CLEVENGER, Circuit Judge.

Recot, Inc. ("Recot") opposes the application of M. C. Becton ("Becton") to register the mark FIDO LAY for natural agricultural products, namely edible dog treats. The Trademark Trial and Appeal Board ("Board") dismissed the opposition on the ground that there was no likelihood that consumers would confuse FIDO LAY with Recot's FRITO-LAY marks. See Recot, Inc. v. M.C. Becton, Opposition No. 96,518, slip op. at 22 (TTAB Dec. 4, 1998). Because the Board improperly discounted the fame of the FRITO-LAY marks, did not consider all of the relevant evidence when determining if the products were related, and improperly dissected the marks, we vacate and remand the case to the Board for further proceedings consistent with this decision.

I

Recot owns six federal registrations for the mark FRITO-LAY and related marks, all of which are incontestable, as illustrated below:

Five of the registrations are for use with a wide range of snack foods, including numerous varieties of chips, pretzels, crisps, crackers, dips, salsas, cookies, and like foods. One registration is for use with several nonfood items, such as housewares, clothing, caps, lapel pins, stationery, and totebags.

Recot, through its predecessors in interest and through its affiliated company Frito-Lay, Inc. (collectively referred to as "Recot"), has manufactured and sold a wide variety of snack food under its mark, FRITO-LAY, for over thirty years. Recot now sells FRITO-LAY products nationwide in supermarkets, grocery stores, mass merchandisers and wholesale clubs, convenience stores, food services, and vending machines. In 1995, retail sales of FRITO-LAY products exceeded $6 billion, and FRITO-LAY products enjoyed a greater than 50 percent market share in the estimated $12.1 billion domestic snack chip industry. In any given year, up to 90 percent of American households purchase at least one FRITO-LAY brand snack. Recot spent about $80 million in 1996 on advertising and promotion for products with the FRITO-LAY mark. Recot also licenses for manufacture, promotion, and sale many widely varying goods bearing the FRITO-LAY mark, such as housewares, stationery, and clothing.

The applicant, Becton, sells natural dog treats — pig's ears, smoked turkey feet and the like-under the FIDO LAY mark in his pet food stores in the Birmingham, Alabama area and in Birmingham-area supermarkets. About 800-1100 units of FIDO LAY products have been sold. Becton also sells T-shirts and hats bearing the FIDO LAY mark.

In January 1994, Becton sought federal registration of its FIDO LAY mark. Recot opposed the mark as likely to cause confusion with its FRITO-LAY marks. The Board dismissed Recot's opposition, holding that there was no likelihood of confusion between the marks. The Board found that applicant's and opposer's goods are "simply not identical, nor otherwise related." Recot, slip op. at 20. It noted that both Recot's and Becton's products are inexpensive and may be purchased on impulse, but held without explanation that these factors were "diminished in importance" because the goods are "so different in nature." Id. at 18. Turning to the similarity or dissimilarity of the marks, the Board concluded that the marks were sufficiently dissimilar in meaning so as to be unlikely to cause confusion. The Board did not separately consider the sound or appearance of the marks, however. The Board noted that FRITO-LAY is a very famous mark, but did not treat the fame as "important" because the goods at issue were so different in nature. Finally, the Board found that there was no evidence of bad faith adoption of FIDO LAY. This appeal followed, conferring jurisdiction pursuant to 28 U.S.C. § 1295(a)(4) (1994).

II

The United States Patent and Trademark Office may refuse to register a trademark that so resembles a registered mark "as to be likely, when used on or in connection with the goods of the applicant, to cause confusion, or to cause mistake, or to deceive." 15 U.S.C. § 1052(d) (1994). Whether a likelihood of confusion exists is a question of law based on underlying factual determinations. See Lloyd's Food Prods., Inc. v. Eli's, Inc., 987 F.2d 766, 767, 25 USPQ2d 2027, 2028 (Fed. Cir. 1993); Kenner Parker Toys Inc. v. Rose Art Indus., Inc., 963 F.2d 350, 352, 22 USPQ2d 1453, 1455 (Fed. Cir. 1992). It is determined on a case-specific basis, applying the factors set out in In re E. I. DuPont DeNemours Co., 476 F.2d 1357, 1361, 177 USPQ 563, 567 (CCPA 1973) (enumerating factors that may be considered when relevant evidence is of record). The DuPont factors are: (1) the similarity or dissimilarity of the marks in their entireties as to appearance, sound, connotation and commercial impression; (2) the similarity or dissimilarity and nature of the goods or services as described in an application or registration or in connection with which a prior mark is in use; (3) the similarity or dissimilarity of established, likely-to-continue trade channels; (4) the conditions under which and buyers to whom sales are made, i.e., "impulse" vs. careful, sophisticated purchasing; (5) the fame of the prior mark (sales, advertising, length of use); (6) the number and nature of similar marks in use on similar goods; (7) the nature and extent of any actual confusion; (8) the length of time during and conditions under which there has been concurrent use without evidence of actual confusion; (9) the variety of goods on which a mark is or is not used (house mark, "family" mark, product mark); (10) the market interface between applicant and the owner of a prior mark; (11) the extent to which applicant has a right to exclude others from use of its mark on its goods; (12) the extent of potential confusion, i.e., whetherde minimis or substantial; and (13) any other established fact probative of the effect of use. See id.

Our review of the Board's ultimate conclusion is plenary. See Giant Food, Inc. v. Nation's Foodservice, Inc., 710 F.2d 1565, 1569, 218 USPQ 390, 394 (Fed. Cir. 1983). We uphold the Board's factual findings unless they are unsupported by substantial evidence. See Dickinson v. Zurko, 527 U.S. 150, 165, 50 USPQ2d 1930, 1937 (1999); In re Gartside, 203 F.3d 1305, 1315, 53 USPQ2d 1769, 1775 (Fed. Cir. 2000).

A

Recot argues that the Board failed to accord proper weight to the fame of the FRITO-LAY mark. The fifth DuPont factor, fame of the prior mark, when present, plays a "dominant" role in the process of balancing the DuPont factors. See Kenner Parker, 963 F.2d at 352, 22 USPQ2d at 1456 (citing Sure-Fit Prods. Co. v. Saltzson Drapery Co., 254 F.2d 158, 160, 117 USPQ 295, 296 (CCPA 1958)). Famous marks thus enjoy a wide latitude of legal protection. See id. at 352-53, 22 USPQ2d at 1456.

The Board acknowledged that the fame of Recot's mark was "unquestionably established," and indeed, Becton conceded this fame. Yet the Board did not treat the fame factor as "important" in this case because the dog treats sold under the FIDO LAY mark are "completely unrelated" to the human snacks sold under the FRITO-LAY marks. Recot, slip op. at 20. The Board noted the law of this circuit stating that the fame of a mark "weighs heavily" in cases involving famous marks, but distinguished those cases because they "generally involved identical or closely related products or services." Id. at 21-22 (listing cases).

The Board erred when it limited the weight accorded to the fame of the FRITO-LAY mark. We think that the Board's rule — that the fame of the FRITO-LAY marks extends no further than the products with which the marks are currently used — undercuts the legal standard of protection for famous marks. Famous marks are accorded more protection precisely because they are more likely to be remembered and associated in the public mind than a weaker mark. See Kenner Parker, 963 F.2d at 352, 22 USPQ2d at 1455-56. For this reason, this court emphasizes:

When an opposer's trademark is a strong, famous mark, it can never be "of little consequence." The fame of a trademark may affect the likelihood purchasers will be confused inasmuch as less care may be taken in purchasing a product under a famous name.

Specialty Brands, Inc. v. Coffee Bean Distribs., Inc., 748 F.2d 669, 675, 218 USPQ 1281, 1284 (Fed. Cir. 1984); cf. Kenner Parker, 963 F.3d at 354, 22 USPQ2d at 1457 (stating that the fame of a mark does not "cut both ways" in the analysis of the likelihood of confusion).

This reasoning applies with equal force when evaluating the likelihood of confusion between marks that are used with goods that are not closely related, because the fame of a mark may also affect the likelihood that consumers will be confused when purchasing these products. Indeed, it is precisely these circumstances which demand great vigilance on the part of a competitor who is approaching a famous mark, for, as the present case illustrates, the lure of undercutting or discounting the fame of a mark is especially seductive. See Recot, slip op. at 19 ("It is applicant's position that opposer's marks are famous for a variety of human food products, but that the fame of opposer's marks does not extend beyond that field. . . ."). Accordingly, we hold that the fame of the mark must always be accorded full weight when determining the likelihood of confusion. When a famous mark is at issue, a competitor must pause to consider carefully whether the fame of the mark, accorded its full weight, casts a "long shadow which competitors must avoid."Kenner Parker, 963 F.3d at 353, 22 USPQ2d at 1457; see also Nina Ricci S.A.R.L. v. E.T.F. Enters., 889 F.2d 1070, 1074, 12 USPQ2d 1091, 1904 (Fed. Cir. 1989) ("There is no excuse for even approaching the well-known trademark of a competitor" (internal quotations omitted).). Although fame alone cannot overwhelm the other DuPont factors as a matter of law,see University of Notre Dame du Lac v. J.C. Gourmet Food Imports Co., Inc., 703 F.3d 1372, 217 USPQ 505 (Fed. Cir. 1983), fame deserves its full measure of weight in assessing likelihood of confusion. Because the Board erred when it discounted the fame of the FRITO-LAY mark, we vacate the Board's decision, and remand for further consideration consistent with this decision.

The Board also erred when it distinguished this court's precedent on the ground that our prior cases concerned products that were identical or closely related. Indeed, this court and its predecessor court have consistently stated that fame of the mark is a dominant factor in the likelihood of confusion analysis for a famous mark, independent of the consideration of the relatedness of the goods. See, e.g., American Sugar Refining Co. v. Andreassen, 296 F.2d 783, 784, 132 USPQ 10, 11 (CCPA 1961) (concluding that DOMINO for sugar is confusingly similar to DOMINO for pet food); Hunt Foods Indus., Inc. v. Gerson Stewart Corp., 367 F.2d 431, 435, 151 USPQ 350, 352 (CCPA 1966) (holding that HUNT'S for canned food products is confusingly similar to HUNT for cleaning products); Kenner Parker, 963 F.3d at 354-56, 22 USPQ2d at 1457-58 (treating fame of mark and relatedness of products as separate inquiries). The Board's holding is also contrary to its own precedent, which has accorded full weight to the fame of a famous mark when analyzing likelihood of confusion between products that were not closely related. See Tiffany Co. v. Classic Motor Carriages Inc., 10 USPQ2d 1835, 1843 (TTAB 1989) (holding likelihood of confusion between CLASSIC TIFFANY for automobiles and opposer's mark TIFFANY for jewelry, silver, and similar items); R.J. Reynolds Tobacco Co. v. R. Seelig Hille, 201 USPQ 856, 860 (TTAB 1978) (holding SIR WINSTON and design for tea is likely to cause confusion with WINSTON for cigarettes).

B

The next DuPont factor we consider is the "relatedness of the goods." DuPont, 476 F.2d at 1361, 177 USPQ at 567; accord In re Martin's Famous Pastry Shoppe, Inc., 748 F.2d 1565, 1566-67, 223 USPQ 1289, 1290 (Fed. Cir. 1984). The Board found that the FIDO LAY dog treats were not identical or closely related to the FRITO-LAY human snacks, and further found that none of Recot's collateral or licensed goods were related to dog treats. See Recot, slip. op at 17. The Board declined to consider the lay testimony of both parties' witnesses that several companies produce and sell both pet and human foods, because it deemed the evidence of no persuasive value. See id. at 21.

The Board erred when it refused to consider the lay evidence that several large companies produce and sell both pet and human food in deciding whether a consumer would reasonably believe that FIDO LAY dog treats originated from the same source as FRITO-LAY human snacks. Indeed, the evidence seems extremely pertinent to the question of whether, absent any evidence of current use of the FRITO-LAY marks for pet food, a consumer would likely think that FRITO-LAY produced, sponsored, or licensed its mark for use for pet snack products. See, e.g., Tuxedo Monopoly, Inc. v. General Mills Fun Group, Inc., 648 F.2d 1335, 1336, 209 USPQ 986, 702 (CCPA 1981) (finding extensive licensing of mark MONOPOLY for real estate game relevant evidence of relatedness of goods); R.J. Reynolds, 201 USPQ at 860 (considering evidence that opposer marketed a wide variety of goods and considered expansion into beverage area relevant to relatedness). Thus, even if the goods in question are different from, and thus not related to, one another in kind, the same goods can be related in the mind of the consuming public as to the origin of the goods. It is this sense of relatedness that matters in the likelihood of confusion analysis. We further note, on the same point, that the Board did not mention in its analysis other evidence presented by Recot, including evidence of Recot's co-merchandising scheme with Walt Disney, Inc., in connection with the release of the Disney film "101 Dalmatians," in which FRITO-LAY-branded products were sold with images of dalmatian puppies on the product bags and associated display racks. On remand, the Board will have the opportunity to consider all of the relevant evidence as it considers the relatedness of the goods.

C

Recot asserts that the Board erred when it considered the fourthDuPont factor, the conditions under which, and buyers to whom, sales are made. The Board acknowledged the evidence that both Recot's and Becton's goods are inexpensive and may be purchased on impulse, but without explanation accorded this factor less weight based on its conclusion that the goods were unrelated. When products are relatively low-priced and subject to impulse buying, the risk of likelihood of confusion is increased because purchasers of such products are held to a lesser standard of purchasing care. See Kimberly-Clark Corp. v. H. Douglas Enter., Ltd., 774 F.2d 1144, 1146, 227 USPQ 541, 542 (Fed. Cir. 1985);Hunt Foods, 367 F.2d at 434, 151 USPQ at 350. Once again, the fact that the goods themselves are different — dog food and human food — does not compel a conclusion that consumers may not confuse the origin of the foods in hasty, economically painless, transactions. Further, the potential for confusion in these circumstances is accentuated by the significance of a famous mark. See Kenner Parker, 963 F.2d at 355, 22 USPQ2d at 1457; Speciality Brands, 748 F.2d at 676, 218 USPQ at 1284. On remand, the Board must accord this factor its full weight when determining the likelihood of confusion.

D

Recot argues that the Board erred because it improperly dissected the mark by considering only the connotation of the marks when considering the first DuPont factor, the similarity or dissimilarity of the marks. We agree with Recot. The similarity or dissimilarity of the marks in their entirety is to be considered with respect to appearance, sound, and connotation. See DuPont, 476 F.2d at 1361, 177 USPQ at 567;Olde Tyme Foods, Inc. v. Roundy's, Inc. 961 F.2d 200, 202-03, 22 USPQ2d 1542, 1544-45 (Fed. Cir. 1992). All relevant facts pertaining to appearance, sound, and connotation must be considered before similarity as to one or more of those factors may be sufficient to support a finding that the marks are similar or dissimilar. See id. at 203, 22 USPQ2d at 1545; In re National Data Corp., 753 F.2d 1056, 1058, 224 USPQ 749, 751 (Fed. Cir. 1985) (stating anti-dissection rule).

In this case, however, the Board only considered the connotation of part of the marks — FRITO having a different connotation than FIDO — before concluding that the marks were very dissimilar, and had different commercial impressions. See Recot, slip op. at 15-16. The Board did not consider the appearance or overall sound of the marks, a conspicuous oversight given the similar block capital letters of the FIDO LAY mark and two of the Recot marks. On remand, the Board will have the opportunity to consider the appearance, sound, and connotation of the marks before reaching its conclusion as to the similarity or dissimilarity of the marks in their entirety.

III

Recot perceives two additional flaws in the Board's decision. First, Recot argues that the Board erred in finding that Becton adopted FIDO LAY in good faith. Instead, Recot thinks that Becton simply copied the FRITO-LAY mark. Before the Board, Becton testified that his stepson came up with the idea of the mark one day at the pet store, when he offered a customer's dog a tasty treat, commanding "Fido! Lay!" to get the dog to behave. Although Becton admitted that he knew of the FRITO-LAY mark, he testified that FIDO LAY "just rang a bell" that day in the pet shop.

The Board found no reason to reject Becton's mark origin explanation. Recot asks us, as it asked the Board, to reject Becton's explanation as far-fetched, and to rule that Becton simply purloined FIDO LAY from Recot's mark. We decline Recot's invitation. Absent a showing that Becton's explanation is untrue, Recot fails to bolster its claim to a likelihood of confusion between the marks. Because we do not reverse the Board's finding, we need not reach Recot's argument that, as a matter of law, evidence of intent to copy should give rise to a presumption of a likelihood of confusion.

Second, Recot asserts that the Board erred in failing to give it any benefit under the DuPont factor that considers channels of trade. Since the goods to which the marks are applied in this case are sold in like channels of trade, such as supermarkets, Recot argues that it should have received credit for its position under the channels of trade consideration. The Board did not weigh this factor in Recot's favor, because the law is that products should not be deemed related simply because they are sold in the same kind of establishments. See Federated Foods, Inc. v. Fort Howard Paper Co., 544 F.2d 1098, 1103, 192 USPQ 24, 26 (CCPA 1976) (stating that the mere existence of modern supermarket containing wide variety or products should not foreclose further inquiry into the likelihood of confusion arising from the use of similar marks on any goods so displayed). Absent some evidence, of which there is none, that the products are sold in close proximity to one another, we discern no error in the Board's finding. See Martin's, 748 F.2d at 1567, 223 USPQ at 1290 (noting that although dairy products generally need to be refrigerated and may be kept in different area of market, not error for Board to take notice that baked goods are stored and sold in refrigerated or frozen form, and that bread and cheese may be displayed in close proximity in the deli counter); Kimberly-Clark, 774 F.2d at 1146, 227 USPQ at 542 (products displayed in same section of store).

CONCLUSION

Because the Board improperly discounted the fame of the FRITO-LAY marks, did not consider all of the relevant evidence when determining if the products were related, and improperly dissected the marks, we vacate the Board's decision and remand the case to the Board for further proceedings consistent with this decision.

COSTS

No costs.

VACATED AND REMANDED.