In this case we consider the breadth of protection afforded a registered trademark that also serves as a model number for a hand-operated staple gun. Specifically, we consider whether the trademark is infringed by another company's use of the letter and number that constitute the mark--in combination with other letters and numbers to form a longer model number--on a higher-priced and functionally different stapler. The United States District Court for the District of Connecticut (Peter C. Dorsey, Chief Judge ) held that the plaintiff, Arrow Fastener Co., Inc. ("Arrow"), manufacturer of the hand-operated staple gun, had a valid trademark in "T-50," and that The Stanley Works ("Stanley") infringed upon that trademark by using "T50" as part of a multi-character designation for its pneumatic staplers (staplers powered by an air compressor). 870 F. Supp. 427 (D.Conn.1994). For the reasons set forth below, we find that the scope of Arrow's trademark protection is not so broad as to bar Stanley's use of "T50" with other letters and numbers as a model number on its pneumatic staplers.

* Unless otherwise indicated, the following facts are not in dispute. Arrow is a manufacturer of staple guns, staplers and other tools. In the early 1950s, it began marketing a hand-operated staple gun identified as T-50. The main market for the Arrow T-50 staple gun is the "do-it-yourself home owner," who uses the device for such tasks as installing, insulating, upholstering, roofing, carpeting, fencing, and wiring. The product retails for approximately $20. The Arrow T-50 stapler has enjoyed extraordinary commercial success and has become the largest seller of its kind in the United States. In 1988, Arrow sought to register the "T-50" identifier as a trademark with the U.S. Patent and Trademark Office ("USPTO"). Although the USPTO initially rejected Arrow's application on grounds that "T-50" merely described an Arrow model, it subsequently granted trademark registration upon receiving evidence that the mark had "become a distinctive indicator of source for the goods." The registration is for "staple gun tackers, in Class 8" of the USPTO's classes of goods and services. Class Eight includes "[h]and tools and instruments; cutlery, forks, and spoons; side arms." See 37 C.F.R. Sec. 6.1(8). Arrow's advertising and packaging for its T-50 stapler, as well as the product itself, feature both the trademark "T-50" and the housemark "Arrow."1

Stanley is a manufacturer of tools and hardware products. It acquired the Bostitch division of Textron, Inc., in February 1986 and now sells Bostitch products under the housemarks "Stanley-Bostitch" and "Bostitch."2 In 1989, Stanley introduced a series of six pneumatic staple guns that it called the "T50 series." Powered by an air compressor, pneumatic staplers are used primarily by individuals in the construction trades for such tasks as sheathing, decking, lathing, and roofing. Although the specific natures of these tasks differ, they generally involve attaching board or plywood to the frame, roof, or floor of a building. The tools retail for about $400, but they cannot be operated without air compressors and various other accessories, which cost approximately $425 and $125, respectively.

When initially marketed, each stapler in the series was packaged in a box that bore the primary model number T50, as well as a secondary model number, S2-1, S2-2, S4-1, S4-2, S5-1, or S5-2. The top and front panels of Stanley's boxes for this series of staplers prominently displayed the legend "model T50 series." The six secondary model numbers appeared on a side panel of the box, each accompanied by a square that could be checked off to indicate which particular model was inside the box. The housemark "Bostitch" appeared on all three panels.

After writing letters to Stanley, protesting Stanley's use of the "T50 series" designation, Arrow filed suit in 1991, seeking declaratory and injunctive relief and damages under the Lanham Act, 15 U.S.C. Sec. 1114(1)(a), for infringement of a registered trademark; 15 U.S.C. Sec. 1125, for false designation of origin; the Connecticut Unfair Trade Practices Act, Conn.Gen.Stat. Secs. 42-110a to 42a-110q; and Connecticut trademark infringement common law. In 1993, Stanley, for reasons disputed by the parties, changed the labeling on its stapler packages, so that the T50 series designation, which had originally appeared unattended by other letters and numbers, now appears only as part of a multi-character designation. Specifically, Stanley joined the six secondary model numbers with the primary number. The resulting model numbers are T50S2-1, T50S2-2, T50S4-1, T50S4-2, T50S5-1, and T50S5-2. The tops of the new boxes do not bear a model number at all, and the front and side panels bear one of these six-digit numbers on a small, computer-generated strip.

After a bench trial, the district court concluded that Stanley's use of T50 as a series identifier infringed on Arrow's rights in the T-50 mark, and the court granted injunctive relief. The November 14, 1994, judgment provided, in pertinent part, that Stanley was:

enjoined from manufacturing, advertising, promoting, selling or offering for sale any product bearing the mark T50 or T-50 and from using in commerce any reproduction, counterfeit, copy, or colorable imitation of the registered trademark T-50 in connection with the sale, offering for sale, distribution, or advertising of any goods or services on or in connection with which such use is likely to cause confusion, or to cause mistake, or to deceive, including by using the mark T50, or any mark comparable or similar to the T-50 trademark, in any manner related to any stapling, tacking or fastening product....

Stanley appealed, seeking reversal and vacatur of the injunction. In the alternative, Stanley sought a modification of the injunction to clarify that it applied only to the use of T50 alone. Stanley argued that the injunction should not apply to the use of T50 as part of a complete model number or to Stanley's use of other number/letter combinations, such as T55 for a certain model of its pneumatic tackers or BT50 for its pneumatic brad tackers.

At oral argument on March 10, 1995, Stanley stated on the record that it was no longer using T50 alone on its packaging, but only in combination with other numbers and letters to form part of longer model numbers. Arrow Fastener Co. v. Stanley Works, 870 F. Supp. 427 (D.Conn.1994) (Appendix at 1, 2), remanded without published opinion, 52 F.3d 311 (2d Cir.1995). It argued that whether the injunction reached this current practice was unclear. Arrow maintained that the injunction did and should reach Stanley's current practice. We agreed with Stanley that the scope of the injunction was unclear, and we remanded the case to the district court to determine in the first instance whether Stanley's current use of the model numbers violated the injunction. Id. On remand, the district court stated that its injunction did reach uses such as T50S2-1, T50S2-2, T50S4-1, T50S4-2, T50S5-1, T50S5-2, as well as BT50-1, BT50-2, and BT50-3, because the record supported a finding that those uses "would suggest a product associated with the T-50 mark." Response to Mandate, at 1 (filed April 28, 1995, doc. # 130). The district court stated that its injunction would not reach the use of T55S4-1, T55S4-2, T55S5-1, and T55S5-2, because T55 "was not found to be a protected mark." Id.

Because Stanley is no longer using T50 alone, and because it represented at oral argument that it will not use T50 standing alone on its packaging, we consider on this appeal only whether Stanley infringes upon Arrow's rights when it uses the alpha-numerical designations described in the preceding paragraph that the district court found to be within the scope of its injunction.3 For the reasons stated below, we find that the district court erred in concluding that these uses violated Arrow's rights.

We begin with a review of the governing legal standards. To establish a trademark infringement claim under the Lanham Act, a plaintiff must show that the defendant used in commerce, without the plaintiff's consent, a "reproduction, counterfeit, copy, or colorable imitation of a registered mark in connection with the sale, offering for sale, distribution, or advertising of any goods or services on or in connection with which such use is likely to cause confusion...." 15 U.S.C. Sec. 1114(1)(a).4 To prevail under this statute, the plaintiff must show that "it has a valid mark entitled to protection and that the defendant's use of it is likely to cause confusion." Gruner + Jahr USA Publishing v. Meredith Corp., 991 F.2d 1072, 1075 (2d Cir.1993). Whether the mark is entitled to protection depends on whether it is inherently distinctive or, if merely descriptive, has acquired "secondary meaning." See Merriam-Webster, Inc. v. Random House, Inc., 35 F.3d 65, 70 (2d Cir.1994), cert. denied, --- U.S. ----, 115 S. Ct. 1252, 131 L. Ed. 2d 133 (1995). Secondary meaning attaches when " 'the name and the business have become synonymous in the mind of the public, submerging the primary meaning of the term in favor of its meaning as a word identifying that business.' " Pirone v. MacMillan, Inc., 894 F.2d 579, 583 (2d Cir.1990) (quoting Abraham Zion Corp. v. Lebow, 761 F.2d 93, 104 (2d Cir.1985)). Whether the mark is registered bears on whether it is distinctive and enjoys protection. See Gruner, 991 F.2d at 1076.

If a mark is entitled to protection, then the inquiry turns to whether "numerous ordinary prudent purchasers are likely to be misled or confused as to the source of the product in question because of the entrance in the marketplace of defendant's mark." Id. at 1077. Under the law of this Circuit, courts deciding whether a plaintiff has established likelihood of confusion must consider the eight factors elaborated by Judge Friendly in Polaroid Corp. v. Polarad Elecs. Corp., 287 F.2d 492, 495 (2d Cir.), cert. denied, 368 U.S. 820, 82 S. Ct. 36, 7 L. Ed. 2d 25 (1961). The factors, explained in greater detail below, are: (1) the strength of the mark; (2) the degree of similarity between the two marks; (3) the proximity of the products; (4) the likelihood that the prior owner will "bridge the gap" [in this case, the likelihood that Arrow will enter Stanley's market under Arrow's mark]; (5) actual confusion; (6) the defendant's good faith in adopting its mark; (7) the quality of the defendant's product; and (8) the sophistication of the buyers. Id. This "list of Polaroid factors is not exclusive and the analysis of the factors is 'not a mechanical process.' " Merriam, 35 F.3d at 70 (quoting Paddington Corp. v. Attiki Importers & Distribs., Inc., 996 F.2d 577, 584 (2d Cir.1993)).

We review the district court's treatment of each Polaroid factor under a clearly erroneous standard. Paddington, 996 F.2d at 584. Whether the plaintiff proved a likelihood of confusion is a legal question, and we review the court's weighing of those factors and its ultimate conclusion under a de novo standard. Id. at 584-85.

On appeal, Stanley does not dispute the validity of the T-50 mark. Therefore, the only question before us is whether Stanley's use of T50 in combination with other letters and numbers results in a likelihood of confusion between Stanley's and Arrow's product. To answer that question, we consider each of the Polaroid factors.

The term "strength" refers to a mark's "tendency to identify the goods sold under the mark as emanating from a particular, although possibly anonymous, source." McGregor-Doniger, Inc. v. Drizzle, Inc., 599 F.2d 1126, 1131 (2d Cir.1979). A court's inquiry regarding the strength of a mark often parallels the inquiry concerning the mark's validity, inasmuch as the "strength or distinctiveness of a mark determines both the ease with which it may be established as a valid trademark and the degree of protection it will be accorded." Id. Here, the parties do not dispute that T-50 is a valid trademark; the question is the degree of protection it enjoys.

Under the law of our Circuit, the protection afforded a mark depends on whether the mark is generic, descriptive, suggestive, arbitrary or fanciful. See id. A generic term can never be a valid trademark (and can never be registered). A descriptive term can be protected only if it has acquired secondary meaning. Suggestive, arbitrary or fanciful marks may be protected without a showing of secondary meaning. See id. A "suggestive" mark offers suggestions of, but does not actually describe, the features of a product. W.W.W. Pharmaceutical Co. v. Gillette Co., 984 F.2d 567, 572 (2d Cir.1993). An arbitrary mark has an actual dictionary meaning, but that meaning does not describe the product. A fanciful mark is a made-up term. See Gruner, 991 F.2d at 1075-76.

Relevant to the inquiry in this case is the fact that T-50 is a model number. Arrow concedes this fact but maintains that T-50 is not solely a model number. Model numbers, while often arbitrary in that they do not refer to characteristics of the item they demark, are nevertheless generally descriptive because they serve to distinguish a single source's products from each other. See J.M. Huber Corp. v. Lowery Wellheads, Inc., 778 F.2d 1467, 1469-70 (10th Cir.1985) (holding that although alphanumeric symbols on wellheads do not describe wellheads' physical characteristics, "symbols distinguish one [company] wellhead from another"); Ideal Indus., Inc. v. Gardner Bender, Inc., 612 F.2d 1018, 1023 (7th Cir.1979) (holding that numbers descriptive of size may become trademarks if they acquire secondary meaning), cert. denied, 447 U.S. 924, 100 S. Ct. 3016, 65 L. Ed. 2d 1116 (1980); CALLMANN, supra, Sec. 17.19, at 19 (Supp.1994) ("The tendency of this type of trademark to cause confusion, however, may be substantially less than that of a conventional trademark."). But cf. Eastman Kodak Co. v. Bell & Howell Document Management Prods. Co., 994 F.2d 1569, 1576 (Fed.Cir.1993) (quoting Neapco Inc. v. Dana Corp., 12 U.S.P.Q.2d 1746, 1748 (T.T.A.B.1989) ("it is possible for an alphanumerical designation, which functions only in part to designate model or grade, to be inherently distinctive")).

Stanley argues that whatever secondary meaning the T-50 model number has acquired extends strictly to the combination "T-50" and Arrow's hand-operated staple guns. Stanley insists that the record does not support a finding of secondary meaning extending to any other Arrow model number, such as HT-50, ETF-50, or ETN-50. Stanley argues further that the mark is not strong enough to extend to products with which Arrow's T-50 stapler does not compete, among them Stanley's pneumatic staplers.

Arrow contends that even though T-50 is used as a model number, "T" and "50" are both arbitrary, and that model numbers may be arbitrary for trademark protection purposes. Arrow argues in the alternative that even if T-50 is descriptive, the record supports a finding that it has acquired a strong secondary meaning. The district court found that T-50 "is not used by Arrow solely as a model number nor is it restricted to the stapling device. T-50 has no descriptive function. It has some suggestive quality but is essentially arbitrary. It has gathered secondary meaning." 870 F. Supp. at 428.

We hold that while the district court correctly concluded that T-50 is a descriptive term that enjoys secondary meaning, it erred in two related conclusions: first, that T-50 is arbitrary, and second, that T-50 extends to a "number of different products." Id.

First, in deeming T-50 to be an arbitrary mark, the district court did not afford appropriate--indeed, any--weight to the fact that T-50 describes a particular Arrow stapler. The district court concluded that "T-50" was "apparently arbitrarily selected." Id. In so concluding, the court appears to have discounted evidence produced by Stanley suggesting that "T" stood for "tacker," and "50" referred to a wire size for staples of .050. On appeal, Stanley does not press the argument that "T" and "50" are descriptive of the product's particular qualities and for that reason cannot be arbitrary.5 Rather, Stanley maintains that, in deeming T-50 to be arbitrary, the district court failed to give proper weight to the fact that T-50 is a model number.

We agree with Stanley. Under these circumstances, T-50 cannot at once be a model number and also be considered an arbitrary mark for purposes of trademark protection. The record shows that customers perceived T-50 to refer to a particular Arrow staple gun. The specific trademark registration that T-50 enjoys extends to a certain class of staplers, namely, "staple gun tackers, in Class 8." Class Eight covers hand tools, whereas Class Seven covers machine tools, such as electric and pneumatic staplers. See 37 C.F.R. Sec. 6.1. Finally, we accord weight to the USPTO's initial conclusion that T-50 "merely describes a ... model designation of the goods," and to the USPTO's decision not to register the mark without evidence of secondary meaning. Cf. Aluminum Fabricating Co. v. Season-All Window Corp., 259 F.2d 314, 316 (2d Cir.1958) (according deference, for purposes of determining validity of trademark, to USPTO's decision whether mark was descriptive or arbitrary and fanciful). In light of these facts, we hold that the district court was clearly erroneous in concluding that T-50 is "not known or regarded as a model number" and is "arbitrary" for purposes of trademark protection. 870 F. Supp. at 428, 430.

We now turn to the issue of "secondary meaning." As applied to Arrow, this inquiry asks whether the T-50 mark suggests that Arrow is the source of the product denoted by the mark. The district court found that it did, and we agree.

Whether secondary meaning has attached to a mark is a question of fact. McGregor-Doniger, 599 F.2d at 1132. The USPTO's decision to register the mark--which it initially deemed descriptive--under 15 U.S.C. Sec. 1052(f), after Arrow submitted evidence of secondary meaning, creates a rebuttable presumption that the mark enjoys secondary meaning.6 See 20th Century Wear, Inc. v. Sanmark-Stardust, Inc., 747 F.2d 81, 88-89 & n.8 (2d Cir.1984), cert. denied, 470 U.S. 1052, 105 S. Ct. 1755, 84 L. Ed. 2d 818 (1985); cf. Gruner, 991 F.2d at 1078 (registration incontestability relevant to question of strength of mark in "likelihood of confusion" inquiry); Evelyn Wood Reading Dynamics Inst. of Am. v. Zimmerman, 134 U.S.P.Q. 475, 483 (N.D.Cal.1962) (administrative decision of secondary meaning carries "strong presumption of correctness").

The defendant fails to rebut this presumption. The factors relevant to the inquiry of secondary meaning include advertising expenditures, consumer studies, sales, competitors' attempts to plagiarize the mark, and length and exclusivity of the mark's use. No single factor is dispositive. Thompson Medical Co. v. Pfizer, Inc., 753 F.2d 208, 217 (2d Cir.1985). With respect to T-50, Arrow submitted evidence of substantial advertising expenditures, high sales volumes, use of the mark for more than 30 years, and recognition of the mark by competitors. Stanley correctly points out that Arrow failed to produce market survey data, and that advertising expenditures are not determinative. See McGregor-Doniger, 599 F.2d at 1133 n.4 ("advertising and sales figures are open to varying interpretations"); J.M. Huber, 778 F.2d at 1470 (affirming district court's finding of no secondary meaning in absence of survey data or "other empirical evidence"). Nevertheless, upon our review of the evidence presented here--particularly when viewed in light of the USPTO's registration of the mark under Sec. 1052(f)--we are not "left with the definite and firm conviction that a mistake [was] committed," Inwood Labs., Inc. v. Ives Labs., Inc., 456 U.S. 844, 855, 102 S. Ct. 2182, 2189, 72 L. Ed. 2d 606 (1982) (internal quotations and citation omitted), when the district court concluded that T-50 enjoyed secondary meaning.

We do take issue, however, with the district court's finding that T-50 had acquired "strong" secondary meaning. 870 F. Supp. at 428. The court appears to have based this conclusion in part on its finding that T-50 is used on, and denotes Arrow as the source of, a number of different products. Id. This finding at best does not find support in the record and at worst is contradicted by it.

The district court apparently reached its conclusion that T-50 is not "limited to identity of a single Arrow product," id. at 430, in response to Arrow's argument that it had established a "family" of marks in which T and 50 were both used, and that this composite use denoted Arrow as the source of the products. See J & J Snack Foods Corp. v. McDonald's Corp., 932 F.2d 1460, 1462-63 (Fed.Cir.1991) (describing "family of marks" where group of marks shared characteristic indicative of common origin). Examples of the model numbers that Arrow claims fall within this family are HT-50, ETF-50, and ETN-50.

The district court does not cite, nor does our independent review of the record reveal, any evidence proving that T-50 indicates product source beyond the Arrow T-50 stapler and the staples that go with it. Arrow itself repeatedly emphasizes that when customers requested a T-50 they received, and accepted for purchase, a particular Arrow model, the T-50. These customers neither desired nor received any other Arrow product, such as the ETF-50 or HT-50.7 Moreover, Arrow offers no evidence that these latter model numbers are registered. Nor does Arrow direct our attention to evidence of secondary meaning, such as advertising expenditures and competitor recognition, for Arrow products other than the T-50 staple gun.8 The lack of evidence of secondary meaning for these model numbers is not surprising, inasmuch as Arrow also has sold products under the model numbers TN-30, T-30, T-30M, JT-21, JT-21C, T-27, T-18, T-25, T-37, T-75, HT-50A, RH200, T-25, RL-100, P-22, P-35, P35S, P-66, and DC-66.

In sum, the district court erred in concluding that T-50 was an arbitrary mark that extended beyond the specific T-50 staple gun. As a registered trademark, T-50 is descriptive and carries secondary meaning, but the mark does not enjoy the broad scope of protection that the district court accorded it.

In applying this factor, courts consider whether the similarity of the marks is likely to cause confusion among potential customers. W.W.W. Pharmaceutical, 984 F.2d at 573. In judging similarity, courts are to consider "all factors that could reasonably be expected to be perceived by and remembered by potential purchasers." McGregor-Doniger, 599 F.2d at 1133. That includes the "context in which the respective marks are generally presented." Id.

The district court stated that Arrow's T-50 mark and Stanley's T50 designation were "close, if not identical (the absence of a hyphen in T50 is found to be an insignificant difference)." 870 F. Supp. at 430. When asked, on remand, to make findings and reach conclusions specifically in light of the multi-digit numbers, the district court stated: "Letter and number combinations with T50 would suggest a product associated with the T-50 mark unless the T50 is so sublimated [sic] as to be indistinguishable to a discerning eye." Response to Mandate, at 1. To the extent that the district court concluded that Stanley's six-digit model numbers, or its other model numbers that used T and 50 in combination with other numbers and letters, were confusingly similar to Arrow's T-50 mark, this conclusion was clearly erroneous.

The similarity of the marks in this case must be weighed in light of two important considerations: first, that the housemarks of "Arrow" and "Stanley-Bostitch" appear on the respective packaging and products, and second, that similarity "is determined on the basis of the total effect of the designation, rather than on a comparison of individual features." Restatement of Torts Sec. 729 cmt. b (1938).

First, both companies prominently display their respective housemarks--Arrow and Stanley-Bostitch--on their packaging. Arrow relies on cases noting that the addition of a trade name does not necessarily alleviate the problem of confusion of marks, and indeed, can aggravate it, as "a purchaser could well think plaintiff had licensed defendant as a second user." A.T. Cross Co. v. Jonathan Bradley Pens, Inc., 470 F.2d 689, 692 (2d Cir.1972) (Friendly, C.J.) (citing Menendez v. Holt, 128 U.S. 514, 521, 9 S. Ct. 143, 144, 32 L. Ed. 526 (1888)). Stanley draws our attention to cases that note that "when a similar mark is used in conjunction with a company name, the likelihood of confusion may be lessened." W.W.W. Pharmaceutical, 984 F.2d at 573 (quoting McGregor-Doniger, 599 F.2d at 1133-34).

We might well have been receptive to Arrow's argument that the Stanley housemark aggravated rather than mitigated the possibility of confusion, had we been concerned with Stanley's initial use of T50 alone. As Arrow points out, Stanley has a history of purchasing other well-known companies, such as Bostitch, and joining its name to that of the acquired company. In light of Stanley's practice, a purchaser viewing the legend "model T50 series" arguably might have thought Stanley-Bostitch was somehow affiliated with Arrow.

Inasmuch as Stanley has represented to this court that it will not use T50 standing alone, the inquiry has shifted. Now we are considering whether Stanley's use of a six-digit model number that contains the symbol "T50" and appears on the same panel as the Stanley-Bostitch housemark is confusingly similar to Arrow's use of "Model T-50," accompanied by its housemark on the top and side panels of Arrow's packaging. (It is unclear from the record what appears on the front panel of the Arrow packaging.) This question brings us to our second consideration--the need to focus on the total effect. According to one authority:

While individual features may be dissimilar, the total effect may be one of similarity. Or the total effect may appear dissimilar despite similarities in individual features.

Restatement of Torts Sec. 729 cmt. b (1938). Here, "T" and "50" both appear in Stanley's designation and Arrow's mark, but the total effect of Stanley's multi-digit model numbers is decidedly different from Arrow's "Model T-50." Moreover, while Arrow's mark is more than a model number, tools often bear model numbers, so it is not likely that purchasers would presume that a "T" and a "50" that are grouped with other letters and numbers on a computer printout are associated with Arrow's "T-50." In sum, when these multi-digit model numbers appear near Stanley's housemark--a clear and familiar source of origin--it is unlikely that consumers will find these marks confusingly similar to Arrow's two-digit mark.

The district court, on remand, stated that the "use of T50 in a discernible manner would allow [Stanley] to secure the benefit of the implication to consumers that its product, so marked, [is] associated with the T-50 mark and the quality of the products manufactured under that mark." Response to Mandate, at 1-2. The court offers no basis for its conclusion that Stanley's use of "T50," along with other letters and numbers as part of a computer printout of a model number, constitutes a "discernible manner" of use that would lead to confusion. In short, this Polaroid factor--similarity of the marks--favors the defendant.

This factor is concerned with the competitive distance between the products. McGregor-Doniger, 599 F.2d at 1134. The district court found that "T-50 products do not directly compete with pneumatic staplers but they have a common purpose and use within the same fields, although not used in precisely the same circumstances." 870 F. Supp. at 429. Although the district court apparently found proximity between the products, its underlying findings and our review of the record do not support that conclusion.

The Arrow and Stanley products at issue here are both staplers and are sold in some of the same stores. Despite these similarities, the district court correctly recognized that the products are functionally different and hence, are not directly competitive. While the T-50 stapler and pneumatic staplers are both staplers, the district court found that "heavy-duty tools" such as pneumatic staplers, "are not purchased by buyers of Arrow's T-50 stapler." Id. The court further found that "homeowners do not use pneumatic staplers. Contractors do not use lightweight small staplers in construction." Id. These significant functional distinctions are reflected in a comparison of the products' prices: the T-50 retails for about $20, while the pneumatic stapler costs approximately $400.

The lack of direct competition between the Arrow T-50 and Stanley's pneumatic staplers is not dispositive. As we have observed, "direct competition between the products is not a prerequisite to relief." Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254, 257 (2d Cir.1987) (internal quotations and citation omitted). Indeed, if we viewed the proximity factor in a vacuum, then it would likely favor Arrow. We have made it clear that "our concern with product proximity relates to 'the likelihood that customers may be confused as to the source of the products, rather than as to the products themselves.' " Spring Mills, Inc. v. Ultracashmere House, Ltd., 689 F.2d 1127, 1134 (2d Cir.1982) (quoting McGregor-Doniger, 599 F.2d at 1134-35) (emphasis in McGregor-Doniger ). For purposes of our analysis of this factor, then, we are not asking whether a person shopping for a hand-operated stapler would accidentally purchase a pneumatic stapler upon seeing the model number. Rather, the question is whether "a purchaser could easily assume that, while the [staplers] themselves are different, they belong to the same genre of products and might well have the same source." Spring Mills, 689 F.2d at 1134. Mobil Oil, however, cautions that when considering this inquiry and the existence of "competitive proximity," we proceed "with reference to the first two Polaroid factors [strength of mark and similarity]." Mobil Oil, 818 F.2d at 258.

In Mobil Oil the court found that the proximity factor favored the plaintiff, Mobil, where a competitor in the specific field of oil trading used a name similar to Mobil's general trademark--even though Mobil did not use its trademark for oil trading. In finding proximity despite the lack of direct competition, the court noted that because Mobil's mark was strong and the defendant's use similar, Mobil's mark was entitled to protection over a "broader range of related products." Id.

The wisdom of Mobil Oil 's reasoning is evident here. We have already concluded that T-50--a model number--is not a strong mark and does not warrant broad protection for a range of products, and that Stanley's multi-digit numbers are not confusingly similar to Arrow's T-50 mark. These conclusions help inform our inquiry as to whether a purchaser would easily confuse the source of the products based on their proximity in the market. Inasmuch as the T-50 mark does not have a broad reach, Stanley's use is not confusingly similar, and these non-competitive products differ in both function and price, we conclude that the district court erred in finding that this factor favored the plaintiff. Cf. Inc. Publishing Corp. v. Manhattan Magazine, Inc., 616 F. Supp. 370, 381-85 (S.D.N.Y.1985) (finding that proximity factor favored defendant where there was "overlap" between two magazines in terms of size, frequency of publication, and demographics, but differences in editorial content, style, geographic distribution and audience appeal), aff'd, 788 F.2d 3 (2d Cir.1986).

In assessing this factor, we consider the likelihood that Arrow will enter Stanley's market. See Lang v. Retirement Living Publishing Co., 949 F.2d 576, 582 (2d Cir.1991). The district court concluded that "Arrow is not likely to produce a pneumatic stapler. Its more powerful electric stapler, being developed, could function comparably and compete with Stanley's pneumatic fasteners." 870 F. Supp. at 430. Despite Stanley's claims of functional differences between its pneumatic staplers and the electric stapler that Arrow is developing, the record does support the conclusion that Arrow is in the process of developing a competitive product. Stanley, however, points out that Arrow's former national sales manager testified that the company would not use the T-50 mark on a product that did not use T-50-sized staples because doing so would be confusing to purchasers.9 Stanley claims, and Arrow does not dispute, that the Arrow model would not use T-50 staples. There appears to be minimal likelihood that Arrow will attempt to carry its T-50 trademark over to a product that is competitive with the product for which Stanley is making an allegedly infringing use. See McGregor-Doniger, 599 F.2d at 1135-36 (affirming finding of unlikelihood that plaintiff, a manufacturer of golf jackets, would enter market of defendant, a manufacturer of women's coats, under plaintiff's mark). This factor, then, favors Stanley.

The district court did not explicitly consider this factor. Arrow did not submit any survey evidence or testimony of customers--retail or wholesale--that tended to prove actual confusion. We conclude that Arrow failed to show "actual confusion affecting purchasers." See Merriam-Webster, 35 F.3d at 72. In any event, the absence of actual confusion alone is not dispositive of the question of likelihood of confusion. Centaur Communications, Ltd. v. A/S/M Communications, Inc., 830 F.2d 1217, 1227 (2d Cir.1987).

This factor considers "whether the defendant adopted its mark with the intention of capitalizing on plaintiff's reputation and goodwill and any confusion between his and the senior user's product." Lang, 949 F.2d at 583 (quoting Edison Bros. Stores, Inc. v. Cosmair, Inc., 651 F. Supp. 1547, 1560 (S.D.N.Y.1987)). A finding of good faith may be supported by, among other things, "[s]election of a mark that reflects the product's characteristics." Id. at 583. Prior knowledge of a senior user's trade mark does not necessarily give rise to an inference of bad faith and may be consistent with good faith. Id. at 583-84. This is particularly true where the presumption of an exclusive right to use a registered mark extends only to the goods and services noted in a registration certificate. McGregor-Doniger, 599 F.2d at 1137. As noted above, Arrow's T-50 staplers are in a different registration class than Stanley's pneumatic staplers would be. See supra, II.B.1.

Here, the district court found that Stanley was aware of Arrow's mark before it made any "significant use" of the T50 series label, and that Stanley failed to respond to Arrow's initial letters of complaint regarding this use. 870 F. Supp. at 429-30. Arrow also points to evidence that Stanley was aware of the value of the mark in 1990-91, when it was considering purchasing Arrow. These discussions occurred before Stanley introduced the T50 series, but after Stanley had chosen the designation. See 870 F. Supp. at 430. Despite these findings, the district court ultimately did not conclude that Stanley acted in bad faith. The court held that Stanley's purported basis for choosing T50 was "not devoid of logic" because, as Stanley had pointed out, "T" could stand for tacker, and "50" was the metric maximum staple length. 870 F. Supp. at 430. Moreover, while the district court did not refer to it, evidence indicated that Bostitch had been using model numbers with the letter T since the 1970s, including T40S2, T31-2SBNK, T29-30, and T31-1B.

The evidence regarding this factor is mixed. But the district court's conclusion that Stanley selected a mark that reflected the product's characteristics is not clearly erroneous. Moreover, the fact that the registration of Arrow's mark extended only to hand-operated staplers, not pneumatic staplers, lends further support to the district court's conclusion that Stanley did not act in bad faith. This factor, then, favors Stanley.10

This factor is primarily concerned with whether the senior user's reputation could be jeopardized by virtue of the fact that the junior user's product is of inferior quality. Lois Sportswear, U.S.A., Inc. v. Levi Strauss & Co., 799 F.2d 867, 875 (2d Cir.1986). The district court found Stanley's products to "have considerable quality," 870 F. Supp. at 430, and Arrow does not contend otherwise.

Relying on Lois Sportswear, however, Arrow argues that the fact that Stanley's products are of high quality is not dispositive because similarities in quality may actually heighten the confusion of consumers after purchase. We agree that similarity in quality does not necessarily hurt Arrow's case. But we cannot go so far as to find the Lois Sportswear result applicable here. There, where similarities in back-pocket stitching on jeans were at issue, we found that similarities in quality could increase the likelihood of confusion after the sale, when passers-by would see similar stitching on precisely the same products--jeans. Id. at 872-73. Here, the pneumatic and hand-held staplers differ in appearance as well as function and price. And while the Lois Sportswear court found that labels identifying the source of the jeans would be discarded after purchase, id. at 873, the housemarks remain on the staplers. Thus, the circumstances creating a risk of post-sale confusion of high quality products in Lois Sportswear court are not present here. This factor does not favor Arrow.

This final factor recognizes that the likelihood of confusion between the products at issue depends in part on the sophistication of the relevant purchasers. W.W.W. Pharmaceutical, 984 F.2d at 575. The district court found that "[b]uyers of Stanley's pneumatic staplers tend to have knowledge and skill as to the tools of their trade. However, they were not shown to be sophisticated." 870 F. Supp. at 430. The court did not elaborate further.

Stanley contends that the district court erred in failing to give sufficient weight to the vast price differences between the relevant products--about $20 for the Arrow T-50 compared with $400 for Stanley's pneumatic staplers--in assessing the sophistication of the purchasers. Arrow minimizes the significance of price and maintains that even knowledgeable purchasers, indeed often knowledgeable purchasers, will likely confuse marks. Such purchasers, Arrow contends, are precisely the individuals who would be familiar with T-50 and other trademarks and seek products from the same source.

Arrow's argument finds some support in the case law. In Wincharger Corp. v. Rinco, Inc., 297 F.2d 261 (C.C.P.A.1962), for example, which concerned the sophistication of technicians in the field of electrical devices, the court stated that while technicians are "a discriminating group of people.... [b]eing skilled in their own art does not necessarily preclude their mistaking one trademark for another...." Id. at 264; see also Lois Sportswear, 799 F.2d at 875 (sophisticated jeans consumer most likely to draw association between sources based on similar marks). Moreover, while price differences are important in determining the sophistication of customers, they are not dispositive. Cf. McGregor-Doniger, 599 F.2d at 1137 (although the "greater the value of an article the more careful the typical consumer can be expected to be," the "general impression" of the consumer is the "touchstone").

In this case, however, the record reflects not only a vast difference in price, but also a "fairly detailed purchasing process" for the purchase of a Stanley pneumatic stapler. See Blue Bell, Inc. v. Jaymar-Ruby, Inc., 497 F.2d 433, 435-36 (2d Cir.1974). The various pneumatic staplers in Stanley's T50 series perform different functions, such as lathing, decking and sheathing, and furniture manufacture. With these various purposes in mind, a purchaser selects a specific model from those in the pneumatic stapler series. A consumer who must possess this high level of knowledge, and who is also paying a substantial amount of money for the product, is not likely to be confused by the use of "T50" in the product's model number. We hold that the district court's conclusion that the relevant purchasers "were not shown to be sophisticated" was clearly erroneous. This factor favors Stanley.

Having considered the district court's treatment of each of the Polaroid factors, we now review the weighing of those factors de novo. The analysis of factors overwhelmingly favors Stanley's defense against this action.

In so concluding, we emphasize what this case is not about. This case is not about whether Stanley's use of T50 alone on its packaging of pneumatic staplers, unattended by other letters or numbers, would infringe upon Arrow's mark, because Stanley has discontinued that use and has represented to this court that it will not resume it. This case is also not about whether Stanley would infringe upon Arrow's protected trademark were Stanley to use model numbers incorporating "T" and "50" for a line of hand-operated staplers or a product similarly competitive with the T-50.

This case addresses only the narrower question of whether Stanley's use of multi-digit model numbers for its pneumatic staplers infringes upon Arrow's use of a mark that also serves as a model number for a particular product, the hand-operated Arrow T-50 stapler. Although T-50 is a model number and, in that sense, is descriptive, the district court correctly concluded that it enjoys secondary meaning. The strength of the T-50 mark, however, is limited to the product that the mark identifies and others that directly compete with it. Here, where the allegedly infringing use concerns inclusion of the symbol T50 in a multi-digit model number used to identify a product different from the plaintiff's in function and in price, we cannot conclude that the plaintiff has established that " 'an appreciable number of ordinarily prudent purchasers are likely to be misled, or indeed simply confused, as to the source of the goods in question.' " W.W.W. Pharmaceutical, 984 F.2d at 571 (quoting McGregor-Doniger, 599 F.2d at 1130). For these reasons, in the circumstances before us, we reverse the judgment of the district court.

Finally, we reiterate the importance of a thorough, delineated analysis of the eight Polaroid factors by district courts in trademark cases. In this case--to be sure a difficult and onerous matter--the district court clearly assimilated and analyzed a vast amount of information. But the court did not always articulate the basis for its conclusions regarding the various Polaroid factors or whether it considered all factors relevant to the case. In several instances it appears that the district court gave undue weight to the question of the mark's strength, and viewed--in error--other Polaroid factors as indicative of the mark's strength. See 870 F. Supp. at 430 ("sophistication of purchasers" factor "lends limited support to a finding of strength"; "good faith" a factor in determining strength of mark).

We have said that the Polaroid factors are not, of course, "exclusive" and should not be applied "mechanically." Paddington, 996 F.2d at 584. No single factor is dispositive, and cases may certainly arise where a factor is irrelevant to the facts at hand. See Orient Express Trading Co., Ltd. v. Federated Dep't Stores, Inc., 842 F.2d 650, 654 (2d Cir.1988) (district courts need not "slavishly recite the litany of all eight Polaroid factors in each and every case"). But it is incumbent upon the district judge to engage in a deliberate review of each factor, and, if a factor is inapplicable to a case, to explain why. See Thompson Medical, 753 F.2d at 214 (analysis of all factors necessary to determination of infringement suit). The steady application of Polaroid is critical to the proper development of trademark law, for it is only when the Polaroid factors are applied consistently and clearly over time that the relevant distinctions between different factual configurations can emerge. Litigants are entitled to the illumination and guidance this common-law process affords, and appellate courts depend on it for the performance of their assigned task of review. As Judge Friendly recognized, the "problem of determining how far a valid trademark shall be protected ... has long been vexing and does not become easier of solution with the years." Polaroid, 287 F.2d at 495. The efficacy of the multi-factor approach that Judge Friendly wisely set out to address this difficult situation depends on thorough, careful, and consistent application of the doctrine by district courts.

For the reasons set forth above, the judgment of the district court, insofar as it extends to Stanley's current practice of using T50 as part of larger model numbers, as described in this opinion, is reversed.