for identical marks on identical goods).
• Consider the dominant element carefully: If filing a composite mark, ensure the dominant element is sufficiently different from prior marks. A distinctive logo will not save a phonetically similar word element.

Common Mistakes

• Searching only in Roman script: KIPO compares marks in Hangul. A search limited to Roman characters will miss phonetic conflicts.
• Underestimating phonetic similarity: Two marks that appear visually different can produce identical or near-identical Hangul transliterations.
• Assuming visual differences will overcome phonetic similarity: KIPO can refuse on phonetic grounds alone, regardless of visual differences.
• Relying on foreign-market coexistence: The fact that two marks coexist in another jurisdiction does not mean KIPO will reach the same conclusion. Korean phonetic assessment is unique.

Key Takeaway

KIPO’s three-dimensional similarity test — visual, phonetic, and conceptual — is applied rigorously, and similarity in any single dimension is sufficient for refusal. For international brands, the most critical dimension is phonetic similarity as assessed through Hangul transliteration. A mark that sounds different in English may sound identical in Korean, creating unexpected conflicts. The introduction of the consent agreement system in 2024 provides a valuable new tool for resolving conflicts, but the best strategy remains a thorough pre-filing search that accounts for the Korean language dimension.