Likelihood of confusion is the reason most trademark applications get refused and the reason most trademark lawsuits exist. The framework for deciding it comes from a 1973 case, In re E. I. du Pont de Nemours & Co., which set out thirteen factors. Almost nobody needs all thirteen.
The useful thing to understand is that the factors are not a checklist and not a score. Any one of them can decide a case on its own, and in the great majority of matters two do nearly all the work.
The two that usually decide it
Similarity of the marks in appearance, sound, connotation and overall commercial impression. Not identity: similarity. The marks are compared as wholes rather than dissected, though within a composite the distinctive element carries more weight than a descriptive one.
Relatedness of the goods or services. The goods do not have to be the same, or even competitive. They have to be related enough that buyers encountering similar marks would assume a common source. This is judged on the identifications as written in the application and registration, not on what the businesses happen to sell today.
These two trade against each other. Very similar marks can coexist on genuinely unrelated goods. Modestly similar marks collide on competing goods. That trade-off is most of the analysis.
The ones that move the result
Trade channels. Where the goods travel and to whom. Two products in the same aisle, the same catalogue, or the same search results page are closer than their descriptions suggest.
Buyer care. Goods bought quickly and cheaply are judged strictly, because nobody studies a two-dollar purchase. Expensive or specialist goods bought by professionals after deliberation get more room, though the Board has repeatedly warned that even careful buyers are not immune to confusion when the marks are close.
Strength of the earlier mark. A coined mark gets a wide berth. A mark built from ordinary words that many businesses already use gets a narrow one. Fame, at the top end, expands protection considerably.
The number of similar marks already in use on similar goods. This is the crowded-field factor, and it is the most useful one for an applicant, because it cuts the earlier owner scope down. If a dozen registrations already share the term, buyers are used to telling them apart by small differences, and your small difference starts to count. See can you trademark a common word.
Actual confusion. Powerful when it exists and rarely available. Its absence proves much less than defendants hope, because meaningful absence requires a long period of genuine side-by-side use.
What examination actually looks like
There is an important practical asymmetry. In examination the examining attorney has the register and little else. There are no customers to survey, no sales figures, no evidence of how the two businesses actually coexist. So the refusal you receive is built almost entirely from the first two factors, read off the identifications.
That is why a refusal is often answerable. The examiner is working from a document, and you know things about the market that the document does not say. Evidence of a crowded field, of different trade channels, or of the genuine distance between the goods is exactly what is missing from the file, and it is the applicant job to put it there.
Litigation is the opposite. In court all thirteen factors are live, survey evidence appears, and intent and actual confusion can dominate.
Things people get wrong
- Counting factors. There is no threshold and no tally. One factor can carry the whole decision.
- Relying on class numbers. Different classes do not mean unrelated goods. Classes are a filing convenience. See the different-industry question and what a class actually decides.
- Adding a word. Adding a descriptive term to somebody’s distinctive mark usually makes things worse, because it reads as a product line from that owner.
- Trusting a disclaimer. Disclaiming a descriptive portion does not remove it from the comparison; the marks are still compared as wholes.
- Assuming a logo saves a word. In a composite mark the words normally dominate, because words are what people say and search.
How to use this before you file
Run the two main factors yourself on every near miss your search turns up. How close are the marks in sound, appearance and meaning? How related are the goods, judged by trade channels and buyers rather than by class? If both answers are uncomfortable, that is your answer. If one is comfortable and the other is not, you are in the space where the other factors decide, and that is the point at which a conversation with a trademark attorney is worth what it costs.