RADER, Circuit Judge.

Palm Bay Imports, Inc. (Palm Bay) appeals from the decision of the United States Patent and Trademark Office Trademark Trial and Appeal Board (Board) refusing registration of the mark VEUVE ROYALE for sparkling wine on the ground of likelihood of confusion with three of opposer Veuve Clicquot Ponsardin's (VCP's) marks. Veuve Clicquot Ponsardin v. Palm Bay Imports, Inc., Opp'n No. 115,438, 2003 WL 21953664 (T.T.A.B. Aug. 4, 2003). This court concludes that the Board did not err in finding a likelihood of confusion, but substantial evidence does not support such a finding for one of the marks. Even though this court reverses the Board's conclusion as to one of the marks, the Board's refusal to register Palm Bay's VEUVE ROYALE mark is affirmed.

I.

In April 1998, Palm Bay filed an intent-to-use trademark application under Section 1(b) of the Lanham Act, 15 U.S.C. § 1051(b), for the mark VEUVE ROYALE for "alcoholic beverages, namely, sparkling wine." Veuve Clicquot Ponsardin, slip op. at 1. The examining attorney found no evidence of a similar mark that would bar registration on the ground of likelihood of confusion under Section 2(d) of the Trademark Act, 15 U.S.C. § 1052(d). In response to the June 22, 1999 Official Gazette publication of Palm Bay's application, VCP filed an opposition with the Board, alleging a likelihood of confusion between VEUVE ROYALE and five of its own marks. Specifically, VCP asserted a likelihood of confusion based on the following marks: (1) VEUVE CLICQUOT PONSARDIN; (2) VEUVE CLICQUOT; (3) VEUVE CLICQUOT PONSARDIN Design; (4) THE WIDOW; and (5) LA VIUDA.

On August 4, 2003, the Board refused registration of VEUVE ROYALE, finding a likelihood of confusion with (1) VEUVE CLICQUOT PONSARDIN, (2) VEUVE CLICQUOT, and (3) THE WIDOW. The Board dismissed VCP's Section 2(d) claim for LA VIUDA finding the doctrine of foreign equivalents inapplicable to marks in two different foreign languages, i.e., Spanish and French. Palm Bay appeals.

II.

Likelihood of confusion under the Lanham Act, 15 U.S.C. § 1052(d), is a legal determination based upon factual underpinnings. On-Line Careline, Inc. v. Am. Online, Inc., 229 F.3d 1080, 1084 (Fed. Cir. 2000). This court determines the question on a case-specific basis, applying the thirteen factors set forth in In re E.I. DuPont DeNemours Co., 476 F.2d 1357, 1361 (C.C.P.A. 1973), without deference. In re Int'l Flavors Fragrances, Inc., 183 F.3d 1361, 1365 (Fed. Cir. 1999). At the same time, this court reviews factual underpinnings for that legal conclusion, namely the DuPont factors, for substantial evidence. Dickinson v. Zurko, 527 U.S. 150, 156, 119 S.Ct. 1816, 144 L.Ed.2d 143 (1999); Bose Corp. v. QSC Audio Prods., 293 F.3d 1367, 1370 (Fed. Cir. 2002). Evidence is substantial if "a reasonable person might find that the evidentiary record supports the agency's conclusion." On-Line Careline, 229 F.3d at 1085.

III. The VEUVE Marks

Palm Bay asserts that the Board erred in its findings on four DuPont factors during analysis of VEUVE CLICQUOT and VEUVE CLICQUOT PONSARDIN: (1) the similarity of the marks; (2) third-party use of the term VEUVE; and (3) the fame of VCP's marks; and (4) purchaser sophistication (corresponding to the first, sixth, fifth and fourth DuPont factors, respectively).

A.

The first DuPont factor requires examination of "the similarity or dissimilarity of the marks in their entireties as to appearance, sound, connotation and commercial impression." DuPont, 476 F.2d at 1361. Palm Bay contends that the Board made two errors in finding that VEUVE ROYALE was confusingly similar to VCP's marks. First, it claims that the Board misstated the legal test of similarity by treating "commercial impression" as the ultimate conclusion rather than as a separate factor. Second, Palm Bay claims the Board erred in finding that VEUVE was the "prominent feature" of VCP's marks. Instead, argues Palm Bay, CLICQUOT is the dominant feature.

1. Test for Similarity

The Board phrased the legal test for similarity as whether the marks "when compared in their entireties in terms of appearance, sound and connotation, are similar or dissimilar in their overall commercial impressions." Veuve Clicquot Ponsardin, slip op. at 26-27. After conducting a thorough analysis of the appearance, sound, meaning, and commercial impression of the marks, the Board concluded that they were "more similar than dissimilar in their overall commercial impression for purposes of the first DuPont factor." Id. at 29. In both instances, the Board treated "commercial impression" as the ultimate conclusion under this prong rather than as a separate factor along with appearance, sound, and meaning. Palm Bay contends that the Board's inaccurate paraphrase of the legal standard infected its analysis and constitutes reversible error.

This court declines to find reversible error merely because the Board, in two instances, made a minor misstatement of the similarity test in an otherwise proper analysis. Moreover, this court's precedent counsels that the phrase "commercial impression" is occasionally used as a proxy for the ultimate conclusion of similarity or dissimilarity of marks resulting from a comparison of their appearance, sound, and meaning. Hewlett-Packard Co. v. Packard Press, Inc., 281 F.3d 1261, 1266-67 (Fed. Cir. 2002) (holding that PACKARD TECHNOLOGIES and HEWLETT PACKARD differ in appearance and sound, but the marks convey a similar commercial impression because consumers would be aware of Hewlett-Packard's heavy involvement in technology-based goods, and therefore the marks are similar in their entireties); Kenner Parker Toys, Inc. v. Rose Art Indus., Inc., 963 F.2d 350, 355 (Fed. Cir. 1992) (holding that in light of the appearance, sound and meaning of the marks PLAY-DOH and FUNDOUGH, consumers may receive the "same commercial impression" from the marks); Morton-Norwich Prods., Inc. v. S.C. Johnson Son, Inc., 531 F.2d 561, 562 (C.C.P.A. 1976) (holding that RAINFRESH is confusingly similar to RAIN BARREL given the close relationship of the goods and "similarity of commercial impressions").

2. Significance of CLICQUOT

Palm Bay next asserts that the Board gave insufficient weight to the relative dominance of CLICQUOT, while placing too much emphasis on the weaker VEUVE portion of VCP's marks. Palm Bay notes that VCP admitted that CLICQUOT was the most distinctive portion of its marks in a World Intellectual Property Organization (WIPO) Internet domain name proceeding, and that VCP has used CLICQUOT in its marketing materials, including radio spots, promotional products, and as the URL of its website (www.clicquot.com). Moreover, Palm Bay argues, VCP has no enforceable trademark rights in the term VEUVE, and has never used the term, standing alone, on its champagne bottles or in its marketing efforts.

The Board correctly weighed the relative importance of VEUVE and CLICQUOT. VEUVE is an arbitrary term as applied to champagne and sparkling wine, and thus conceptually strong as a trademark. See, e.g., Nautilus Group, Inc. v. Icon Health Fitness, Inc., 372 F.3d 1330, 1340 (Fed. Cir. 2004) (defining an arbitrary mark as "a known word used in an unexpected or uncommon way" and observing that such marks are typically strong). Indeed, VCP has made VEUVE a theme of its marketing efforts because of the history, and the widow Clicquot herself, connected with the origins of the company. To be sure, CLICQUOT is an important term in the mark, but VEUVE nevertheless remains a "prominent feature" as the first word in the mark and the first word to appear on the label. Not only is VEUVE prominent in the commercial impression created by VCP's marks, it also constitutes "the dominant feature" in the commercial impression created by Palm Bay's mark. Veuve Clicquot Ponsardin, slip op. at 31-32 (VEUVE is first word in mark; significance of ROYALE is more laudatory and less source-indicating than VEUVE or "widow"; arbitrary term VEUVE contributes more to commercial impression of product than ROYALE). The presence of this strong distinctive term as the first word in both parties' marks renders the marks similar, especially in light of the largely laudatory (and hence non-source identifying) significance of the word ROYALE.

This case is distinguishable from cases upon which Palm Bay relies where this court found no likely confusion between marks involving a common term. See, e.g., Kellogg Co. v. Pack `em Enters., Inc., 951 F.2d 330 (Fed. Cir. 1991) (FROOTEE ICE for flavored ice bars not likely to cause confusion with FROOT LOOPS for breakfast cereal and related products); Keebler Co. v. Murray Bakery Prods. Inc., 866 F.2d 1386 (Fed. Cir. 1989) (PECAN SHORTIES not likely to cause confusion with PECAN SANDIES for cookies). In these cases, the common term is a generic or non-distinctive term. Here, the common term — VEUVE — is distinctive, and as such its presence in both parties' marks enhances the likelihood of confusion. Substantial evidence therefore supports the Board's finding that the marks are similar under the first DuPont factor.

B.

Palm Bay next contends that the Board erred in rejecting evidence of third-party use of at least five different alcoholic beverages that use the term VEUVE (or a foreign equivalent). This assertion requires analysis of the sixth DuPont factor, which considers "the number and nature of similar marks in use on similar goods." DuPont, 476 F.2d at 1361. Evidence of third-party use of similar marks on similar goods is relevant to show that a mark is relatively weak and entitled to only a narrow scope of protection. General Mills, Inc. v. Kellogg Co., 824 F.2d 622, 626-27 (8th Cir. 1987); J. Thomas McCarthy, 2 McCarthy on Trademarks and Unfair Competition § 11:88 (4th ed. 2001) (hereinafter McCarthy on Trademarks). In particular, Palm Bay argues that third-party evidence consists of at least five different alcoholic beverages that use the term VEUVE: (1) VIUDA DE ROMERO (tequila); (2) VEUVE CASTARDE BAS ARMAGNAC (brandy); (3) VEUVE ROTH BRANDY (brandy); (4) VEUVE DE LALANDE (sparkling wine); and (5) VEUVE DU VERNAY (sparkling wine). Each of these brands has at various points of time appeared in an industry trade publication, the Beverage Media Guide, which lists all beverage products in the wine and spirits categories sold by wholesalers to restaurants and retail stores in New York State. The Board rejected this evidence on the ground that the Beverage Media Guide is only distributed to the trade, and thus does not show the extent to which consumers actually encounter these brands in the marketplace.

The probative value of third-party trademarks depends entirely upon their usage. E.g., Scarves by Vera, Inc. v. Todo Imports, Ltd., 544 F.2d 1167, 1173 (2d Cir. 1976) ("The significance of third-party trademarks depends wholly upon their usage. Defendant introduced no evidence that these trademarks were actually used by third parties, that they were well promoted or that they were recognized by consumers."). While the Beverage Media Guide is compelling evidence that distributors were aware that the term VEUVE was used for other alcoholic products, it is not evidence that the consuming public was likewise aware. At best, the Beverage Media Guide is evidence that the consuming public could potentially be cognizant of third-party use of the term VEUVE. Absent evidence of the consuming public's awareness, however, the Beverage Media Guide standing alone does not suffice. As this court has previously recognized where the "record includes no evidence about the extent of [third-party] uses . . . [t]he probative value of this evidence is thus minimal." Han Beauty, Inc. v. Alberto-Culver Co., 236 F.3d 1333, 1338 (Fed. Cir. 2001) (emphasis added).

Lloyd's Food Products, Inc. v. Eli's, Inc., 987 F.2d 766 (Fed. Cir. 1993), does not alter that reasoning. In Lloyd's, this court held that the Board erred in failing to consider evidence of third-party use of service marks in phone directories. Id. at 768. Unlike the Lloyd's phone directories that were distributed to the public, the Beverage Media Guide is only sent to distributors, not to the consuming public. Additionally, the issue in Lloyd's involved service marks, for which advertising in phone directories is evidence of use. Here, by contrast, the marks are trademarks applied to goods for which a mere listing in a directory is not evidence of such use. Thus, the Lloyd's holding is inapposite to the facts of this case.

Palm Bay further argues that the Board erred in minimizing the import of retail marketing evidence it presented for the third-party mark VEUVE DE VERNAY. Specifically, Palm Bay notes that its private investigator discovered that six New York stores displayed for sale the VEUVE DE VERNAY sparkling wine as well as internet web sites and restaurant lists. Though the Board determined that such evidence exhibited more than de minimis use, it did not credit the single mark use with much weight. As the Board observed, "[t]he purpose of a defendant introducing third party uses is to show that customers have become so conditioned by a plethora of such similar marks that customers `have been educated to distinguish between different [such] marks on the bases of minute distinctions.'" Veuve Clicquot Ponsardin, slip op. at 25 (citing McCarthy on Trademarks, at § 11:88). Palm Bay's evidence does not rise to the level of demonstrating that the single third-party use was so widespread as to "condition" the consuming public.

Accordingly, substantial evidence supports the Board's finding that the strength of VCP's mark was not undermined by third-party use in either the Beverage Media Guide or the retail market.

C.

Palm Bay argues that the Board erred in concluding that VCP's marks are famous, per the fifth DuPont factor, and therefore entitled to a wide scope of protection. In its view, the Board applied an incorrect legal standard for measuring fame by focusing on a narrow class of consumers limited to purchasers of champagne and sparkling wine. The proper standard for fame, Palm Bay asserts, is whether a mark has achieved "extensive public recognition and renown" among the general public. Kenner Parker Toys, Inc., 963 F.2d at 353. Palm Bay further contends there was insufficient evidence to support a finding of fame.

Fame of an opposer's mark, if it exists, plays a "dominant role in the process of balancing the DuPont factors." Recot, Inc. v. M.C. Becton, 214 F.3d 1322, 1327 (Fed. Cir. 2000). Famous marks enjoy wide latitude of legal protection since they are more likely to be remembered and associated in the public mind than weaker marks, and are thus more attractive as targets for would-be copyists. Id. Indeed, "[a] strong mark . . . casts a long shadow which competitors must avoid." Kenner Parker Toys, 963 F.2d at 353. Fame for likelihood of confusion purposes and fame for dilution purposes, however, are distinct concepts. The Toro Co. v. ToroHead, Inc., 61 U.S.P.Q.2d 1164, 1170 (T.T.A.B. 2001). While dilution fame is an either/or proposition — fame either does or does not exist — likelihood of confusion fame "varies along a spectrum from very strong to very weak." In re Coors Brewing Co., 343 F.3d 1340, 1344 (Fed. Cir. 2003).

This court's statement in Kenner Parker Toys that fame for likelihood of confusion purposes is to be measured by whether a mark has achieved "extensive public recognition and renown" was not intended to require public awareness among every segment of the U.S. population. Indeed, the court in Kenner Parker Toys did not consider the issue of the extent of public renown necessary to qualify for treatment as a famous mark. Palm Bay's argument for a general public standard would be contrary to the trend of our case law and improperly elevate likelihood of confusion fame to the higher and more rigorous standard for dilution fame required under the FTDA. In this case, some classes of consumers would not have occasion to be exposed to VCP's champagne products, such as children or adolescents below the drinking age, or adults who for religious or other reasons choose not to consume alcoholic beverages. Thus, a general public awareness standard does not adequately reflect the mark's fame amongst the purchasing public. Fame for confusion purposes arises as long as a significant portion of the relevant consuming public, namely, purchasers of champagne and sparkling wine, recognizes the mark as a source indicator. Although this court has not directly addressed the question of what segment of the consuming public must be aware of a mark in order for it to be considered famous in a likelihood of confusion analysis, it has indirectly suggested that a mark's renown within a specific product market is the proper standard. See Bose Corp., 293 F.3d at 1376 (2002) ("Large market shares of product sales or large percentages of advertising expenditures in a product line would buttress claims to fame."). Similarly, this court's precedent has defined the relevant product market for purposes of determining likelihood of confusion as customers and potential customers. Elec. Design Sales, Inc. v. Elec. Data Sys. Corp., 954 F.2d 713, 716 (Fed. Cir. 1992) (holding that purchaser confusion is the "primary focus" and, in case of goods and services that are sold, "the inquiry generally will turn on whether actual or potential `purchasers' are confused"). Accordingly, this court holds that the proper legal standard for evaluating the fame of a mark under the fifth DuPont factor is the class of customers and potential customers of a product or service, and not the general public. The Board did not err in so holding.

Given this clarification of the proper relevant market for evaluating the fame of a mark, this court finds that there was sufficient factual evidence of fame within the market of purchasers of champagne and sparkling wine to support the Board's conclusion. The record indicates that VCP's sales volume and advertising expenditures since 1990 have been substantial. VEUVE CLICQUOT champagne is the second leading brand sold in the U.S. The brand is sold in 8,000 restaurants nationwide, and in liquor stores, wine shops, and other retail establishments. VCP advertises in general interest magazines such as Vanity Fair and in wine specialty magazines, radio ads, point-of-sale displays, through in-store and in-restaurant wine tastings and events, through sponsorship of events, and on its Internet site. VCP's products have been featured in articles and reviews in both specialized and general interest magazines. According to an April 2001 issue of Wine and Spirits, VEUVE CLICQUOT was the most-ordered wine in the "sparkling wine" category of the 363 survey respondents. VCP's products have also been featured in Business Week, American Way (in-flight magazine of American Airlines), The New York Times, the Boston Globe, Money magazine, and the Detroit News. Moreover, Palm Bay's President, David Taub, admitted that the VEUVE CLICQUOT mark is famous. His later qualification that such fame was limited to the "top-end" segment of the market does not diminish the significance of his admission in view of the Board's finding that high-end champagne and less-expensive sparkling wines are marketed in the same channels of trade to the same consumers. Finally, the Board noted that several WIPO domain name arbitration decisions had found VCP's marks to be famous. While acknowledging Palm Bay's argument that a mark must be famous among purchasers in the United States, whereas WIPO examined VCP's marks worldwide, the Board properly noted that such evidence nonetheless provided a "confirmatory context" for VCP's other evidence of fame. Considering this evidence in its totality and in context, this court finds that substantial evidence supports the Board's finding of fame.

D.

The fourth DuPont factor examines the conditions under which, and to whom, sales are made. DuPont, 476 F.2d at 1361. Purchaser sophistication may tend to minimize likelihood of confusion. Conversely, impulse purchases of inexpensive items may tend to have the opposite effect. Recot, Inc., 214 F.3d at 1329.

In concluding that this factor was "neutral, at best," Veuve Clicquot Ponsardin, slip op. at 15, the Board found that champagne and sparkling wines are not necessarily expensive goods which are always purchased by sophisticated purchasers who exercise a great deal of care in making their purchases. This court agreed with the Board's finding. Although some champagne can be expensive, many brands sell for around $25 a bottle, and sparkling wines for less than $10 a bottle. Moreover, general consumers, not just connoisseurs, occasionally purchase champagne or sparkling wines on celebratory occasions, with little care or prior knowledge. And even more sophisticated purchasers might be aware that champagne houses offer both types of products under similar marks, and could easily conclude that VEUVE ROYALE was Veuve Clicquot's sparkling wine. This market expansion rationale undercuts Palm Bay's argument that because VCP has never sold a sparkling wine product or low-priced champagne, and that it has always used its CLICQUOT house mark in connection with its sales, consumers would be unlikely to confuse Palm Bay's inexpensive sparkling wine and VCP's slightly more expensive champagne products. Considering this evidence in its entirety, this court concludes that substantial evidence supports the Board's finding under this DuPont factor.

IV.

The Board held that Palm Bay's VEUVE ROYALE was confusingly similar to VCP's mark THE WIDOW, in part because under the doctrine of foreign equivalents, an appreciable number of purchasers in the U.S. speak and/or understand French, and they "will translate" applicant's mark into English as "Royal Widow." Veuve Clicquot Ponsardin, slip op. at 36. The Board erred in so finding.

Under the doctrine of foreign equivalents, foreign words from common languages are translated into English to determine genericness, descriptiveness, as well as similarity of connotation in order to ascertain confusing similarity with English word marks. See In re Sarkli, Ltd., 721 F.2d 353 (Fed. Cir. 1983); In re Am. Safety Razor Co., 2 U.S.P.Q.2d 1459, 1460 (T.T.A.B. 1987) (finding BUENOS DIAS for soap confusingly similar to GOOD MORNING for shaving cream). When it is unlikely that an American buyer will translate the foreign mark and will take it as it is, then the doctrine of foreign equivalents will not be applied. In re Tia Maria, Inc., 188 U.S.P.Q. 524 (T.T.A.B. 1975) (no likelihood of confusion between TIA MARIA for a Mexican restaurant and AUNT MARY's for canned vegetables).

In comparing VEUVE ROYALE with VEUVE CLICQUOT PONSARDIN and VEUVE CLICQUOT, the Board found that "an appreciable number of purchasers are unlikely to be aware that VEUVE means `widow' and are unlikely to translate the marks into English." Veuve Clicquot Ponsardin, slip op. at 11 (emphasis added). In comparing VEUVE ROYALE with THE WIDOW, however, the Board found that "[A]n appreciable number of purchasers in the United States speak and/or understand French, and they will translate applicant's mark into English as ROYAL WIDOW." Id., slip op. at 14 (emphasis added). An appreciable number of U.S. consumers either will or will not translate VEUVE into "widow," and the Board was inconsistent in its application of the doctrine of foreign equivalents.

Although words from modern languages are generally translated into English, the doctrine of foreign equivalents is not an absolute rule and should be viewed merely as a guideline. In re N. Paper Mills, 20 C.C.P.A. 1109, 64 F.2d 998, 999 (1933); McCarthy on Trademarks, at § 11:34. The doctrine should be applied only when it is likely that the ordinary American purchaser would "stop and translate [the word] into its English equivalent." In re Pan Tex Hotel Corp., 190 U.S.P.Q. 109, 110 (T.T.A.B. 1976). This court agrees with the T.T.A.B. that it is improbable that the average American purchaser would stop and translate "VEUVE" into "widow." Substantial evidence does not support the Board's finding regarding the doctrine of foreign equivalents. This court, therefore, reverses the Board's finding of likelihood of confusion for THE WIDOW.

CONCLUSION

This court affirms the Board's decision that a likelihood of confusion exists between applicant's VEUVE ROYALE mark and opposer's marks VEUVE CLICQUOT PONSARDIN and VEUVE CLICQUOT. We reverse the Board's conclusion of the likelihood of confusion as to the THE WIDOW mark and we affirm the Board's refusal to register Palm Bay's VEUVE ROYALE mark.

COSTS

Each party shall bear its own costs.

AFFIRMED.