Factors and Fictions: The Empirical Collapse of the Likelihood-of-Confusion Test Across the Federal Courts

Abstract

Two factors predict 93.50% of federal trademark confusion outcomes.

Across 3,461 comparisons drawn from nearly 11,000 opinions spanning 1970 to 2025, the bottom-line confusion finding tracks similarity of the marks and competitive proximity of the goods or services. The remaining factors matter, but as structured exceptions: narrow, directionally asymmetric overrides that flip the two-factor prediction in identifiable patterns. The pattern holds in every circuit, across five decades, in district and appellate courts, and in the TTAB’s parallel DuPont framework.

The finding reframes the multifactor test. What doctrine describes as flexible, open-ended balancing is, in operation, a two-factor gate with bounded exceptions. The elaborate apparatus that fills federal opinions and drives millions of dollars in annual litigation costs is largely confirmatory rather than independently decisive. The pattern also speaks beyond trademark: multifactor balancing tests pervade American law, and this Article supplies a factor-level census of what one of them actually does at scale, along with a method for auditing the others.

Courts have already separated the factors from the fictions. This Article proposes that doctrine catch up. Acknowledging the hierarchy courts already apply would reduce cost, improve predictability, focus discovery, briefing, and opinions on the factors most likely to matter, and bring the official account into alignment with observed practice.

Keywords: trademark, likelihood of confusion, multifactor test, trademark infringement, trademark confusion, mark similarity, goods relatedness, services relatedness, trade dress, Lanham Act, DuPont, Sleekcraft, Polaroid, Lapp, Digits of Confusion, Frisch, Interspace, SquirtCo, federal circuit, TTAB, USPTO, trademark registration, trademark opposition, trademark cancellation, multifactor collapse, multifactor balancing, judicial decision-making, doctrinal reform, predictive, balancing test, totality of the circumstances, intellectual property, unfair competition, consumer protection, brand protection, trademark litigation, legal scholarship, law and technology, computational law

JEL Classification: K11, K41, C25, C38, C81, O34, K40