OPINION OF THE COURT

SCIRICA, Circuit Judge.

In this Lanham Act case, plaintiff Checkpoint Systems, Inc. alleges Check Point Software Technologies, Inc. infringed on its trademark and engaged in unfair competition in violation of 15 U.S.C. §§ 1114, 1125(a). Finding no likelihood of confusion between Checkpoint Systems's and Check Point Software's marks, the District Court found no Lanham Act violation. We will affirm.

I. A.

Plaintiff Checkpoint Systems, Inc. has been manufacturing and distributing commercial electronic security control systems since 1967. Its devices are designed to track the physical location of goods and are sold to retailers to prevent merchandise theft. It is one of the two dominant manufacturers in the retail security products market. Since 1967 Checkpoint Systems has used the "CHECKPOINT" mark, which is registered with the United States Trademark office.

Checkpoint Systems primarily sells four types of security monitoring devices: (1) electronic article surveillance systems; (2) access control systems; (3) closed circuit television systems; and (4) radio frequency identification devices. Its principal and most successful products are electronic article surveillance systems designed to alert retailers when items are removed from confined areas. Primarily used to prevent theft of merchandise from stores and books from libraries, these systems consist of circuited tags, electronic sensors, and deactivation equipment. The systems work by placing circuited tags on merchandise which are deactivated at the time of sale. If the tags are not deactivated, an alarm sounds. Electronic article surveillance systems cost between $2000 and $5000, plus the cost of individual tags. A large retail chain, including many of Checkpoint Systems's current clients, may spend over $20 million a year on this technology.

Checkpoint Systems also manufactures electronic access control systems in the form of security cards that permit selected personnel to have access to restricted areas. Checkpoint Systems intends to use these electronic access control systems to make "smart cards" that will enable log-on access to computers and will facilitate monitoring of physical access to restricted areas. This product was in the development stage when this lawsuit commenced.

Checkpoint Systems also sells closed circuit television systems that monitor activity within confined areas and transfer data to remote locations for employee audit and inventory management security teams. Although these systems currently transfer data and images over traditional telephone lines, Checkpoint Systems intends to transfer this data over customers' computer networks. These closed circuit products can be integrated with other computer systems, including other security application programs.

In 1997, Checkpoint Systems began developing radio frequency identification devices with the view to creating "intelligent tags" for merchandise to carry information on "merchandise history." This radio technology would permit greater information storage than currently available on traditional bar code technology. These systems cost from $80,000 to $140,000.

The sale of electronic article surveillance systems to retail, industrial, institutional and government users comprises 90% of Checkpoint Systems's business. Revenues for article surveillance systems amounted to $365-380 million in 1999; revenues for access control systems were $15 million. Checkpoint Systems's electronic article surveillance systems have a 30% market share, but its systems dominate the drug store segment with a 70% market share. Its largest competitor is Sensormatic Electronic Corporation, which has a 45% share of the overall market.

Checkpoint Systems promotes its products primarily through its direct sales force, trade shows, direct mailings to security systems dealers and its internet website. It spends $8 million annually on advertising and marketing. The trade shows in which it advertises are targeted to security professionals and retailers. Checkpoint Systems concentrates its print advertising in two magazines: Security and Today's Facility Manager.

The "CHECKPOINT" trademark is displayed prominently on all of Checkpoint Systems's products including all security tags, sensors, computer screens and access control cards. Checkpoint Systems is a public company whose stock is traded on the New York Stock Exchange under the symbol "CKP."

B.

Defendant Check Point Software Technologies, Inc. writes computer programs that protect and manage access to information. Check Point Software's principal product is "firewall" technology. Firewalls are computer systems that prevent unauthorized internal or external entry into computer networks. Check Point Software's firewall systems regulate data by acting as a screen between a business's private computer networks (intranets) and the wider Internet. These systems are sophisticated software applications that cost $2,999 for a single firewall and hundreds of thousands of dollars for more complicated systems. Check Point Software firewall products must be installed and maintained by computer network information specialists.

At the time Check Point Software was founded in Israel in 1993, its founders were unaware of Checkpoint Systems. They adopted the "Check Point" mark because they believed their computer firewall technology, which limited access to data, resembled military check points that prevent access to restricted areas.

Check Point Software is recognized as a leader in the firewall market with 40% of the worldwide market share. In 1997, it expanded its product line beyond network security products to include "network management products." This software enables customers to design their own computer network security programs that can be integrated with Check Point Software programs. Check Point Software also sells "private networking software" which enables consumers to encrypt internal data so that outside users in the wider internet community will be unable to access this data when transmitted over unsecured public channels.

In its promotional materials, Check Point Software markets itself as active in the corporate security industry. Its products are installed as a part of the overall network design or "architecture" of a business's computer system. They are not functional software applications that can be purchased by a consumer and installed at will.

Check Point Software uses "Check Point" as its trade name. It has used this mark in several manifestations including, "CHECK POINT," "Checkpoint," and "CHECKPOINT."

Check Point Software promotes its network security products through trade shows, direct advertising, educational seminars, and its web site. The trade shows and seminars in which it advertises are targeted to the Internet and computer network markets. These exhibitions do not include physical security manufacturers or their products. It also advertises in two widely circulated computer trade magazines, PC Week and Network World. Its computer security programs are sold predominately to institutional and corporate clients. Check Point Software's principal competitors are computer companies like Cisco Systems and Network Associates. Check Point Software is a public company whose stock is traded on the NASDAQ under the symbol "CHKP."

C.

In early 1996, Checkpoint Systems attempted to register the internet domain name www.checkpoint.com, but discovered it was registered by Check Point Software. After its request to discontinue use was rebuffed, Checkpoint Systems filed suit alleging trademark infringement and unfair competition under the Lanham Act. In a non-jury trial, the District Court held Check Point Software neither violated the Lanham Act nor infringed Checkpoint Systems's trademark. Examining likelihood of confusion, the District Court applied the factors enumerated in Interpace Corp. v. Lapp, Inc., 721 F.2d 460, 463 (3d Cir. 1983), finding that only one factor — the strong similarity between the parties' marks — favored plaintiff Checkpoint Systems. With respect to the other Lapp factors, the court found the litigants operated in different and non-competing segments of the broad corporate security industry and that the level of consumer care in making purchasing decisions counseled against a finding of likely confusion. After determining the parties employed different marketing techniques in different industries, the court found there was no evidence that the two markets are converging. The court also found the few instances of misdirected correspondence and "minor degree of investor confusion" along with the failure to prove the existence of a single instance of a mistaken purchasing decision based on mark confusion weighed heavily against finding a Lanham Act violation. Additionally, the District Court held there was insufficient evidence of reverse confusion.

In this appeal, Checkpoint Systems contends the District Court improperly evaluated the Lapp factors in its likelihood of confusion analysis. In particular, Checkpoint Systems argues the District Court improperly required evidence of actual consumer confusion at the point of sale and disregarded evidence of initial interest and investor confusion.

II.

The District Court had jurisdiction over plaintiff's Lanham Act claims under 28 U.S.C. §§ 1331, 1338(a) and 15 U.S.C. § 1051 et seq. We have jurisdiction over the final judgment of the District Court under 28 U.S.C. § 1291. We exercise plenary review over the District Court's legal conclusions concerning the Lanham Act. Express Servs., Inc. v. Careers Express Staffing Servs., 176 F.3d 183, 185 (3d Cir. 1999), cert. denied, 531 U.S. 1052, 121 S.Ct. 656, 148 L.Ed.2d 559 (2000). We review factual findings on the likelihood of confusion for clear error. A H Sportswear Inc. v. Victoria's Secret Stores, Inc., 237 F.3d 198, 210 (3d Cir. 2000).

III.

Checkpoint Systems alleges Check Point Software engaged in unfair competition and trademark infringement in violation §§ 32 and 43(a) of the Lanham Act when it adopted the Checkpoint mark. To prove trademark infringement and unfair competition under the Lanham Act, Checkpoint Systems must prove: (1) it owns the Checkpoint mark; (2) the mark is valid and legally protectable; and (3) Check Point Software's use of the mark to identify goods or services is likely to create confusion. Commerce Nat'l Ins. Servs., Inc. v. Commerce Ins. Agency, Inc., 214 F.3d 432, 437-38 (3d Cir. 2000). Neither party disputes the first two elements. The issue on appeal is whether Check Point Software's use of the "CHECKPOINT" mark is likely to create confusion as to the source of the parties' products. Lapp, 721 F.2d at 462 ("The law of trademark protects trademark owners in the exclusive use of their marks when use by another would be likely to cause confusion."). To prove likelihood of confusion, plaintiffs must show that "consumers viewing the mark would probably assume the product or service it represents is associated with the source of a different product or service identified by a similar mark." Scott Paper Co. v. Scott's Liquid Gold, Inc., 589 F.2d 1225, 1229 (3d Cir. 1978). In Interpace Corp. v. Lapp, Inc., 721 F.2d 460, 463 (3d Cir. 1983), we set forth the relevant standards on likelihood of confusion in noncompeting goods cases, commonly known as the Lapp factors:

(1) [The] degree of similarity between the owner's mark and the alleged infringing mark;

(2) the strength of the owner's mark;

(3) the price of the goods and other factors indicative of the care and attention expected of consumers when making a purchase;

(4) the length of time the defendant has used the mark without evidence of actual confusion;

(5) the intent of the defendant in adopting the mark;

(6) the evidence of actual confusion;

(7) whether the goods, though not competing, are marketed through the same channels of trade and advertised through the same media;

(8) the extent to which the targets of the parties' sales efforts are the same;

(9) the relationship of the goods in the minds of consumers because of the similarity of functions; and

(10) other facts suggesting that the consuming public might expect the prior owner to manufacture a product in the defendant's market or that he is likely to expand into that market.

None of these factors is determinative in the likelihood of confusion analysis and each factor must be weighed and balanced one against the other. Fisons, 30 F.3d at 473 ("[N]ot all factors must be given equal weight. The weight given to each factor in the overall picture, as well as the weighing for plaintiff or defendant, must be done on an individual fact-specific basis."); see also A H Sportswear, 237 F.3d at 215 ("[T]he Lapp test is a qualitative inquiry. Not all factors will be relevant in all cases; further, the different factors may properly be accorded different weights depending on the particular factual setting."). We stated,

[W]here the plaintiff and defendant deal in non-competing goods or services, the court must look beyond the trademark to the nature of the products themselves, and to the context in which they are marketed and sold. The closer the relationship between the products, and the more similar their sales contexts, the greater the likelihood of confusion.

Fisons, 30 F.3d at 473.

As noted, the District Court applied the appropriate Lapp factors to find there was no likelihood of confusion. We will address the District Court's application of each factor.

A. Similarity of Marks — Lapp Factor (1)

Although the degree of similarity between the owner's mark and the alleged infringing mark is but one factor in the multi-factor confusion analysis, we have recognized that when products directly compete, mark similarity "may be the most important of the ten factors in Lapp." Fisons, 30 F.3d at 476; see also A H Sportswear, 237 F.3d at 216 ("The single most important factor in determining likelihood of confusion is mark similarity."). As we held in Versa Prods. Co. v. Bifold Co. (Mfg.), 50 F.3d 189, 202 (3d Cir. 1995), "unless the allegedly infringing mark . . . is substantially similar to the protectable mark . . ., it is highly unlikely that consumers will confuse the product sources represented by the different marks." "[I]f the overall impression created by marks is essentially the same, `it is very probable that the marks are confusingly similar.'" Opticians Assn. of Am., 920 F.2d at 195 (quoting J. Thomas McCarthy, 2 McCarthy on Trademarks and Unfair Competition, § 23:7 (2d ed. 1984)). In applying this test, courts attempt to "move into the mind of the roving consumer," A H Sportswear, 237 F.3d at 216, and determine "whether the labels create the same overall impression when viewed separately." Fisons, 30 F.3d at 476; see also Banff, Ltd. v. Federated Dep't Stores, Inc., 841 F.2d 486, 492 (2d Cir. 1988). Courts must "compare the appearance, sound and meaning of the marks," Harlem Wizards Entm't. Basketball, Inc. v. NBA Props., Inc., 952 F.Supp. 1084, 1096 (D.N.J. 1997), to determine whether the "average consumer, on encountering one mark in isolated circumstances of marketplace and having only [a] general recollection of the other, would likely confuse or associate the two." Fisons, 30 F.3d at 477-78.

Here, the District Court found Check Point Software's mark was very similar to Checkpoint Systems's mark. The court noted that "[w]hen the dominant portions of the two marks are the same and the overall impression created by the marks is essentially the same `it is very probable that the marks are confusingly similar.'" Checkpoint Sys., 104 F.Supp.2d at 457-58 (quoting Opticians Assn. of Am., 920 F.2d at 195). The court found the dominant portion of each parties' trademark is the word "Checkpoint." Even though Check Point Software has at times used the mark in slightly different forms including, "Checkpoint," "CheckPoint," or "Check Point," the dominant features of its mark are still the words "check" and "point." Additionally because "Systems," "Software," "Technologies," and "Inc.," are generic or descriptive terms, the District Court found their addition to the dominant terms "check" and "point" would not lead the average consumer to disassociate the products. J. Thomas McCarthy, 3 McCarthy on Trademarks and Unfair Competition, § 23:50 (4th ed. 2000) ("The Trademark Board has said that the general rule is that a subsequent user may not avoid likely confusion by appropriating another's entire mark and adding descriptive or non-descriptive matter to it."). We agree. The District Court did not clearly err in finding the overall commercial impression of the parties' marks was confusingly similar.

On appeal, Checkpoint Systems contends the District Court ignored its finding of mark similarity when it ultimately determined there was no likelihood of confusion. But mark similarity is not necessarily determinative of likely confusion, particularly when the products do not directly compete. A H Sportswear, 237 F.3d at 214 ("[W]hen the marks directly compete . . . the factor regarding the similarity of marks may increase in importance, but it does not eliminate the other factors entirely."); Fisons, 30 F.3d at 473 ("Where the goods or services are not competing the similarity of the marks is only one of a number of factors the court must examine to determine likelihood of confusion."). Here, the District Court determined that in weighing all the Lapp factors, there was not a strong likelihood of confusion. We see no error.

B. Strength of Mark — Lapp Factor (2)

Courts have recognized that

[a] strong trademark is . . . one that carries widespread, immediate recognition that one producer (even if unknown) is associated with the mark, and so with the product. If a second comer adopts a mark substantially identical to a strong mark, there is a correspondingly high likelihood that consumers will mistakenly associate the newcomer's product with the owner of the strong mark.

Versa Prods., 50 F.3d at 203.

The strength of a mark is determined by (1) the distinctiveness or conceptual strength of the mark and (2) its commercial strength or marketplace recognition. Fisons, 30 F.3d at 478-79. Under distinctiveness, we look at the inherent features of the mark. Trademarks protected under the Lanham Act are divided into four categories:

[1] arbitrary or fanciful marks [that] use terms that neither describe nor suggest anything about the product; they "bear no logical or suggestive relation to the actual characteristics of the goods." [2] Suggestive marks [that] require consumer "imagination, thought or perception" to determine what the product is.[3] Descriptive terms [that] "forthwith convey an immediate idea of the ingredients, qualities or characteristics of the goods" [and][4] generic marks . . . that "function as the common descriptive name of a product class."

A H Sportswear, 237 F.3d at 221-22 (quoting A.J. Canfield Co. v. Honickman, 808 F.2d 291, 296-97 (3d Cir. 1986)).

In Ford Motor Co., we further explained the categorization of marks stating,

Arbitrary marks are "those words, symbols, pictures, etc., which are in common linguistic use but which, when used with the goods or services in issue, neither suggest nor describe any ingredient, quality or characteristic of those goods or services." . . . Suggestive marks are virtually indistinguishable from arbitrary marks, but have been defined as marks which suggest a quality or ingredient of goods. . . . A mark is considered descriptive if it describes the "intended purpose, function, or use of the goods; of the size of the goods, of the class of users of the goods, or of the end effect upon the user."

930 F.2d at 292 n. 18 (quoting 1 McCarthy, Trademarks and Unfair Competition at § 11:3-5, 20).

While generic marks do not receive trademark protection, arbitrary, suggestive and descriptive marks with a demonstrated secondary meaning are entitled to trademark protection.

The District Court found Checkpoint Systems's mark was either suggestive or descriptive with a secondary meaning within the electronic article surveillance market. We see no error. The mark is suggestive to the extent it requires consumer imagination to determine that Checkpoint Systems's security products serve as "check points" to prevent unauthorized access and theft of merchandise. But the mark is descriptive insofar as its products serve as checkpoints. That is, they occupy a point at which customers are checked for stolen merchandise or people are checked for authorization to access restricted areas. In either event, the mark has achieved a strong secondary meaning in the physical article security market. Checkpoint Systems has spent millions of dollars in advertising its physical article security systems and the mark has achieved high recognition among consumers in this market. The District Court was therefore correct in stating, "Checkpoint [Systems] has used its name and mark in connection with its electronic article surveillance products for more than thirty years continuously, extensively, and in a substantially exclusive manner, and that long use renders the mark strong within the physical surveillance field." Checkpoint Sys., 104 F.Supp.2d at 458.

The mark has not, however, attained a strong secondary meaning in other segments of the corporate security market, particularly in the information technology market. There is no evidence that Checkpoint Systems spent a substantial amount of resources in advertising in this market, nor is there any evidence that Checkpoint Systems has achieved mark recognition in this segment of the industry. Because Checkpoint Systems's descriptive mark has not attained a secondary meaning within the information technology market, it is not conceptually strong within this sector of the corporate security market. The District Court found that Checkpoint Systems's mark "may be strong, but that strength does not appear to extend beyond its own subfield of physical security with respect to electronic article surveillance." 104 F.Supp.2d at 460. This finding is adequately supported in the record. Although Checkpoint Systems manufactured other products aside from physical article surveillance systems, the court found these products constituted a very small portion of its business. Its mark was not readily recognizable by consumers outside the electronic article surveillance market and there is "no evidence that any witness from the information security field had ever heard of [Checkpoint Systems]." Id.

Plaintiffs claim the District Court erred in its commercial strength analysis because it undervalued the overwhelming strength of the Checkpoint Systems mark in the article surveillance market. But the District Court held, properly in our view, that courts must look at the strength of the mark in the industry in which infringement is alleged. See, e.g., Homeowners Group, Inc., 931 F.2d at 1107-08 ("[A mark] may . . . have high recognition which is limited to a particular product or market segment."); Mead Data Cent., Inc. v. Toyota Motor Sales, U.S.A., Inc., 875 F.2d 1026, 1031 (2d Cir. 1989) (luxury car maker could not demonstrate that its powerful mark extended to the completely unrelated market of legal research tools); IDV N. Am., Inc. v. S M Brands, Inc., 26 F.Supp.2d 815, 824 (E.D.Va. 1998) ("A mark may be strong in the market in which it is used but weak in another market in which it is not used."). Here, the alleged infringement extends beyond the physical article security field into other, broader segments of the security industry, including computer security. The District Court did not clearly err in finding that Checkpoint Systems's mark's strength was limited to the physical article security market. Because the mark was not strong in the network access security market, it was not entitled to heightened protection within that market.

C. Factors Indicative of the Care and Sophistication of Purchasers — Lapp Factor (3)

When consumers exercise heightened care in evaluating the relevant products before making purchasing decisions, courts have found there is not a strong likelihood of confusion. Where the relevant products are expensive, or the buyer class consists of sophisticated or professional purchasers, courts have generally not found Lanham Act violations. Versa Prods., 50 F.3d at 204 ("Inexpensive goods require consumers to exercise less care in their selection than expensive ones. The more important the use of the product, the more care that must be exercised in its selection."). As a leading treatise notes,

Obviously, the price level of the goods or services is an important factor in determining the amount of care the reasonably prudent buyer will use. If the goods or services are relatively expensive, more care is taken and buyers are less likely to be confused as to source or affiliation.

3 McCarthy on Trademarks, § 23:95; see also Astra Pharm. Prods., Inc. v. Beckman Instruments, Inc., 718 F.2d 1201, 1206-07 (1st Cir. 1983) (expensive health care equipment elevated concern of purchasers). Similarly,

[w]here the relevant buyer class is composed solely of professional, or commercial purchasers, it is reasonable to set a higher standard of care than exists for consumers. Many cases state that where the relevant buyer class is composed of professionals or commercial buyers familiar with the field, they are sophisticated enough not to be confused by trademarks that are closely similar. That is, it is assumed that such professional buyers are less likely to be confused than the ordinary consumer.

3 McCarthy on Trademarks, § 23:101; see also Ford Motor Co., 930 F.2d at 293 ("Professional buyers, or consumers of very expensive goods, will be held to a higher standard of care."); Perini Corp. v. Perini Constr., Inc., 915 F.2d 121, 128 (4th Cir. 1990) ("[I]n a market with extremely sophisticated buyers, the likelihood of consumer confusion cannot be presumed on the basis of the similarity in trade name alone."); Oreck Corp. v. U.S. Floor Sys., Inc., 803 F.2d 166, 173-74 (5th Cir. 1986) (business purchasers of expensive products not likely to confuse goods with similar marks).

When the purchasing class is mixed, courts normally do not hold the general class to a high standard of care. Ford Motor Co., 930 F.2d at 293 ("When the buyer class is mixed, the standard of care to be exercised by the reasonably prudent purchaser will be equal to that of the least sophisticated consumer in the class."). If there is evidence that both average consumers and specialized commercial purchasers buy goods, there is a lower standard of care because of the lack of sophistication of some of the relevant purchasers. Id. ("Where the buyer class `consists of both professional buyers and consumers then the issue will center on the consumers, for confusion within the lowest stratum of reasonably prudent buyers may give rise to liability even if professional buyers in the market are not confused.'") (quoting Worthington Foods, Inc. v. Kellogg Co., 732 F.Supp. 1417, 1448 (S.D.Ohio 1990)).

Evaluating the care and sophistication exercised by consumers of Checkpoint Systems's and Check Point Software's products, the District Court found,

[B]oth plaintiff's products and defendant's products are expensive. The purchasers of these respective products are, for the most part, highly sophisticated, and the sale process is lengthy. . . . Plaintiff's and defendant's products are not impulse purchases, but rather are subject to long sales efforts and careful customer decision making.

Checkpoint Sys., 104 F.Supp.2d at 460.

We agree. The consumers of Checkpoint Systems's and Check Point Software's products place great importance on, and take great care in, purchasing these products. Checkpoint Systems's consumers exercise a heightened standard of care in their purchasing decisions because they need to ensure the article surveillance equipment can be quickly and easily repaired and is readily available for new merchandise. Similarly, purchasers of Check Point Software's firewall technology are highly technical computer and information specialists that must ensure the software is compatible with other programs. Because the security provided is integral to loss prevention in the one case, and confidentiality of communications in the other, the products here are essential to the customers' business needs. Because of the respective products' importance to their buyers' security needs and their high cost, consumers take care in making purchasing decisions and are not likely to be confused by the parties' similar marks.

Checkpoint Systems contends that not all the relevant consumers are sophisticated purchasers and that many "mom and pop stores," which exercise a lower standard of care, purchase both parties' products. The District Court acknowledged these smaller commercial purchasers were part of the relevant consumer class and adjusted its evaluation of consumer sophistication accordingly. Because the systems are expensive and require on-going maintenance, the court found these smaller commercial consumers view the purchase of both parties' products as an important investment decision. The court also noted that many small businesses turn to outside experts to assist in the purchase of complicated computer and article surveillance systems. Additionally, only computer specialists with significant training can install and make informed purchasing decisions about Check Point Software's complicated firewall technology. Furthermore, expert witnesses testified that smaller "mom and pop stores" would likely not have need for the complicated security software manufactured by Check Point Software.

The District Court properly evaluated the care exercised by consumers of these products in making purchasing decisions. The evidence supports the finding that the purchase of both parties' products involve a significant investment and even smaller commercial consumers exercise a heightened degree of care in evaluating the products and making purchasing decisions. We see no clear error.

D. Intent of the Defendant in Adopting the Mark — Lapp Factor (5)

While evidence of a party's intentional use of another party's mark to cause confusion is not a prerequisite to proving a Lanham Act violation, see generally Lois Sportswear, U.S.A., Inc. v. Levi Strauss Co., 799 F.2d 867, 875 (2d Cir. 1986), courts have recognized that evidence of "intentional, willful and admitted adoption of a mark closely similar to the existing marks" weighs strongly in favor of finding the likelihood of confusion. National Football League Props., Inc. v. New Jersey Giants, Inc., 637 F.Supp. 507, 518 (D.N.J. 1986). Here, the District Court found, "[t]here is no evidence or even inference that defendant chose its name with plaintiff's name or products in mind." Checkpoint Sys., 104 F.Supp.2d at 465. We agree.

E. Relationship of goods in the minds of consumers because of the similarity of product functions — Lapp Factor (7)

Under this prong, courts examine whether buyers and users of each parties' goods are likely to encounter the goods of the other, creating an assumption of common source affiliation or sponsorship. Fisons, 30 F.3d at 481 ("The question is whether the consumer might . . . reasonably conclude that one company would offer both of these related products."). The test is whether the goods are similar enough that a customer would assume they were offered by the same source. Wynn Oil Co. v. Thomas, 839 F.2d 1183, 1187 (6th Cir. 1988).

Checkpoint Systems contends that both parties operate in the overlapping business of corporate security. While each serves a different branch of this broad industry (i.e., Checkpoint Systems primarily focuses on physical security while Check Point Software focuses on information and computer security), each is a smaller piece in the overall corporate security industry. Checkpoint Systems contends that in the increasingly technical and sophisticated business world, a reasonable business consumer might believe that Checkpoint Systems, which had developed a niche in physical article security, had expanded into the computer technology market. See Dreamwerks Prod. Group, Inc. v. SKG Studio, 142 F.3d 1127, 1131 (9th Cir. 1998) ("[R]elatedness of each company's prime directive isn't relevant. Rather, we must focus on [plaintiff's] customers and ask whether they are likely to associate the[plaintiff's products] with [the defendant]."). Checkpoint Systems argues the increasing awareness among security professionals of both physical and computer information security systems has led professionals to be cognizant of both companies' products.

Checkpoint Systems also contends it has developed integrated circuit technology in its radio frequency and closed circuit television products that allow consumers to link Checkpoint Systems hardware with their computer systems. It argues these physical security products employ secure information lines similar to the firewall products that Check Point Software manufactures. On the basis of these products, it contends its physical security products overlap with computer and information security products making consumers likely to associate Checkpoint Systems's mark with Check Point Software's mark.

But the District Court found the parties operate in separate and distinct segments of the overall field of corporate security. Plaintiff's sphere includes physical security and control of flow of corporate goods and people. . . . Defendant's sphere includes electronic information security on computer networks at the point of connection to the Internet and within the customer's intranet. Their products are not substitutes for each other. . . . [Their products] are advertised in different magazines and are promoted in entirely different trade shows.

Checkpoint Sys., 104 F.Supp.2d at 466. The court continued,

That there is no overlap in the places these parties market their products tends to diminish the chance that someone who is a specialist in network security will even come across an advertisement for plaintiff, and vice versa.

Id.

The District Court did not clearly err. The "relatedness" analysis is intensely factual. Goods may fall under the same general product category but operate in distinct niches. When two products are part of distinct sectors of a broad product category, they can be sufficiently unrelated that consumers are not likely to assume the products originate from the same mark. See, e.g., Commerce Nat'l Ins. Servs., Inc., 214 F.3d at 441 (holding marks held by company operating in banking industry and company operating in insurance industry did not create consumer confusion because the two companies were involved in distinct highly regulated industries); Astra Pharm. Prods., 718 F.2d at 1207 (finding no product similarity in medical technology sold to different departments in hospital because "`the hospital community' is not a homogeneous whole, but is composed of separate departments with diverse purchasing requirements, which, in effect, constitute different markets for the parties' respective products."); Harlem Wizards, 952 F.Supp. at 1095 (finding no product similarity between professional competitive basketball team and "show basketball" team).

The District Court held that Checkpoint Systems's products are simply different from Check Point Software's products. While Checkpoint Systems's access control, closed circuit television and radio frequency products may employ similar technology, their purpose is physical article surveillance or personal access. On the other hand, Check Point Software's firewall technology is not intended to prevent theft of merchandise or limit physical access. Its purpose is to prevent third parties from accessing information from unsecure computer lines. Because the products serve different functions, and there is only "minimal overlap" in the product technology, it is unlikely consumers would be confused by the similar marks. Fisons, 30 F.3d at 481. The District Court did not clearly err in its application of the product similarity prong of the confusion analysis. See, e.g., Nutri/System, Inc. v. Con-Stan Indus., 809 F.2d 601, 606 (9th Cir. 1987) (weight loss centers and weight loss counseling differed in methods, customer groups and facilities); Plus Prods. v. Plus Disc. Foods, Inc., 722 F.2d 999, 1008 (2d Cir. 1983) (bargain food and health food were not so similar as to create likelihood of confusion).

F. The extent to which the parties' goods are marketed through the same channels of trade — Lapp Factor (9)

Courts have recognized that "the greater the similarity in advertising and marketing campaigns, the greater the likelihood of confusion." Acxiom Corp. v. Axiom, Inc., 27 F.Supp.2d 478, 502 (D.Del. 1998). Applying this factor, courts must examine the trade exhibitions, publications and other media the parties use in marketing their products as well as the manner in which the parties use their sales forces to sell their products to consumers. This is a fact intensive inquiry. Id.

Here, there is no evidence that a single security expert has sufficient knowledge in both the physical security and information security realms that he purchases both of these products for his corporation.

Here, the District Court found the parties "products are advertised in different magazines and are promoted in entirely different trade shows." Checkpoint Sys., 104 F.Supp.2d at 466. Checkpoint Systems advertises in trade shows and trade publications that are marketed to physical and retail security specialists. Check Point Software advertises in publications and at trade shows that are marketed to computer information specialists. As the District Court found,

There is no evidence that plaintiff's products and defendant's products were offered at the same time in any magazine, trade show, or distribution network. . . . That there is no overlap in the places these parties market their products tends to diminish the chance that someone who is a specialist in network security will even come across an advertisement for plaintiff, and vice versa.

Id.

Additionally, Check Point Software's products are only sold to consumers through its "specialized value added resellers." These "resellers" do not sell physical article security systems. On the other hand, consumers can purchase Checkpoint Systems's physical security products from its direct mailings and product catalogues or through its direct sales force. For these reasons, the District Court concluded, "[T]he parties do not market their products through the same trade channels or markets, and . . . their target customers generally belong to distinctly different groups." Id. at 467. We see no clear error.

G. The extent to which targets of the parties' sales efforts are the same — Lapp Factor (8)

We have recognized that when parties target their sales efforts to the same consumers, there is a stronger likelihood of confusion. Lapp, 721 F.2d at 463-64. This analysis too is intensely factual.

The District Court found the parties market their products to different users. Checkpoint Systems markets its products to physical security consumers. Check Point Software markets its products to MIS professionals. Even though many companies often purchase both types of products, most rely on information specialists to make purchasing decisions about network security systems, especially since Check Point Software's firewall systems must be installed as part of the overall architecture of a business's total computer system. But information specialists are not essential to make purchasing decisions about physical article security systems.

On the basis of this finding, the court concluded there is not a strong likelihood that users and consumers of the parties' products were likely to come across advertisements about the other party's products. The court stated, "That there is no overlap in the places these parties market their products tends to diminish the chance that someone who is a specialist in network security will even come across an advertisement for plaintiff, and vice versa." Checkpoint Sys., 104 F.Supp.2d at 466. The court did not clearly err in its determination.

H. Evidence of Converging markets — Lapp Factor (10)

We have held that evidence that the markets in which the parties sell their goods are converging is "pivotal in non-competing products cases." Lapp, 721 F.2d at 463. "One of the chief reasons for granting a trademark owner protection in a market not his own is to protect his right someday to enter that market. When it appears extremely likely . . . that the trademark owner will soon enter the defendant's field, this . . . factor weighs heavily in favor of injunctive relief." Id. (internal citation omitted). Under this factor we look not only to evidence that a plaintiff has actually moved into the defendant's market, but also to "other facts suggesting that the consuming public might expect the prior owner to manufacture a product in the defendant's market, or that it is likely to expand into that market." Id. at 463. Evaluating this factor, courts look to evidence that other companies sell products in both markets, as well as evidence that the products at issue are so closely related that the consuming public might find it natural for one company to do so. Here, the District Court found the physical article security market in which Checkpoint Systems sold its products was not converging with the computer information security market in which Check Point Software sold its firewall products. Nor would consumers expect that Checkpoint Systems would naturally expand into the network information security market. The court stated, "[T]he weight of the evidence indicates that, if anything, the fields are diverging as network security becomes more and more specialized, and the interrelatedness of Internet and intranets becomes more complex." Checkpoint Sys., 104 F.Supp.2d at 452. The court noted that some of Checkpoint Systems's competitors in the physical article security field had attempted to move into the information security field. But these competitors had been largely unsuccessful in expanding into this field thereby

undermin[ing] . . . the idea that network security may be within the zone of natural expansion for companies that once had their focus on physical security. . . . Network security . . . is a highly specialized field dominated by companies such as [Check Point Software] which are experts in the intricacies of electronic networks and the techniques of electronic intrusion. Nothing in the record indicates that plaintiff had the requisite expertise or the interest in achieving entry into [Check Point Software's] field.

Id. at 467 n. 17.

Checkpoint Systems disputes this finding contending that modern security specialists look to purchase network security systems that work hand in hand with physical security systems. It argues these specialists recognize that computer security products are worthless if not physically secure. Consumers, they claim, are likely to assume that a company that sells products in the physical security market is likely to expand into the network security market to provide businesses with complete security packages. But the District Court found that the increasingly complex market for network security systems requires expertise to understand evolving technology. Because of the highly specialized and technical nature of network security systems, consumers would not likely assume that a company that has operated for over 30 years in the distinct physical security market would have the expertise to move into the complicated network security market. Indeed Checkpoint Systems's own security expert witness testified that he would turn to a network security professional to assist in the purchase of this type of equipment because the knowledge required to make an informed purchasing decision was highly technical and specialized. Thus, although consumers may require security products from both the physical security market and the information security market, it is reasonable to assume they would turn to different companies to purchase these products.

Additionally, there is little evidence that either party currently operates in both the physical security and network security markets. On appeal, Checkpoint Systems points to several of its products, including its physical access control systems, its "smart card" technology, its radio frequency identification devices, and its closed circuit television systems as examples of products that operate in both the physical security and network information security markets. It contends these products employ similar secure network lines to Check Point Software's network security technology in order to provide physical article security. But even though its products employ some overlapping computer technology, Checkpoint Systems's products are intended to provide physical article and access security and are sold to consumers looking to purchase physical security products. Check Point Software's products are sold to consumers looking to purchase network information security products. The District Court's finding is supported by the evidence. We see no clear error.

I. Evidence of actual confusion and the length of time the defendant has used the mark without evidence of actual confusion — Lapp Factors (4), (6)

Evidence of actual confusion is not required to prove likelihood of confusion. Versa Prods., 50 F.3d at 205; Fisons, 30 F.3d at 476 ("[W]hile evidence of actual confusion would strengthen plaintiff's case, it is not essential."). We have recognized that it is difficult to find evidence of actual confusion because many instances are unreported. For this reason, evidence of actual confusion may be highly probative of the likelihood of confusion. Frisch's Rests., Inc. v. Elby's Big Boy of Steubenville, Inc., 670 F.2d 642, 648 n. 5 (6th Cir. 1982); see also Tisch Hotels, Inc. v. Americana Inn, Inc., 350 F.2d 609, 612 (7th Cir. 1965) ("[S]ince reliable evidence of actual confusion is difficult to obtain in trademark and unfair competition cases, any such evidence is substantial evidence of likelihood of confusion."). "If a defendant's product has been sold for an appreciable period of time without evidence of actual confusion, one can infer that continued marketing will not lead to consumer confusion in the future. The longer the challenged product has been in use, the stronger this inference will be." Versa Prods., 50 F.3d at 205.

Here the District Court found, "Though there is some evidence in the record of temporary initial confusion between the companies, by experts, the media, investors and consumers, there is no evidence of actual customer confusion in connection with the purchase of defendant's products." Checkpoint Sys., 104 F.Supp.2d at 447. The court continued stating,

[N]o customer or potential customer for defendant's products has believed, to any meaningful degree, that the products originated with the plaintiff. Conversely, of the few customers of plaintiff, or other persons in the media, who momentarily thought plaintiff might have entered the firewall business, not one took any meaningful action while under that mistaken belief.

Id. at 450. The court concluded,

Here, there is some evidence of initial confusion by the media and investors when defendant first gained success in the stock market, but that confusion has dissipated, rather than grown, over the years, such that there appears to have been none in the last few years. Moreover, the significance of such initial confusion by the media and investors, like initial confusion by consumers, lies only in assessing whether it supports the conclusion that confusion in consumer purchasing decision is likely.

Id. at 464.

On appeal Checkpoint Systems argues the District Court erred because it limited its actual confusion inquiry to evaluating evidence of consumer confusion at the point of sale rather than according weight to evidence of investor confusion and other initial interest confusion. It contends the Lanham Act affords protection against other types of confusion, including initial interest confusion, investor confusion and post-sale confusion.

Several courts of appeals have found initial interest confusion and post-sale confusion actionable under the Lanham Act. We agree and hold initial interest confusion is actionable under the Lanham Act. We also hold the District Court properly evaluated the evidence of initial interest confusion and did not clearly err in finding this evidence did not weigh heavily in favor of finding likely confusion. We first turn to a discussion of the relevance of initial interest to Lanham Act claims.

1.

A leading treatise on trademark infringement states:

[Trademark infringement] can be based upon confusion that creates initial customer interest, even though no actual sale is finally completed as a result of the confusion.

* * *

[For example,] the likelihood that a potential purchaser of a specialized computer program may be drawn to the junior user, thinking it was the senior user, is actionable "confusion" even if over the course of several months of the purchasing decision-making process, the buyer's confusion is dissipated. Such a senior user who is the opposer may suffer injury if a potential purchaser is initially confused between the parties' respective marks in that the opposer may be precluded from further consideration by a potential purchaser in reaching his or her buying decision.

3 McCarthy on Trademarks and Unfair Competition, § 23:6 (internal quotations and footnotes omitted). Several courts of appeals, including the Second, Seventh, and Ninth Circuits have found initial interest confusion actionable under the Lanham Act. In Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254 (2d Cir. 1987), the United States Court of Appeals for the Second Circuit found actionable initial interest confusion when it enjoined the defendant from using the trade name Pegasus because the name was confusingly similar to the plaintiff's "flying horse" mark. The court noted that in the oil industry, companies often attract potential consumers by making "cold calls" to determine interest in purchasing products. The court also noted that on the basis of the similar name and mark, it was likely that "Pegasus Petroleum would gain crucial credibility during the initial phases of a deal. For example, an oil trader might listen to a cold phone call from Pegasus Petroleum . . . when otherwise he might not, because of the possibility that Pegasus Petroleum is related to Mobil." Id. at 259. Because the similar marks could ultimately affect a consumer's consideration of the defendant's product as well as affect the plaintiff's goodwill with its customers, the court found actionable confusion under the Lanham Act. Id. at 260 ("[P]otential purchasers would be misled into an initial interest in Pegasus Petroleum. Such initial confusion works a sufficient trademark injury."); see also Lois Sportswear, 799 F.2d at 867; Grotrian, Helfferich, Schulz, Th. Steinweg Nachf v. Steinway and Sons, 365 F.Supp. 707, 717 (S.D.N.Y. 1973) ("Misled into an initial interest, a potential Steinway buyer may satisfy himself that the less expensive Grotrian Steinweg is at least as good, if not better, than a Steinway. Deception and confusion thus work to appropriate [Steinway's] good will."), aff'd, 523 F.2d 1331 (2d Cir. 1975).

Similarly in Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036, 1057 (9th Cir. 1999), the United States Courts of Appeals for the Ninth Circuit recognized that initial interest confusion is actionable under the Lanham Act. In Brookfield Communications, plaintiffs alleged that defendant West Coast Entertainment's use of the web site www.moviebuff.com infringed on its "MovieBuff" mark. Brookfield used the "MovieBuff" mark on computer software that it sold to industry professionals. This software contained a searchable database of entertainment related data, including information about film release schedules, box office receipts, films in development and other entertainment news. West Coast's "moviebuff.com" website was a free searchable entertainment database that assisted its consumers in making educated decisions about the rental and purchase of movies. The website also sold movie and other entertainment accessories. The court found actionable confusion under the Lanham Act reasoning:

People surfing the Web for information on "MovieBuff" may confuse "MovieBuff" with the searchable entertainment database at "moviebuff.com" and simply assume that they have reached Brookfield's website. . . . Alternatively, they may incorrectly believe that West Coast licensed "MovieBuff" from Brookfield . . . or that Brookfield otherwise sponsored West Coast's database. Other consumers may simply believe that West Coast bought out Brookfield or that they are related companies.

Yet other forms of confusion are likely to ensue. Consumers may wrongly assume that the "MovieBuff" database they were searching for is no longer offered, having been replaced by West Coast's entertainment database, and thus simply use the services at West Coast's web ste. And even where people realize, immediately upon accessing "moviebuff.com," that they have reached a site operated by West Coast and wholly unrelated to Brookfield, West Coast will still have gained a customer by appropriating the goodwill that Brookfield has developed in its "MovieBuff" mark. A consumer who was originally looking for Brookfield's products or services may be perfectly content with West Coast's database (especially as it is offered free of charge); but he reached West Coast's site because of its use of Brookfield's mark as its second-level domain name, which is a misappropriation of Brookfield's goodwill by West Coast.

Id. at 1057 (internal citations omitted); see also Interstellar Starship Servs., Ltd. v. Epix Inc., 184 F.3d 1107, 1110 (9th Cir. 1999) ("We recognize a brand of confusion called `initial interest' confusion, which permits a finding of a likelihood of confusion although the consumer quickly becomes aware of the source's actual identity and no purchase is made as a result of the confusion."), cert. denied, 528 U.S. 1155, 120 S.Ct. 1161, 145 L.Ed.2d 1073 (2000); but see Bryce J. Maynard, Note, The Initial Interest Confusion Doctrine and Trademark Infringement on the Internet, 57 Wash Lee L.Rev. 1303 (urging caution in finding actionable initial interest confusion on the Internet).

In Elvis Presley Enters., Inc. v. Capece, 141 F.3d 188 (5th Cir. 1998), the United States Court of Appeals for the Fifth Circuit also recognized initial interest confusion. The court there found the plaintiff assignees of the rights of the estate of Elvis Presley were entitled to injunctive relief against the defendant bar owner's use of the name "The Velvet Elvis" because it was likely to create confusion. In finding initial interest confusion the court reasoned,

The witnesses all testified that, upon entering and looking around the [defendants'] bar, they had no doubt that [plaintiff was] not affiliated with it in any way. Despite the confusion being dissipated, this initial-interest confusion is beneficial to the Defendants because it brings patrons in the door; indeed, it brought at least one of the [plaintiff's] witnesses into the bar. Once in the door, the confusion has succeeded because some of the patrons may stay, despite realizing that the bar has no relationship with [plaintiff]. This initial-interest confusion is even more significant because the Defendants' bar sometimes charges a cover charge for entry, which allows the Defendants to benefit from initial-interest confusion before it can be dissipated by entry into the bar.

Id. at 204.

The United States Court of Appeals for the Seventh Circuit has equated initial interest confusion to a "bait and switch scheme." In Dorr-Oliver, Inc. v. Fluid Quip, Inc., 94 F.3d 376, 382 (7th Cir. 1996), the court stated,

[T]he Lanham Act forbids a competitor from luring potential customers away from a producer by initially passing off its goods as those of the producer's, even if confusion as to the source of the goods is dispelled by the time any sales are consummated. This "bait and switch" of producers, also known as "initial interest" confusion, will affect the buying decisions of consumers in the market for the goods, effectively allowing the competitor to get its foot in the door by confusing consumers.

Id. (internal citation omitted); see also Eli Lilly Co. v. Natural Answers, Inc., 233 F.3d 456, 464 (7th Cir. 2000) ("[Initial interest] confusion, which is actionable under the Lanham Act, occurs when a consumer is lured to a product by its similarity to a known mark, even though the consumer realizes the true identity and origin of the product before consummating a purchase.").

We join these circuits in holding that initial interest confusion is probative of a Lanham Act violation. Without initial interest protection, an infringer could use an established mark to create confusion as to a product's source thereby receiving a "free ride on the goodwill" of the established mark. Mobil Oil Corp., 818 F.2d at 260 (likelihood that "potential purchasers would be misled into an initial interest" justifies finding of infringement). Confining actionable confusion under the Lanham Act to confusion present at the time of purchase would undervalue the importance of a company's goodwill with its customers.

Congress recognized the relevance of initial interest confusion and its effect on a company's goodwill when it amended the Lanham Act in 1962. In its original form, the Lanham Act only applied where the use of similar marks was "likely to cause confusion or mistake or to deceive purchasers as to the source of origin of such goods or services." Esercizio v. Roberts, 944 F.2d 1235, 1245 (6th Cir. 1991) (quoting 1946 Lanham Act.) (emphasis added). In 1962 Congress deleted the term "purchasers," affording Lanham Act protection where a mark is "likely to cause confusion or mistake or to deceive." Id. Several courts have found the amendment expanded the reach of the Lanham Act beyond mere purchasers to recognize pre-sale confusion as well as post-sale confusion. See, e.g., id. ("Since Congress intended to protect the reputation of the manufacturer as well as to protect purchasers, the Act's protection is not limited to confusion at the point of sale."); Koppers Co. v. Krupp-Koppers GmbH, 517 F.Supp. 836, 843-44 (W.D.Pa. 1981) (holding Congress intended to broaden the scope of protections with the 1962 Amendment to the Lanham Act); see also 3 McCarthy on Trademarks Unfair Competition, § 23.7 ("In 1962, Congress struck out language in the Lanham Act which required confusion, mistake or deception of `purchasers as to the source of origin of such goods and services.' Several courts have noted this expansion of the test of infringement and held that it supports a finding of infringement when even non-purchasers are deceived."). We agree with the view that Congress's amendment of the Lanham Act in 1962 expanded trademark protection to include instances in which a mark creates initial interest confusion.

Check Point Software notes that some circuits have only found "initial interest" confusion probative of likely confusion when the products at issue are direct competitors or where there is a strong interrelatedness between the products. It claims these circuits have recognized that initial interest confusion is probative of a Lanham Act violation when the defendant attempts to use the plaintiff's more recognizable mark to "get a foot in the door" with potential customers. But if the initial interest confusion does not ultimately result in a purchasing decision, this factor counsels against finding the likelihood of confusion under the Lanham Act.

Some courts have looked at the relatedness of the relevant products in evaluating whether initial interest confusion is probative of likely confusion. See Hasbro, Inc., 232 F.3d at 2 ("Certainly in a case involving such disparate products and services as this, the court's refusal to enter the `initial interest confusion' thicket is well taken given the unlikelihood of `legally significant' confusion."); Brookfield Communications, 174 F.3d at 1056 (when products are unrelated "the likelihood of confusion would probably be remote."). Other courts look at evidence that the defendant intentionally adopted the plaintiff's mark to create confusion among consumers making purchasing decisions. See, e.g., Dorr-Oliver, 94 F.3d at 382-83. Still other courts view the level of care consumers exercise in their purchasing decisions to determine whether initial interest confusion has dissipated prior to a purchasing decision. See, e.g., Brookfield Communications, 174 F.3d at 1057 (finding initial interest confusion actionable because "[I]n the Internet context, . . . entering a web site takes little effort — usually one click from a linked site or a search engine's list; thus, Web surfers are more likely to be confused as to the ownership of a web site than traditional patrons of a brick-and-mortar store would be of a store's ownership."); Rust Env't Infrastructure, Inc. v. Teunissen, 131 F.3d 1210, 1217 (7th Cir. 1997) (initial interest confusion unlikely with sophisticated consumers). Safeway Stores, Inc. v. Safeway Disc. Drugs, 675 F.2d 1160, 1167 (11th Cir. 1982) ("Short-lived confusion or confusion of individuals casually acquainted with a business is worthy of little weight.").

Product relatedness and level of care exercised by consumers are relevant factors in determining initial interest confusion. When products are similar, a firm is more likely to benefit from the goodwill of a firm with an established mark. And when consumers do not exercise a high level of care in making their decisions, it is more likely that their initial confusion will result in a benefit to the alleged infringer from the use of the goodwill of the other firm. Conversely, in the absence of these factors, some initial confusion will not likely facilitate free riding on the goodwill of another mark, or otherwise harm the user claiming infringement. Where confusion has little or no meaningful effect in the marketplace, it is of little or no consequence in our analysis. Nonetheless, we decline to issue a blanket rule limiting the probative value of initial interest. Its significance will vary, and must be determined on a case-by-case basis. As with all cases involving the likelihood of confusion under the Lanham Act, courts should employ all the relevant Lapp factors and weigh each factor to determine whether in the totality of the circumstances marketplace confusion is likely. See A H Sportswear, 237 F.3d at 215; Fisons, 30 F.3d at 476 n. 11.

2.

We now turn to whether the District Court properly evaluated the evidence of actual confusion, including Checkpoint Systems's evidence of initial interest confusion. It is undisputed that Checkpoint Systems presented no evidence of actual consumer confusion in the form of mistaken purchasing decisions. But it presented the following evidence of initial interest confusion: (1) e-mails that Checkpoint Systems received inquiring about firewall products; (2) customer inquiries to Checkpoint Systems employees, including its CEO, that expressed an interest in purchasing firewalls; (3) service calls made to Checkpoint Systems that were intended for Check Point Software; (4) misdirected calls to Checkpoint Systems's receptionist that were intended for Check Point Software; and (5) reports in the media that confused Checkpoint Systems and Check Point Software. The District Court found this evidence insufficient to support a finding of likely confusion reasoning,

[A]ny evidence by the plaintiff of initial interest confusion by consumers and initial confusion by some reporters or other observers, is de minimis and weak, and the record is completely devoid of any evidence of an actual mistaken purchase or attempt to purchase one company's products believing they originated with the other company. Especially when viewed in light of the parties' co-existence for over five years, this lack of evidence of any confusion in purchasing decisions of either company's products falls strongly in defendant's favor.

Checkpoint Sys., 104 F.Supp.2d at 464-65.

On appeal Checkpoint Systems claims the District Court erred as a matter of law by requiring it to produce evidence of actual consumer confusion in the form of mistaken purchasing decisions. It also contends the court erred in discounting its evidence of initial interest confusion as de minimis. See Tisch Hotels, Inc., 350 F.2d at 612.

But the District Court properly evaluated the evidence of initial interest confusion within the context of the relevant Lapp factors stating,

This Court does not read [the relevant case law] as holding that the initial interest confusion doctrine is never applicable if the parties are not competitors. However, Hasbro, and numerous other cases, correctly emphasize the importance of the parties' competition or strong interrelatedness.

A review of the cases which have applied the initial interest doctrine indicates that the parties are either direct competitors, or strongly interrelated such that it could be expected that plaintiff would expand into defendant's market. Other cases have also pointed out that one element of the applicability of the initial interest confusion doctrine is an intent by the defendant to deceive.

Checkpoint Sys., 104 F.Supp.2d at 462 (citations omitted). The court concluded,

[W]hile evidence of temporary initial interest confusion is not completely irrelevant, such evidence is entitled to less weight than it might be in a case where the parties compete or are strongly interrelated, or where there is evidence or even an inference that defendant was trying to trade on plaintiff's goodwill. These factors are utterly absent here. Moreover, the evidence of initial interest confusion in this case is de minimis.

Id. at 462-63.

This is the proper analysis. Checkpoint Systems's and Check Point Software's products are unrelated and are sold in different markets to consumers who exercise a heightened degree of care in making purchasing decisions. The markets for the parties' products are not converging, nor is there evidence that Check Point Software intentionally adopted the Check Point mark to create confusion. Given these factors, the District Court did not clearly err in finding no likelihood of confusion.

Nor did the District Court err in finding the evidence of initial interest confusion was de minimis. At oral argument, Checkpoint Systems averred that at most there was evidence of twenty instances of initial interest confusion over the period of time the parties operated in the United States. Some of this confusion consisted of misdirected calls and therefore it is uncertain whether the consumers were confused by the parties' similar names or whether directory assistance erred in connecting consumers with the parties. Duluth News-Tribune v. Mesabi Publ'g Co., 84 F.3d 1093, 1098 (8th Cir. 1996) (misdirected communications are not evidence of confusion where the sender knows which party he or she wished to send his communication but erred in the delivery of the message). Furthermore, the District Court properly took into account the potential bias of Checkpoint Systems's employees who testified they had been approached by consumers interested in Check Point Software's firewall products as well as Checkpoint Systems's expert security witnesses who testified they erroneously believed that Check Point Software was affiliated with Checkpoint Systems. A H Sportswear, 237 F.3d at 227; Reed-Union Corp. v. Turtle Wax, Inc., 77 F.3d 909, 911-12 (7th Cir. 1996) (fact finder's credibility judgments about confusion testimony are virtually never clearly erroneous).

We agree with the District Court that this limited initial interest confusion (i.e., the handful of e-mails and other anecdotal evidence of mistaken consumer inquiries) was de minimis when viewed in light of the length of time the parties operated together in the United States without significant evidence of confusion. Versa Prods., 50 F.3d at 205 ("If a defendant's product has been sold for an appreciable period of time without evidence of actual confusion, one can infer that continued marketing will not lead to consumer confusion in the future. The longer the challenged product has been in use, the stronger this inference will be."); Universal Money Ctrs. v. AT T, 22 F.3d 1527, 1535 (10th Cir. 1994) (a few instances of confusion over a long period of time are insufficient to support a finding of confusion); Barre-National, Inc. v. Barr Labs., Inc., 773 F.Supp. 735, 744 (D.N.J. 1991) ("Use of similar marks for a substantial period of time with no confusion among consumers may create a presumption that there is little likelihood of confusion."). Nor were the isolated instances of confusion in the media probative of likely confusion. Centaur Communications, Ltd. v. A/S/M Communications, Inc., 830 F.2d 1217, 1227 (2d Cir. 1987) (newspaper mis-attributions were isolated incidents not probative of actual confusion). Given the size of these companies, and the large number of e-mails, customer inquiries, and other communications they receive on a daily basis, these isolated instances of initial interest confusion counsel against a finding of likely confusion. A H Sportswear, 237 F.3d at 227 ("isolated" instances of confusion insufficient to support finding of likely confusion); Scott Paper Co., 589 F.2d at 1231 (nineteen misdirected letters over four years weighed against finding of confusion); Nippondenso Co. v. Denso Distribs., Civ. No. 86-3982, 1987 WL 10703 at *5 (E.D.Pa. May 11, 1987) ("The weight attributed to instances of actual confusion depends to some extent upon the circumstances of the case, including the volume of sales of the goods involved. If isolated instances of actual confusion result from the sale of a substantial number of goods, the weight given to the evidence of actual confusion is markedly reduced.") (citing Scott Paper Co., 589 F.2d at 1225). The District Court did not clearly err.

3.

On appeal, Checkpoint Systems contends the District Court erred in evaluating evidence of investor confusion. While a handful of district courts have found evidence of investor confusion probative of a Lanham Act violation, see, e.g., Acxiom Corp. v. Axiom, Inc., 27 F.Supp.2d 478 (D.Del. 1998); Koppers Co. v. Krupp-Koppers GmbH, 517 F.Supp. 836 (W.D.Pa. 1981), no appellate court has yet so held.

In Acxiom, the district court held that "investor confusion may be considered by the trial court as part of its application of the Lapp factors." 27 F.Supp.2d at 501. In support of its likelihood of confusion claim, plaintiff in that case presented evidence of two investors and a potential investor who testified that they were confused by the similar names. The court held this evidence counseled in favor of finding likely confusion. It reached this conclusion on the basis of its view that the 1962 amendment to the Lanham Act evinced "a clear congressional intent to `outlaw the use of trademarks which are likely to cause confusion, mistake, or deception of any kind, not merely of purchasers nor simply as to source of origin.'" Id. at 501 (quoting Koppers, 517 F.Supp. at 843-45).

Checkpoint Systems contends investor confusion is probative of the likelihood of confusion and the District Court erred in discounting this evidence. But given the parties limited briefing on this issue as well as the scant evidence of investor confusion here, we will reserve decision on this issue. We make a few observations nonetheless. Arguably, the 1962 amendments to the Lanham Act extended actionable confusion beyond purchasers to other instances affecting a party's business or goodwill. Investor confusion may well threaten a party's business or goodwill if it would likely deter or inhibit a company's ability to attract investors and raise capital.

Here it seems unlikely that investors would be confused by the parties' similar marks or somewhat similar stock symbols ("CKP" and "CHKP" for Checkpoint Systems and Check Point Software, respectively). Many investors are sophisticated parties who exercise a high degree of care in making investment decisions. Also, the parties' shares are traded on different exchanges (Checkpoint Systems's stock is traded on the New York Stock Exchange and Check Point Software's stock is traded on the NASDAQ). Given the paucity of evidence of investor confusion here, it is unnecessary to resolve the issue here, and we believe it is prudent to reserve judgment at this time. Cf. Hoover Co. v. Citicorp Venture Capital Ltd., 674 F.Supp. 460, 461 (S.D.N.Y. 1987) (finding no Lanham Act violation where there was "only scant evidence of investor confusion in connection with a single advertisement for a securities offering.").

J. Totality of Lapp Factors

The District did not err in its factual findings nor in its application of the individual Lapp factors to those findings. Furthermore, the District Court's conclusion that the combination of the Lapp factors do not establish substantial likelihood of confusion in the marketplace is consistent with its findings. Even though mark similarity — a factor favorable to Checkpoint Systems — is ordinarily the most important element in the multi-factor Lapp analysis, it is not necessarily determinative, especially where, as here, the markets for competing marks are highly distinct. Because Checkpoint Systems's and Check Point Software's products are not related, and are marketed to different consumers through different channels, the evidence demonstrates it is unlikely the relevant consumers, who exercise a high degree of care in purchasing and selecting among these products, were likely to be confused by the marks. Additionally, there is no evidence that Check Point Software intended to use its mark to gain a competitive advantage. Also, the evidence supports the District Court's finding that the markets are unlikely to converge in the future. The minimal evidence of actual initial interest confusion is insufficient to override this conclusion. Because the District Court did not clearly err in its application of the Lapp factors — individually and in combination — we will affirm its holding there was no likelihood of (direct) marketplace confusion. See A H Sportswear, 237 F.3d at 237 ("The question of likelihood of confusion is ultimately one of fact, and [the appellate court] cannot roll up [its] sleeves and engage in the balancing [itself].").

IV. A.

On appeal, Checkpoint Systems challenges the District Court's analysis of its "reverse confusion" claim, contending the mark of Check Point Software (the junior user) has become so strong in the relevant market that the consuming public might assume that Check Point Software is actually the source of the products of Checkpoint Systems (the senior user).

In Fisons, 30 F.3d at 475, we expressly recognized a cause of action for reverse confusion, stating,

Ordinarily, one expects that the new or junior user of the mark will use to its advantage the reputation and good will of the senior user by adopting a similar or identical mark. Reverse confusion occurs when a larger, more powerful company uses the trademark of a smaller, less powerful senior owner and thereby causes likely confusion as to the source of the senior user's goods and services.

30 F.3d at 474 (citing Sands, Taylor Wood Co., 978 F.2d at 947).

Courts have recognized that "failure to recognize reverse confusion would essentially immunize from unfair competition liability companies that have well established trade names and the financial ability to advertise a senior mark taken from smaller, less powerful competitors." Harlem Wizards, 952 F.Supp. at 1091-92 (citing Big O Tire Dealers, Inc., v. Goodyear Tire Rubber Co., 561 F.2d 1365, 1372 (10th Cir. 1977)); see also Dreamwerks, 142 F.3d at 1130. "Were reverse confusion not a sufficient basis to obtain Lanham Act protection, a larger company could with impunity infringe the senior mark of a smaller one." Fisons, 30 F.3d at 475.

In a direct confusion case, the consuming public may assume that the established, senior user is the source of the junior user's goods. In a reverse confusion case, the consuming public may assume the converse; that is, that the junior, but more powerful, mark user is the source of the senior user's products. The harm to the senior user is not just that its goodwill is appropriated, but that the value of its mark is diminished. The fundamental issue in both cases, however, is essentially the same: whether the consuming public is likely to be confused about the source of products of the respective mark users.

In A H Sportswear, we held that in a typical reverse confusion case, courts should employ the same Lapp factors in assessing likelihood of confusion. But in a reverse confusion case, some of the Lapp factors must be analyzed differently. The strength of the parties' marks, the intent in adopting the marks, and the evidence of actual confusion are analyzed differently from the method employed in a typical direct confusion case. A H Sportswear, 237 F.3d at 236. On appeal, Checkpoint Systems argues the District Court did not follow the analysis of actual confusion and strength of mark outlined in A H Sportswear.

B.

In a typical direct confusion claim, the stronger the senior user's mark, the greater the likelihood of confusion caused by a junior user's use of a similar mark. But in a reverse confusion situation, the senior user's claim may be strengthened by a showing that the junior user's mark is commercially relatively strong. The greater relative strength of the junior mark allows the junior user to "overwhelm" the marketplace, diminishing the value of the senior user's mark. While analysis of the strength of the senior user's mark is relevant, the more important inquiry focuses on the junior user's mark. "[T]he lack of commercial strength of the smaller senior user's mark is to be given less weight in the analysis because it is the strength of the larger, junior user's mark which results in reverse confusion.'" A H Sportswear, 237 F.3d at 231 (quoting Commerce Nat'l Ins. Servs., Inc., 214 F.3d at 444). Courts should look at the commercial strength of the junior user's mark to determine whether its advertising and marketing has resulted in a "saturation in the public awareness of the junior user's mark." Id.

Checkpoint Systems contends the District Court improperly evaluated the strength of the marks. Because the court found its mark weak in the electronic access control products (the product most similar to Check Point Software's firewall technology), and by inference that Check Point Software's mark was relatively strong in that market, plaintiffs claim the court should have found actionable reverse confusion. In other words, as a stronger junior user in the network information security market, Check Point Software's mark allegedly overwhelmed Checkpoint Systems's senior, but weaker mark in the relevant market.

Unlike its other findings, the District Court did not scrutinize the reverse confusion claim, perhaps because it was not extensively argued. It did not expressly analyze the relative strengths of the marks for purposes of this claim, nor did it directly examine the strength of Check Point Software's mark. As noted, it is the strength of the junior user's mark that is crucial in establishing a reverse confusion claim, because it is the strength of its mark that might lead consumers to mistakenly identify "Checkpoint" — labeled products with Check Point Software. A H Sportswear, 237 F.3d at 231. In its limited discussion of the reverse confusion claim, the court stated,

[R]everse confusion threatens harm when the junior user "overwhelms" the marketplace such that the senior user begins to lose its identity and profile. This concern is of greatest force with competing or related products. In the present case, involving non-competing, unrelated products, such "overwhelming" has not even remotely occurred.

Checkpoint Sys., 104 F.Supp.2d at 461.

Had the District Court the benefit of our recent opinion in A H Sportswear, it doubtless would have been guided by our analysis. But its failure to do so does not require reversal. In light of its Lapp-factors findings, a focused analysis under A H Sportswear could not have changed the court's conclusion of no reverse confusion.

The District Court properly noted that reverse confusion is most likely in cases involving competing or related products. When products are related, consumers may be likely to assume they are produced by the same company. But when not related, the likelihood of confusion diminishes dramatically. Given the strong distinctiveness of the respective markets as found by the District Court, any confusion would appear unlikely. For this reason, a focused analysis of the strength of Check Point Software's mark would not affect the court's conclusion.

To establish a claim of reverse confusion, there must be evidence that consumers of Checkpoint Systems's products were likely to assume that Check Point Software was the source of those products. See Dreamwerks, 142 F.3d at 1130 ("The question . . . is whether consumers doing business with the senior user might mistakenly believe that they are dealing with the junior user."). But as noted, the court found the decision makers who purchased Checkpoint Systems's products were not likely to be the target audience of Check Point Software's advertising or marketing. See Checkpoint Sys., 104 F.Supp.2d at 466 ("That there is no overlap in the places these parties market their products tends to diminish the chance that someone who is a specialist in network security will even come across an advertisement for plaintiff, and vice versa."). And even if purchasers received such marketing materials, the high level of care with which they ordinarily made their purchasing decisions substantially lessened the likelihood of confusion. Put another way, the strength of Check Point Software's mark within the corporate information security market would not likely affect the consumers within Checkpoint Systems's market, given the level of distinctiveness of the relevant markets found by the District Court.

In light of these findings, the strength of Check Point Software's mark could have a determinative effect on the outcome of the reverse confusion analysis, if at all, only if extremely powerful. Only then could it "overwhelm" the marketplace diminishing the value of Checkpoint Systems's mark. Although the District Court did not analyze this factor, it was correct in finding that, "[i]n the present case, involving noncompeting, unrelated products, such `overwhelming' has not even remotely occurred." Checkpoint Sys., 104 F.Supp.2d at 461.

As we held in Fisons, the court's application of the Lapp factors is not mechanical; rather it must weigh each factor against the others. 30 F.3d at 473. When the reviewing court determines that almost all the relevant factors favor one party, it may discount one factor. See, e.g., Jet, Inc. v. Sewage Aeration Sys., 165 F.3d 419, 423 (6th Cir. 1999) (finding no reversible error when the district court erred in its application of the "converging markets" factor but correctly applied other factors); Freedom Sav. and Loan Ass'n v. Way, 757 F.2d 1176, 1186 (11th Cir. 1985) (finding no reversible error where District Court erred in its application of the "similarity of products or services" factor but correctly applied other factors). We see no legal error here.

C.

Checkpoint Systems also contends the District Court erred in its evaluation of the evidence of actual confusion in its reverse confusion claim. In A H Sportswear, we stated that in most trademark cases, parties allege both direct and reverse confusion since "the manifestation of consumer confusion as `direct' or `reverse' may merely be a function of the context in which the consumer first encountered the mark. Isolated instances of `direct' confusion may occur in a reverse confusion case, and vice-versa." 237 F.3d at 233. We declined "to create a strict bar to the use of `direct' confusion evidence in a `reverse' confusion case, or vice versa." Id. Under this view, evidence that the public thought Check Point Software was the origin of Checkpoint Systems's products is just as probative of likely confusion as evidence that the public thought Checkpoint Systems was the origin of Check Point Software's products. Here, the District Court properly examined the evidence of actual confusion and found it to be de minimis. We see no error.

Finally, Checkpoint Systems contends the District Court erred in requiring evidence of actual confusion of purchasers at the point of sale to prove its reverse confusion claim. But as we have discussed, plaintiffs have mischaracterized the court's analysis. The court factored Checkpoint Systems's evidence of initial interest and investor confusion into its confusion analysis but found the evidence insubstantial. We see no clear error.

V.

For the foregoing reasons, we will affirm the judgment of the District Court.