United States of America

United States Court of Appeals, Federal Circuit
Converse, Inc. v. International Trade Commission Skechers U.S.A., Inc.

909 F.3d 1110

2018

DYK, Circuit Judge.

Converse, Inc., appeals from a final determination of the International Trade Commission ("ITC") that held invalid Converse's trademark in the midsole design of its Chuck Taylor All Star shoes, U.S. Trademark Registration No. 4,398,753 ("the '753 trademark"). Because it found the registered mark invalid and that Converse could not establish the existence of common-law trademark rights, the ITC determined there was no violation of section 337 of the Tariff Act of 1930, as amended, 19 U.S.C. § 1337 (2012), by the importation of the accused products. The ITC nonetheless addressed infringement, finding that various accused products would have infringed Converse's mark if valid.

We hold that the ITC erred in applying the wrong standard in aspects of both its invalidity and infringement determinations. We vacate and remand for further proceedings.

BACKGROUND

This case involves alleged infringement of Converse's rights in trade dress arising from the common law and its trademark registration. The '753 trademark was issued to Converse on September 10, 2013, and describes the trade-dress configuration of three design elements on the midsole of Converse's All Star shoes. In particular, as described in the registration, "the mark consists of the design of the two stripes on the midsole of the shoe, the design of the toe cap, the design of the multi-layered toe bumper featuring diamonds and line patterns, and the relative position of these elements to each other." The mark is depicted in a single drawing in the registration:

Converse asserts common-law rights in the same mark predating its registration.

Section 337 provides a remedy at the ITC for, among other things, "[t]he importation into the United States, the sale for importation, or the sale within the United States after importation by the owner, importer, or consignee, of articles that infringe a valid and enforceable United States trademark registered under the Trademark Act of 1946." 19 U.S.C. § 1337(a)(1)(C). On October 14, 2014, Converse filed a complaint with the ITC alleging violations of section 337 by various respondents in the importation into the United States, the sale for importation, and the sale within the United States after importation of shoes that infringe its trademark. The ITC instituted an investigation on November 17, 2014. Although some of the respondents defaulted, several appeared and actively participated in the ITC proceedings, asserting that the accused products did not infringe the mark and that, in any event, it was invalid. These respondents have now intervened in Converse's appeal and are referred to herein as the intervenors.

A central issue was whether the mark had acquired secondary meaning. Converse asserted that it had acquired secondary meaning, the mark having been used by Converse since 1932. The intervenors, on the other hand, disputed secondary meaning, claiming that Converse's use of the mark had not been substantially exclusive and offering a survey (the Butler survey) concluding that consumers did not associate the Converse mark with a single source. The parties also disputed infringement. Both the ITC Administrative Law Judge ("ALJ") and the ITC treated Converse as claiming two separate marks—a common-law mark and a registered mark.

On November 17, 2015, the ALJ issued an initial determination finding violations of section 337 by the intervenors because the registered '753 trademark was infringed and not invalid, relying on the presumption of secondary meaning afforded to the registered mark. However, the ALJ found that Converse had not established secondary meaning for the common-law mark (but that, if protectable, the common-law mark was infringed). Converse, the intervenors, and the ITC staff petitioned for review.

On June 23, 2016, the ITC issued its final determination. The ITC reversed the ALJ's finding of no invalidity of the registered mark. The ITC found the registered mark invalid in light of its determination that the mark had not acquired secondary meaning. With respect to the common-law mark, the ITC affirmed the ALJ's finding that the mark had not acquired secondary meaning. The ITC determined that, if either trademark was not invalid or protectable, it was infringed, affirming the ALJ's finding in this respect. The ITC refused to enter an exclusion order with respect to any of the respondents, including those who had defaulted. Converse timely appealed, and we have jurisdiction under 28 U.S.C. § 1295(a)(6).

The court held oral argument on February 8, 2018. On June 7, 2018, the court requested supplemental briefing on the following questions:

1. Was Converse required to show priority in the mark (i.e., secondary meaning at the time of first infringement) without regard to the presumption of validity that would exist if the trademark registration is valid?

2. What significance does the registration of the mark or its validity have in these proceedings?

3. Was it necessary or appropriate for the ITC to address the validity of the registered mark?

Converse, Inc. v. Int'l Trade Comm'n, 726 F. App'x 818, 819 (Fed. Cir. 2018) (per curiam) (nonprecedential order). Each of the parties filed supplemental briefs in response.

DISCUSSION

We review the ITC's legal determinations de novo and its factual findings for substantial evidence. Cisco Sys., Inc. v. Int'l Trade Comm'n, 873 F.3d 1354, 1360-61 (Fed. Cir. 2017). We conclude that the ITC made a series of errors that require a remand. In Part I, we discuss the relevant date for assessing secondary meaning, the significance of Converse's trademark registration, and the benefits arising from that registration. In Part II, we define the factors to be weighed in determining whether a mark has acquired secondary meaning. And in Part III, we address the standard for evaluating likelihood of confusion for the purposes of determining infringement.

I. The Timing of the Secondary Meaning Inquiry and the Relevance of Trademark Registration

The ITC's first error was failing to distinguish between alleged infringers who began infringing before Converse obtained its trademark registration and those who began afterward. This error was not identified as such in Converse's briefing, no doubt because the error was beneficial to Converse. The intervenors argued in their principal and supplemental briefs that the ITC erred in this respect.

In addressing these issues, we think that it is confusing and inaccurate to refer to two separate marks—a registered mark and a common-law mark. Rather, there is a single mark, as to which different rights attach from the common law and from federal registration. E.g., In re Int'l Flavors & Fragrances Inc., 183 F.3d 1361, 1366 (Fed. Cir. 1999) ("The federal registration of a trademark does not create an exclusive property right in the mark. The owner of the mark already has the property right established by prior use .... However, those trademark owners who register their marks with the [Patent and Trademark Office (`PTO') ] are afforded additional protection not provided by the common law."); In re Deister Concentrator Co., 289 F.2d 496, 501 (CCPA 1961) ("[T]he Lanham Act does not create trademarks. While it may create some new substantive rights in trademarks, unless the trademarks pre-exist there is nothing to be registered. Neither does it create ownership, but only evidence thereof."); 3 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition, § 19:3 (5th ed. 2017 & Supp. 2018) ("Although a federal registration gives the owner of a mark very important and valuable legal rights and benefits, the registration does not create the trademark."); see also Matal v. Tam, ___ U.S. ___, 137 S.Ct. 1744, 1751, 198 L.Ed.2d 366 (2017) (quoting B & B Hardware, Inc. v. Hargis Indus., Inc., ___ U.S. ___, 135 S.Ct. 1293, 1299, 191 L.Ed.2d 222 (2015)) ("[F]ederal law does not create trademarks." (alteration in original) ).

Converse secured trademark registration for its trade dress on September 10, 2013. Converse alleges that before and after the date of registration the respondents infringed that mark. To establish infringement of a trademark under the Lanham Act, Converse must prove "(1) it has a valid and legally protectable mark; (2) it owns the mark; and (3) the defendant's use of the mark to identify goods or services causes a likelihood of confusion." ERBE Elektromedizin GmbH v. Canady Tech. LLC, 629 F.3d 1278, 1287 (Fed. Cir. 2010) (quoting A & H Sportswear, Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198, 210 (3d Cir. 2000) ).

All trademarks, in order to be valid or protectable, must be distinctive of a product's source, and "courts have held that a mark can be distinctive in one of two ways." Wal-Mart Stores, Inc. v. Samara Bros., Inc., 529 U.S. 205, 210, 120 S.Ct. 1339, 146 L.Ed.2d 182 (2000). "First, a mark is inherently distinctive if `[its] intrinsic nature serves to identify a particular source.'" Id. (alteration in original) (quoting Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992)). "Second, a mark has acquired distinctiveness, even if it is not inherently distinctive, if it has developed secondary meaning, which occurs when, `in the minds of the public, the primary significance of a [mark] is to identify the source of the product rather than the product itself.'" Id. at 211, 120 S.Ct. 1339 (alteration in original) (quoting Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844, 851 n.11, 102 S.Ct. 2182, 72 L.Ed.2d 606 (1982)); see also 15 U.S.C. § 1052(f).

Converse here seeks protection as to the intervenors for a mark in the form of unregistered product-design trade dress. The Supreme Court has held that unlike word marks and product-packaging trade dress, product-design trade dress can never be inherently distinctive. Wal-Mart, 529 U.S. at 216, 120 S.Ct. 1339. As a result, "a product's design is distinctive, and therefore protectable, only upon a showing of secondary meaning." Id. Accordingly, Converse must show that its mark has acquired distinctiveness, i.e., secondary meaning.

The ITC's decision never determined the relevant date for assessing the existence of secondary meaning. The ITC utilized 2003 (characterized as the "date of first infringement") as the relevant date in certain instances, but also extensively cited to evidence from after 2003. On appeal, the ITC contends that its finding should be read to mean that Converse's trademark had not acquired secondary meaning at any time.

Because the relevant date is so important to the secondary-meaning analysis, we find that a specific determination of secondary meaning as of the relevant date must be made. In any infringement action, the party asserting trade-dress protection must establish that its mark had acquired secondary meaning before the first infringing use by each alleged infringer. See, e.g., Braun, Inc. v. Dynamics Corp. of Am., 975 F.2d 815, 826 (Fed. Cir. 1992) (holding that "[a] claim of trade dress infringement fails if secondary meaning did not exist before the infringement began" and placing the burden of proof on the plaintiff); 2 McCarthy, supra, § 16:34 (noting that the purported "senior user must prove the existence of secondary meaning in its mark at the time and place that the junior user first began use of that mark" and collecting cases); Restatement (Third) of Unfair Competition § 19 cmt. b. (Am. Law Inst. 1995 & Supp. 2018). In this respect, Converse argues that it is entitled to rely on the presumption of validity afforded to registered marks. We do not agree that this presumption applies to infringement that began before registration.

For infringement in the period after registration, the Lanham Act entitles the owner of the registered mark to a presumption that the mark is valid, see 15 U.S.C. §§ 1057(b), 1115(a), including that it has acquired secondary meaning, see 2 McCarthy, supra, § 11:43 (citing Lovely Skin, Inc. v. Ishtar Skin Care Prods., LLC, 745 F.3d 877, 882 (8th Cir. 2014); Cold War Museum, Inc. v. Cold War Air Museum, Inc., 586 F.3d 1352, 1356 (Fed. Cir. 2009) ); 2 McCarthy, supra, § 15:34; 6 McCarthy, supra, § 32:134. In the context of cancellation proceedings, we have held that this presumption shifts both the burden of persuasion and the initial burden of production to the challenger to rebut the presumption. Cold War, 586 F.3d at 1358; see also Schaffer ex rel. Schaffer v. Weast, 546 U.S. 49, 56, 126 S.Ct. 528, 163 L.Ed.2d 387 (2005) (describing the difference between these "two distinct burdens"). We see no reason why the effect of the presumption should be any different in the infringement context, and we join with the majority of circuits that have held that the presumption shifts both burdens to the party challenging secondary meaning. See 6 McCarthy, supra, § 32:138 n.12 (collecting cases); Restatement (Third) of Unfair Competition § 13 cmt. a. This result is strongly supported by the legislative history of the Lanham Act as thoroughly documented by Charles L. Cook and Theodore H. Davis Jr., Litigating the Meaning of "Prima Facie Evidence" Under the Lanham Act: The Fog and Art of War, 103 Trademark Rep. 437, 459-86 (2013) (showing that Congress repeatedly considered this question and expressly chose to shift the burden of persuasion).

However, under the statute, the registration and its accompanying presumption of secondary meaning operate only prospectively from the date of registration, i.e., the date on which the Patent and Trademark Office ("PTO") determined that secondary meaning had been acquired. See 15 U.S.C. §§ 1052(f), 1057(a)-(b), 1115(a). This was recognized by the Eighth Circuit in Aromatique, Inc. v. Gold Seal, Inc., 28 F.3d 863, 870 (8th Cir. 1994)[1] and has been approved by the relevant treatises. See 7 Louis Altman & Malla Pollack, Callmann on Unfair Competition, Trademarks and Monopolies § 26:101 (4th ed. 2012 & Supp. 2018); 2 McCarthy, supra, §§ 11:53, 15:34.

This rule is a straightforward application of the Lanham Act, pursuant to which the PTO examines whether secondary meaning has been acquired at the time of registration. See 15 U.S.C. § 1052(f). The resulting registration confers a presumption of secondary meaning from that point in time, see id. §§ 1057(b), 1115(a), but at the time of registration the PTO is not asked to determine whether secondary meaning had been acquired at some previous date, and therefore registration cannot support a presumption for the period before registration. Indeed, on a record such as this, with a multiyear gap between infringement and registration, registration cannot even be probative of secondary meaning at the time of infringement.

Converse argued to the ITC that its registration entitled it to a presumption of secondary meaning at any and all times, including before registration. But Converse identified no circuit court decisions in which this question was squarely presented and decided in favor of Converse's position, and it relies on legislative history that predates the Lanham Act by two decades and fails to address the question at issue.

We conclude that Converse's registration confers a presumption of secondary meaning beginning only as of the date of registration and confers no presumption of secondary meaning before the date of registration. Thus, with respect to infringement by those respondents whose first uses came before the registration (including all of the intervenors), Converse must establish without the benefit of the presumption that its mark had acquired secondary meaning before the first infringing use by each respondent.

The intervenors contend that Converse has waived any argument that it could prevail on its claims of pre-registration infringement without the benefit of a presumption of secondary meaning. We do not think a finding of waiver is appropriate here, given that our opinion is clarifying and in some ways changing the legal landscape with respect to proving secondary meaning. Thus, we conclude that on remand, Converse has not waived the argument that its mark acquired secondary meaning even before the date of registration under the appropriate standards, an issue we discuss in Part II below.

The question remains whether the issue of trademark validity needs to be addressed by the ITC on remand. The dissent argues that since the intervenors' first infringement in all cases began before registration and the remaining respondents defaulted, trademark validity need not be addressed in granting relief. In this respect the dissent relies on the provisions of 19 U.S.C. § 1337.[2] Dissenting Op. at 1132-33. We decline to decide that issue at this stage since it has not been addressed by either the ITC or the parties. But Converse and the intervenors assume that on remand further issues of validity remain because the order sought is a general exclusion order. Converse Suppl. Br. 11-12 ("The ITC also needed to address Converse's registered trademark when evaluating Converse's infringement claims as to infringements that began after Converse registered the trademark. For those infringements, as discussed above, Converse may rely on its federal registration's presumption of secondary meaning to satisfy the first element of its infringement claims."); Intervenors' Suppl. Br. 12 ("With respect to parties other than the Intervenors who may be affected by a GEO, it was both necessary and appropriate for the ITC to consider the validity of the '753 Registration"). On their face, the 1988 amendments to the Tariff Act of 1930 (now section 1337(g) ) only appear to authorize the entry of an exclusion order "limited to that person," i.e., the defaulting party and not a general exclusion order.[3] Whether the 1988 amendments, as the dissent urges, require the entry of a general exclusion order without addressing trademark validity or infringement is best addressed on remand.[4]

If validity remains an issue, of course, the secondary-meaning analysis should look to the date of registration. See McCormick & Co. v. Summers, 354 F.2d 668, 674 (CCPA 1966) ("[R]egistrability of a mark must be determined on the basis of facts as they exist at the time when the issue of registrability is under consideration."). In this case, that date was September 10, 2013.

II. The Standards for Determining Whether A Mark Has Acquired Secondary Meaning

Converse argues that the ITC erred in its secondary meaning analysis in a number of respects. We agree with Converse in part, concluding that in some of the claimed respects the ITC applied the wrong legal standard in its determination of secondary meaning.

A

We first address the relevant factors. In assessing whether the '753 trademark had acquired secondary meaning, the ITC weighed seven factors: "(1) the degree and manner of use; (2) the exclusivity of use; (3) the length of use; (4) the degree and manner of sales, advertising, and promotional activities; (5) the effectiveness of the effort to create secondary meaning; (6) deliberate copying; and (7) association of the trade dress with a particular source by actual purchasers (typically measured by customer surveys)." J.A. 21. The ITC affirmed the ALJ's determination that factors 1, 3, 4, and 6 supported a finding of secondary meaning; that factor 5 was neutral; and that factor 7 weighed against such a finding. The ITC determined that the ALJ had erred, however, in finding factor 2 (exclusivity of use) to be neutral; instead, the ITC found that it weighed against a finding of secondary meaning. Weighing all these factors, the ITC then determined that the '753 trademark had not acquired secondary meaning and was, therefore, invalid.

Each circuit that has addressed second-ary meaning—11 circuits in all—has formulated some version of a multifactor test similar to the test adopted by the ITC.[5] This Court has previously discussed certain factors that are relevant to the analysis, which overlap to an extent with those identified by the ITC. See, e.g., Coach Servs., Inc. v. Triumph Learning LLC, 668 F.3d 1356, 1379 (Fed. Cir. 2012) ("To determine whether a mark has acquired secondary meaning, courts consider: advertising expenditures and sales success; length and exclusivity of use; unsolicited media coverage; copying of the mark by the defendant; and consumer studies."); see also Real Foods Pty Ltd. v. Frito-Lay N. Am., Inc., 906 F.3d 965, 978 (Fed. Cir. 2018) (quoting Coach for same proposition); In re Steelbuilding.com, 415 F.3d 1293, 1300 (Fed. Cir. 2005) ("In determining whether secondary meaning has been acquired, the Board may examine copying, advertising expenditures, sales success, length and exclusivity of use, unsolicited media coverage, and consumer studies (linking the name to a source).").

Today we clarify that the considerations to be assessed in determining whether a mark has acquired secondary meaning can be described by the following six factors: (1) association of the trade dress with a particular source by actual purchasers (typically measured by customer surveys); (2) length, degree, and exclusivity of use; (3) amount and manner of advertising; (4) amount of sales and number of customers; (5) intentional copying; and (6) unsolicited media coverage of the product embodying the mark. While the ITC's test set forth length, degree, and exclusivity of use as separate factors, we think that these considerations are substantially interrelated and should be evaluated together. All six factors are to be weighed together in determining the existence of secondary meaning.

B

Next, we address the significance of the trademark owner's and third parties' prior uses of the mark. We conclude that the ITC relied too heavily on prior uses long predating the first infringing uses and the date of registration. The secondary meaning analysis primarily seeks to determine what is in the minds of consumers as of the relevant date,[6] and factor 2 must be applied with this purpose in view. The most relevant evidence will be the trademark owner's and third parties' use in the recent period before first use or infringement.

The Lanham Act itself sheds light on what constitutes the most relevant period. Section 2(f) provides that in assessing secondary meaning:

The Director [of the PTO] may accept as prima facie evidence that a mark has become distinctive, as used on or in connection with the applicant's goods in commerce, proof of substantially exclusive and continuous use thereof as a mark by the applicant in commerce for the five years before the date on which the claim of distinctiveness is made.

15 U.S.C. § 1052(f) (emphasis added). A somewhat different, ten-year rule had previously been enacted in 1905. See Act of Feb. 20, 1905, ch. 592, § 5, 33 Stat. 724, 726. Describing marks subject to that 1905 law, the Supreme Court noted, "Their exclusive use as trademarks for the stated period was deemed, in the judgment of Congress, a sufficient assurance that they had acquired a secondary meaning as the designation of the origin or ownership of the merchandise to which they were affixed." Thaddeus Davids Co. v. Davids, 233 U.S. 461, 470, 34 S.Ct. 648, 58 L.Ed. 1046 (1914).

Today's five-year rule was enacted in substantially its present form in 1946. See Act of July 5, 1946 (Lanham Act), ch. 540, § 2(f), 60 Stat. 427, 429. Remarking on the shift from ten to five years and relying on Davids, the Eighth Circuit found "it must follow that Congress in establishing the new and different prerequisite of five years ... now deems that period of such use adequate." Curtis-Stephens-Embry Co. v. Pro-Tek-Toe Skate Stop Co., 199 F.2d 407, 413 (8th Cir. 1952).

To be sure, section 2(f) sets up an evidentiary rule for the Director rather than courts. But the Supreme Court has noted "that the general principles qualifying a mark for registration under § 2 of the Lanham Act are for the most part applicable" in other trademark proceedings, such as suits seeking to enforce unregistered marks under 15 U.S.C. § 1125(a). Two Pesos, 505 U.S. at 768, 112 S.Ct. 2753. And several other courts of appeals, drawing on section 2(f), have found five years' substantially exclusive and continuous use to weigh strongly in favor of a finding of secondary meaning. See Thomas & Betts Corp. v. Panduit Corp., 138 F.3d 277, 295-96 (7th Cir. 1998); Sunbeam Prods., Inc. v. W. Bend Co., 123 F.3d 246, 255 (5th Cir. 1997), abrogated on other grounds by Traf-Fix Devices, Inc. v. Mktg. Displays, Inc., 532 U.S. 23, 121 S.Ct. 1255, 149 L.Ed.2d 164 (2001); Stuart Hall, 51 F.3d at 789-90. But see FN Herstal, 838 F.3d at 1084 (declining to rely on section 2(f)).

We agree with those circuits that recognize the importance of looking to this five-year period. While section 2(f) cannot be read as limiting the inquiry to the five years before the relevant date, it can and should be read as suggesting that this period is the most relevant. As a result, in evaluating factor 2, the ITC should rely principally on uses within the last five years. The critical issue for this factor is whether prior uses impacted the perceptions of the consuming public as of the relevant date. See Palm Bay Imps., Inc. v. Veuve Clicquot Ponsardin Maison Fondee en 1772, 396 F.3d 1369, 1373-74 (Fed. Cir. 2005) (making a similar point in the likelihood-of-confusion context). Consumers are more likely to remember and be impacted in their perceptions by third-party uses within five years and less likely with respect to older uses. We recently applied similar, common-sense reasoning in the trademark opposition context, finding that survey results were probative, at best, of the public's perception five years after the survey was conducted. Royal Crown Co. v. Coca-Cola Co., 892 F.3d 1358, 1370 (Fed. Cir. 2018).

Therefore, uses older than five years should only be considered relevant if there is evidence that such uses were likely to have impacted consumers' perceptions of the mark as of the relevant date. For example, this might be the case where a particular advertising campaign has been in use for longer than five years. The ITC's determination relied heavily on evidence— both as to Converse's use and the use by competitors—far predating the relevant timeframe. Evidence older than this five-year period should be reevaluated on remand to determine whether it is relevant.

C

In considering exclusivity of use—that is, whether the trademark owner's use of the mark was substantially exclusive—it appears the ITC considered prior third-party uses of shoes with designs that were not substantially similar to the '753 trademark. The ITC cited historical examples as a "third-party use of the ['753 trademark]" or "shoes bearing the [trademark]." J.A. 28. But several of the instances of third-party use cited in the ITC's decision are shoes that bear at most a passing resemblance to the '753 trademark. Many more are missing at least one of the elements of the '753 trademark. Still others are reproduced in such poor resolution that no reasonable comparison can be made. Examples of prior uses that may not be substantially similar are set forth in the Appendix to this opinion.

Although we agree with the ITC that evidence of the use of similar but not identical trade dress may inform the secondary-meaning analysis, we think such uses must be substantially similar to the asserted mark. See Echo Travel, Inc. v. Travel Assocs., Inc., 870 F.2d 1264, 1266, 1269 (7th Cir. 1989) (weighing prior use of marks described as "substantially similar" and "substantially identical"); Sweetheart Plastics, Inc. v. Detroit Forming, Inc., 743 F.2d 1039, 1045 (4th Cir. 1984) (approving of the admission of evidence of "substantially identical" and "substantially similar" designs as "probative of the extent and nature of exclusivity of use"); Scarves by Vera, Inc. v. Todo Imports Ltd., 544 F.2d 1167, 1174 (2d Cir. 1976) (criticizing reliance on prior uses of marks that were insufficiently similar). Here, the ITC made no determination as to which of the prior uses were substantially similar to the '753 trademark and relied on at least some that were not. On remand, the ITC must constrain its analysis of both Converse's use and the use by its competitors to marks substantially similar to Converse's registered mark.

D

The ITC also placed considerable weight on survey evidence submitted by the intervenors (the CBSC only Butler survey) to support its determination that the mark had not acquired secondary meaning.[7] We agree with amici that surveys are sometimes difficult to use as evidence of historic secondary meaning. Amicus Br. of All Market Inc. 29. This is because "[t]he relevant consumer population for assessing consumer attitudes at a point in the past is a group of consumers at that point in the past" and "[a] contemporaneous survey commissioned for litigation obviously cannot access such a pool of respondents." Id. at 29. Thus, the ITC should only give such survey evidence "weight appropriate to the extent that it sheds light on consumer perceptions in the past." 2 McCarthy, supra, § 16:34.

We think the Butler survey likely has little relevance with respect to the issue of secondary meaning for the intervenors. The intervenors' expert, Sarah Butler, surveyed respondents in the spring of 2015 to determine whether they associated the '753 trademark with a single source. Converse did not dispute that a survey taken two years after the registration is relevant to determining secondary meaning as of the date of registration. On remand, however, the ITC must consider whether Converse's mark had acquired secondary meaning as of each first infringing use by each intervenor, the earliest of which is more than ten years before the date of the Butler survey and the latest of which is likely more than five years before the Butler survey. Thus, with respect to Converse's claims of infringement against the intervenors, the ITC should give the Butler survey little probative weight in its analysis, except to the extent that the Butler survey was within five years of the first infringement by one of the intervenors. See generally Commerce Nat'l. Ins. Servs., Inc. v. Commerce Ins. Agency, Inc., 214 F.3d 432, 440 (3d Cir. 2000) (recency of survey contributed to its being "wholly irrelevant to whether CBI established secondary meaning in the `Commerce' mark as of 1983").

If on remand, secondary meaning at the time of the registration remains an issue, the Butler survey may have relevance since it was conducted within two years of the registration; indeed, Converse does not object to the use of the Butler survey on the ground that it is too distant in time. Surveys that are conducted within five years of the relevant date may provide evidence as to secondary meaning although "[n]o single factor is determinative." In re Steelbuilding.com, 415 F.3d at 1300.

Butler concluded that the '753 trademark had a 21.5% net rate of association with a single source, a rate of association that the ALJ found "insufficient to establish secondary meaning" at the time of registration. J.A. 88. The ITC found that the Butler survey "weighs against a finding of secondary meaning," J.A. 26, relying on the ALJ's conclusion that the survey was "insufficient to establish secondary meaning," J.A. 88. We see no error in the conclusion that the survey does not establish secondary meaning, but we are unclear as to the ITC's reasoning as to why the survey supports the opposite—a lack of secondary meaning. On appeal the intervenors argue that the Butler survey affirmatively supports their position—showing lack of secondary meaning. Converse argues that the Butler survey is flawed and should be given no weight. We do not resolve this issue which is a matter for the ITC in the first instance. In any remand where secondary meaning as of the time of registration is a relevant issue, the Board should analyze whether the survey shows lack of secondary meaning as of the date of registration. Unless the survey affirmatively shows a lack of secondary meaning, there is simply a lack of survey evidence of secondary meaning—which is a neutral factor favoring neither party.

III. The Standards for Determining Likelihood of Confusion

A

The intervenors also contend that the ITC erred in finding a likelihood of confusion —and, therefore, infringement—with respect to accused products that lacked one or more elements of the '753 trademark. For example, the intervenors point out that the ITC found no infringement by an accused product missing one of the '753 trademark's elements (two stripes on the midsole) but did find infringement by two other accused products missing a different element (the multi-layered toe bumper featuring diamonds and line patterns). The ALJ's only explanation for these different results was that, in the former case, "[t]he differences in these shoe models are not drastic enough to overcome the similarities." J.A. 151.

The likelihood-of-confusion analysis for determining infringement turns in part on the similarity of the accused products to the asserted mark. See, e.g., Nautilus Grp., Inc. v. Icon Health & Fitness, Inc., 372 F.3d 1330, 1334-35 (Fed. Cir. 2004). We described earlier that, in the invalidity determination, marks that are not substantially similar cannot be considered. In the context of trade-dress infringement, we also hold that accused products that are not substantially similar cannot infringe. See Versa Prods. Co. v. Bifold Co. (Mfg.), 50 F.3d 189, 202 (3d Cir. 1995) ("[S]ubstantial similarity of appearance is necessarily a prerequisite to a finding of likelihood of confusion in product configuration cases."); see also Eng'g Dynamics, Inc. v. Structural Software, Inc., 26 F.3d 1335, 1350 (5th Cir. 1994) ("The Lanham Act prohibits passing off goods or services as those of a competitor by employing substantially similar trade dress which is likely to confuse consumers as to the sources of the product."), modified on other grounds, 46 F.3d 408 (5th Cir. 1995). We have applied an analogous requirement in the design-patent context, where infringement cannot be found unless an ordinary observer would perceive that the "two designs are substantially the same." Egyptian Goddess, Inc. v. Swisa, Inc., 543 F.3d 665, 670 (Fed. Cir. 2008) (en banc) (quoting Gorham Co. v. White, 81 U.S. 14 Wall. 511, 528, 20 L.Ed. 731 (1871)).

On remand, the ITC should reassess the accused products to determine whether they are substantially similar to the mark in the infringement analysis.

B

Finally, the intervenors present other arguments in favor of affirmance, none of which we find persuasive.

First, the intervenors assert that the brand-name labeling of the accused products was dispositive of the likelihood of confusion as a matter of law. Contrary to the intervenors' argument, we have not held that such labeling is always legally sufficient to avoid likelihood of confusion but rather that those labels may be highly probative evidence. See, e.g., Conopco, Inc. v. May Dep't Stores Co., 46 F.3d 1556, 1570-71 (Fed. Cir. 1994); Braun, 975 F.2d at 828; Litton Sys., Inc. v. Whirlpool Corp., 728 F.2d 1423, 1446-47 (Fed. Cir. 1984), abrogated on other grounds by Two Pesos, 505 U.S. 763, 112 S.Ct. 2753, 120 L.Ed.2d 615; see also 1 McCarthy, supra, § 8:16; 4 id. § 23:53. The ALJ did not misapply this rule.

Second, one intervenor contends that the ITC erred by finding infringement absent evidence that its products have harmed Converse's reputation. Neither case cited by the intervenor establishes such a prerequisite to a finding of infringement. See Gen. Motors Corp. v. Keystone Auto. Indus., Inc., 453 F.3d 351, 356-59 (6th Cir. 2006); Payless Shoesource, Inc. v. Reebok Int'l Ltd., 998 F.2d 985, 989-90 (Fed. Cir. 1993).

Third, the intervenors argue that the '753 trademark is functional and hence not protectable. We find no error in the ITC's determination that the mark is non-functional. Any functional benefit is derived from the presence of toe caps and bumpers generally, not the particular design of the '753 trademark, and there are numerous commercial alternatives to that design. See, e.g., In re Morton-Norwich Prods., Inc., 671 F.2d 1332, 1338-39 (CCPA 1982).

CONCLUSION

We vacate and remand to the ITC for further proceedings consistent with this opinion.

VACATED AND REMANDED

COSTS

No costs.

Appendix

J.A. 40,928.

J.A. 41,738.

J.A. 42,824.

J.A. 43,133.

__________________________________

O'MALLEY, Circuit Judge, concurring-in-part and dissenting-in-part.

I agree with the majority that the International Trade Commission ("ITC" or "Commission") erred in its legal analysis in this matter and that its decision must be vacated. I also agree that a remand is appropriate. I even agree with many of the ways in which the majority finds that the ITC erred.[1] I cannot join in the majority's reasoning, however, because I believe that the majority overlooks important procedural facts and binding statutory authority to reach issues that are not properly before us. Thus, I do not agree with the majority's decision to remand questions regarding the validity of the registered mark for further consideration or its decision to even address questions of infringement. And, I cannot accede to the majority's failure to order the ITC to enter a remedy against the parties found in default in the proceedings below or to explain why public interest concerns would justify not doing so.

Specifically, I believe that the majority: (1) misperceives the scope of the ITC's authority to invalidate duly issued intellectual property rights when it addresses the issue of the validity of a registered mark; (2) blurs the line between the concepts of priority of use under common law and the validity of a registered mark; (3) espouses advisory—and unnecessary—opinions on the weight to be given certain survey evidence and the question of infringement; and (4) ignores the ITC's statutory obligation to enter remedies against defaulting parties. Thus, except where otherwise noted, I concur only in the conclusion that the ITC's findings must be vacated, and the matter remanded to the ITC for further analysis. I dissent from the fact that the majority discusses matters not properly before us and from the nature of the remand the majority outlines.

I begin with a discussion of the ITC's authority and how that impacts the scope of our review on appeal. The majority rightly finds fault with the ITC's failure "to distinguish between alleged infringers who began infringing before Converse obtained its trademark registration and those who began afterward." Majority Op. at 1115. But, the majority makes the same mistake when it considers and passes judgment on the ITC's determination regarding the validity of the registered mark. The ITC has no authority to invalidate a trademark—or patent for that matter— except and to the extent the validity of either is asserted as a defense to and is, thus, relevant to the question of whether an accused infringer can be liable for infringement. See Kinik Co. v. Int'l Trade Comm'n., 362 F.3d 1359, 1367 (Fed. Cir. 2004) (concluding that the ITC has "no authorization to determine patent validity when that defense was not raised"). Indeed, where a respondent seeks a declaration of invalidity by way of a counterclaim, rather than as a mere defense to infringement, it is statutorily required to remove that counterclaim to district court. 19 U.S.C. § 1337(c) ("Immediately after a counterclaim is received by the Commission, the respondent raising such counterclaim shall file a notice of removal with a United States district court in which venue for any of the counterclaims raised by the party would exist under section 1391 of Title 28."); see 28 U.S.C. § 1368 ("The district courts shall have original jurisdiction of any civil action based on a counterclaim raised pursuant to section 337(c) of the Tariff Act of 1930, to the extent that it arises out of the transaction or occurrence that is the subject matter of the opposing party's claim in the proceeding under section 337(a) of that Act.")

The validity of Converse's registered trademark is not relevant to the question of whether any remaining party in this case is liable for infringement or has violated 19 U.S.C. § 1337. Only two types of accused infringers remain in this case—(1) the defaulting parties, who acquiesced to Converse's claims and whose first uses began after the date of registration, and (2) the intervenors, who actively participated in the proceeding and whose first uses began before the date of registration. The validity of the registered mark is not relevant to Converse's claims against any of those parties.

The defaulting parties, by virtue of defaulting, never challenged Converse's claims of infringement or the validity of the registered mark. Even the majority admits that our opinion in Lannom Mfg. Co. v. Int'l Trade Comm'n, 799 F.2d 1572 (Fed. Cir. 1986) stands for the proposition "that the ITC cannot consider validity as to defaulting parties." Majority Op. at 1119 n.4. Because the ITC may only assess validity when raised as a defense to a claim of infringement, the ITC's ruling with respect to the registered mark was neither necessary nor appropriate as to those parties. The same is true with respect to the claims against the intervenors, but for different reasons. Because Converse concedes that the intervenors all began their allegedly infringing uses well before the date of registration (as the majority correctly concludes), the intervenors cannot be liable for infringement of the registered mark. That means that the ITC's discussion of the validity of the registered mark was neither necessary nor appropriate with respect to those respondents either.

The majority recognizes that registration is irrelevant to uses beginning before the date of registration when it concludes that "Converse's registration confers a presumption of secondary meaning beginning only as of the date of registration and confers no presumption of secondary meaning before the date of registration." Majority Op. at 1118. It even admits that Converse only "seeks protection as to the intervenors for a mark in the form of unregistered product-design." Majority Op. at 1116 (emphasis added). But it then questions the logical impact of its own conclusion when it asks "whether the issue of trademark validity needs to be addressed by the ITC on remand." Majority Op. at 1118. And, it does so again when it concludes that the question of whether the validity of the mark needs to be addressed on remand is, itself, "best addressed on remand." Majority Op. at 1119. But, whether the validity of the registered mark has any impact on pre-registration uses is a legal question. Indeed, it is one the majority decides, but from which it then backs away.

The majority then adds to the uncertainty of its holding by continuing at some length to consider the ITC's determination regarding the validity of the registered mark—implying it sees some relevance to that analysis—and ultimately remanding for further consideration of that question. [2] Majority Op. at 1119-24. It does so without identifying a single respondent in this appeal, or who was actively involved in the proceedings when the ITC issued its decision, for whom the validity of the registered mark is relevant to Converse's claims of infringement. In fact, the record is devoid of any indication that any such party remains.

The majority glosses over the procedural facts in this case that establish that no active respondent remains whose first use began after registration. When discussing the procedural background, the majority states that "Converse filed a complaint with the ITC alleging violations of section 337 by various respondents," and that, "[a]lthough some of the respondents defaulted, several appeared and actively participated in the ITC proceedings." Majority Op. at 1114 (emphases added). What the majority fails to mention is that these "several" respondents who "appeared and actively participated" comprise only the four intervenors whose earliest first uses predate the date of registration. A careful review of the procedural history and record below confirms this.

On October 14, 2014, Converse filed a complaint with the ITC alleging violations of § 1337 by 31 respondents. On January 12, 2015, an additional party, New Balance Athletic Shoe, Inc., intervened, bringing the total number of respondents to 32.[3] J.A. 54. Five of these respondents, identified above as the defaulting parties, acquiesced to Converse's claims by defaulting.[4] As noted, none of the defaulting parties contested Converse's allegations that they violated and continue to violate § 1337 by the importation of products that infringe a registered mark. J.A. 56. And none asserted defenses—based on the invalidity of the registered mark or otherwise—to Converse's claims of infringement. Exactly 21 of the remaining 27 respondents settled with Converse.[5] Finally, two other respondents moved to terminate based on a consent order or for good cause pursuant to ITC Rule 210.21(a)(1).[6] These motions were granted. Thus, the arithmetic shows that, at the time that the Administrative Law Judge ("ALJ") issued the Initial Determination on November 17, 2015, 23 of the total 32 initial respondents no longer remained in the case and five of the remaining nine respondents had defaulted. J.A. 52-56. Only Skechers U.S.A., Inc., Wal-Mart Stores, Inc., HU Liquidation, LLC, f.k.a. Highline United LLC, New Balance Athletics, Inc., f.k.a. New Balance Athletic Shoe, Inc., identified above as the intervenors, actively participated in the Investigation. J.A. 56.[7] And, as noted, these intervenors are "those respondents whose first uses came before the registration." Majority Op. at 1118. It is clear then that no parties remain whose earliest first uses began after registration and who also actively participated in the case. Thus, the relevant and only question is who has priority of use of a common law trade dress. The validity of the registered mark is irrelevant to that question.[8]

This leads me to my second concern with the majority's opinion—that it blurs the line between the concept of priority of use under common law and the concept of the validity of a registered mark. The above timeline establishes that the only issue properly before this court as it relates to the mark is the priority of use of an alleged common law right to the midsole trade dress. Yet the majority goes on to assess the validity of the registered mark even though no respondents remain for whom the registered mark is relevant. The natural, but demonstrably false implication of the majority's position is that a later-obtained registration is somehow relevant to establishing priority of use at an earlier date. This is contrary to our precedent stating that "[a] claim of trade dress infringement fails if secondary meaning did not exist before the infringement began." Braun Inc. v. Dynamics Corp. of Am., 975 F.2d 815, 826 (Fed. Cir. 1992). Indeed, a party with priority of use may continue to use a mark without infringing even if the mark later acquires distinctiveness —demonstrated through registration or otherwise. See Saratoga Vichy Spring Co., Inc. v. Lehman, 625 F.2d 1037, 1043 (2d Cir. 1980) ("Even if Saratoga Vichy has rights in the name `Saratoga' because its use of the name has acquired a secondary meaning, it could not prevent the use of that term by one whose use had begun before the secondary meaning was acquired.").

Here, if the intervenors are proven to have priority of use, they cannot be found to infringe even if Converse's later-obtained registered mark were valid and the intervenors continued to use the mark today. The intervenors argued exactly this point below, stating that they, as "junior user[s] cannot infringe rights that do not exist," and that, "the circumstances do not change if the senior user[, Converse,] later acquires secondary meaning (or a registration)" because "a junior user's permissible use does not become infringing from mere continuation." Suppl. J.A. 19. Thus, the only question that could properly be before the ITC on remand is whether Converse can show that its mark acquired distinctiveness as of each first use by each intervenor. The majority's attempt to expand the scope of our review by considering the validity of the registered mark conflates the concepts of the validity of a registered mark and priority of use. That Converse mischaracterized the effect of the registered mark as it relates to its claims against the intervenors, or that the ITC seemed to misunderstand the extent of its authority to invalidate a registered mark, cannot justify the majority's decision to compound the confusion at the heart of the ITC's judgment.

Another related concern is with the majority's decision to pass judgment on the relevance or adequacy of the Butler survey. Because the validity of the registered mark is not at issue, any discussion of a clearly post-dated survey is pure dicta. I agree with the majority that surveys significantly postdating historical uses are not relevant to the question of priority of use. This should end any further discussion of the Butler survey. The majority's extended consideration of the merits of the survey, including its suggestion that "the Butler survey may have relevance since it was conducted within two years of the registration," is inappropriate.[9] Majority Op. at 1123.

My next concern with the majority opinion is in its assessment of the ITC's infringement analysis. Like its discussion of the merits of the Butler survey, the majority's discussion of the ITC's infringement findings are dicta. Indeed, that portion of the opinion is no more than an advisory opinion. Again, because Converse's claims alleging infringement of the registered mark are relevant only to Converse's claims against the defaulting parties, and the defaulting parties did not challenge Converse's infringement claims, the ITC must proceed on the assumption that those parties did infringe the registered mark at the time of their respective uses of the trade dress covered by the mark. As to the intervenors, unless and until a judgment of priority of use is established, questions of infringement are not at issue. The question of infringement—by any party—is either not at issue in this appeal or not ripe for our review. Even the majority acknowledges that the ITC's infringement findings were alternative findings and, thus, dicta, at the Commission level itself. See Majority Op. at 1115 ("The ITC determined that, if either trademark was not invalid or protectable, it was infringed...." (emphasis added)); see also J.A. 35 ("The Commission affirms the ID's finding that there is a likelihood of confusion with respect to the [registered mark] for specific accused footwear products if the [registered mark] is not invalid ...." (emphasis added)).

Even if the issue of infringement were not dicta, moreover, it is hard to see how the majority could conceivably vacate the Commission's findings. The ALJ made detailed findings under the factors set forth In re E.I. DuPont DeNemours & Co., 476 F.2d 1357, 1361 (CCPA 1973). He specifically analyzed the question of substantial similarity to the mark for each allegedly infringing product, making factual findings as to each. The ITC then adopted those findings. J.A. 119-55. Those findings are findings of fact to which we owe deference. Coach, 668 F.3d at 1365-66.

My final concern with the majority's opinion, and specifically the scope of its remand, is with its failure to instruct the ITC to enter a remedy against all the defaulting parties, or to justify its failure to do so by reference to any relevant public interest concerns. The ITC is required by statute to grant relief to Converse against the defaulting parties. Specifically, 19 U.S.C. § 1337(g)(1) states that, if any parties are found to be in default, then:

the Commission shall presume the facts alleged in the complaint to be true and shall, upon request, issue an exclusion from entry or a cease and desist order, or both, limited to that person unless, after considering the effect of such exclusion order upon the public health and welfare, competitive conditions in the United States economy, the production of like or directly competitive articles in the United States, and United States consumers, the Commission finds that such exclusion or order should not be issued.

This language requires the ITC to presume the facts alleged in the complaint to be true and to provide some form of relief against any parties found to be in default —be it an exclusion order (general or limited), a cease and desist order, or both. SAS Inst., Inc. v. Iancu, ___ U.S. ___, 138 S.Ct. 1348, 1354, 200 L.Ed.2d 695 (2018) (finding that the word "shall" means "must" because it "generally imposes a nondiscretionary duty."); Boomerang Tube LLC v. United States, 856 F.3d 908, 912 (Fed. Cir. 2017) (explaining that "shall" connotes a requirement). The ITC may only refuse to do so based on public interest factors unrelated to the validity of the mark or whether that mark was infringed by the defaulting parties.

Such a presumption was not always the law. Congress added the statutory section providing for default judgments in the 1988 amendment to the Trade Act. Prior to the amendment, the ITC could not issue a default judgment without first finding a violation under § 1337. See Certain Attache Cases, Inv. No. 337-TA-49, 1979 WL 61026, at *2-4 (USITC Mar. 1, 1979) (denying motion for default judgment because record evidence failed to demonstrate "that the importation of the infringing articles does not have the effect or tendency to destroy or substantially injure the domestic industry"). Thus, complainants faced the same burden of proof even when the named respondents had defaulted. In passing the 1988 amendment, Congress acknowledged that, without the participation of a party in default, a complainant faced difficulties proving facts sufficient to establish a violation of § 1337. See In the Matter of Certain Elec. Skin Care Devices, Inv. No. 337-TA-959, 2017 WL 8683854, at *14 (USITC Feb. 13, 2017) ("[D]iscovery is usually difficult, if not impossible, to obtain from named respondents who have chosen not to participate in an investigation."). Section 1377(g) addresses this by requiring the ITC to "presume the facts alleged in the complaint to be true," and "upon request, issue appropriate relief solely against that person." See H.R. Rep. No. 100-576, at 636 (1988) (Conf. Rep.).

Against that backdrop, the majority's remand instruction simultaneously does too much and too little. It does too much by directing the ITC to further address the validity and infringement of the registered mark, even though the statute requires that the ITC presume that Converse's infringement allegations against the defaulting parties are true and that its registered mark is valid. And it does too little by not directing the ITC to grant a remedy against the defaulting parties unless it explains why public interest factors would justify not doing so.

Converse alleged that importation of the defaulting parties' accused products would violate § 1337, in part, because the accused products infringe Converse's registered mark. These parties undisputedly defaulted, and Converse moved for default judgment. Thus, there is no need to assess the strength of Converse's allegations against the defaulting parties; rather, we must instruct the ITC to presume as true the facts alleged in Converse's complaint and to issue an exclusion order, a cease and desist order, or both limited to the defaulting parties unless the ITC determines that the public interest weighs against such relief.

For the reasons stated above, I believe the majority stretches the scope of this court's review beyond that which is appropriate to express its views on issues that are not properly before us. I also believe the majority ignores its and the ITC's obligation to order relief against the defaulting parties. Therefore, I cannot join the majority's reasoning and must concur only in the decision to vacate the ITC's findings and to remand for further proceedings. I dissent from both the majority's consideration of matters not before us and the nature of the remand the majority orders.