The Keys to Keyword Advertising

Until recently, courts have had relatively little to say about the practice of keyword advertising ' ie, triggering Internet advertisements to appear when users search for a keyword identical to a competitor's trademark. Practitioners could look only to a single decision denying Playboy Enterprises, Inc.'s (“PEI”) motion for a preliminary injunction against Netscape Communications Corp. (“Netscape”) and Excite, Inc. (“Excite”). Now, four courts have recently issued decisions reaching starkly different results on keyword advertising practices, including a Ninth Circuit decision reversing summary judgment that had been entered against PEI in its litigation with Netscape and Excite. While much remains to be resolved, certain factors have been particularly influential.

Practices at Issue

In PEI's litigation, PEI sued Netscape and Excite for selling to online adult-oriented Web sites the right to have banner advertisements displayed near search results when users of those search engines searched for the keywords “playboy” and “playmate,” which are identical in spelling to PEI's trademarks. According to the appellate decision, Netscape and Excite did not merely give such adult advertisers the option of linking their advertisements to these keywords, but required them to do so. These advertisements were found to be graphic in nature and either confusingly labeled or not labeled at all.

The other cases all involve a program distributed by WhenU.com, Inc. (“WhenU.com”) called SaveNow. SaveNow was distributed to computer users in conjunction with free screen savers and other software applications. SaveNow caused pop-up advertisements to appear over a Web site being viewed. SaveNow targeted advertisements toward particular categories that the program determined to be of interest to the computer user. The program would determine that the user was interested in a particular category, based on algorithms that took into account, among other things, particular keyword search terms and Internet addresses. Thus, if a user searched for the Web site of Wells Fargo or typed in the domain name www.wellsfargo.com, upon arriving at Wells Fargo's Web site, the user would be presented with a pop-up advertisement of one of WhenU.com's clients from the category of “finance.mortgage” in WhenU.com's directory. Thus, the advertisement appearing over the Web site of the trademark owner was likely to be competitive with the trademark owner. One decision noted that WhenU.com did not actually sell the right to have advertisements appear in conjunction with searches for particular trademarked terms, but only sold the right to have one's advertisements triggered by categories of interest, which, in turn, are triggered by the algorithms.

In the WhenU.com cases, one decision focused on an earlier version of the SaveNow program, whereby advertisements were branded with a green “$” mark and the text “SaveNow!” at the top of the pop-up window, as well as with a link at the bottom right of the window with the text “A WhenU offer – click? for info.” Other decisions emphasized a disclaimer that later versions of the program placed in the advertisements, which included the following text with a link: “This is a WhenU offer and is not sponsored or displayed by the websites you are visiting. …” One decision highlighted a license agreement whereby users must assent to a service described as one that would generate contextually relevant advertisements and coupons through pop-up advertising and other formats.

Use in Commerce and Fair Use

One point of conflict in these decisions is the question of whether the defendants' practices constituted a “use” of the plaintiffs' trademarks in commerce. In the first decision in PEI's litigation, a federal district court held that Netscape and Excite did not use PEI's trademarks in displaying banner advertisements of competing adult entertainment products and services in response to searches for the terms “playboy” and “playmate.” Although the court did not elaborate on the reason for this holding, other parts of the court's decision emphasized that these terms are generic words that are used by others and that PEI did not have the right to all commercial utility associated with these terms. Alternatively, the court deemed Netscape's and Excite's use of PEI's marks to be a fair use on the grounds that these words were used as words in the English language rather than as trademarks. Playboy Enters., Inc. v. Netscape Communications Corp., 55 F. Supp. 2d 1070, 1073, 1079, 1080-82, 1086-87 (C.D. Cal. 1999), aff'd, 202 F.3d 278 (9th Cir. 1999).

More recently, however, the Ninth Circuit reversed a subsequent summary judgment in favor of Netscape and Excite. The court found that the use of the marks was not functional because nothing about the marks had any impact on the functionality of the products with which they were used. The court found that the fair use defense for comparative advertising was not available where a likelihood of confusion is established. Finally, the court found that the use was not defensible as a “nominative” fair use, because the defendants did not limit themselves to using no more of the marks than necessary to identify products that could not otherwise be identified without such marks. Playboy Enters., Inc. v. Netscape Communications Corp., 2004 U.S. App. LEXIS 442, *22-*27 (9th Cir. Jan. 14, 2004).

In the WhenU.com cases, two courts each concluded that there was no use in commerce. These courts agreed that merely positioning advertisements near the plaintiffs' trademarks on the plaintiffs' Web sites did not constitute use in commerce. These courts distinguished cases involving the “framing” of a Web site inside branding supplied from a different source on the Internet than the source of the Web site being framed, where the boundaries between framing and framed sites are blurred. By contrast, these courts found that a clearly branded advertisement in a window over part of a Web site is distinguishable from the site over which the ad appears. These courts also found that WhenU.com's practices constituted legitimate comparative advertising. They further held that WhenU.com's practices were distinguishable from practices such as using a trademark in a domain name, which may hinder access to the trademark owner's Web site. Finally, these courts concluded that using the plaintiffs' domain names containing the plaintiffs' trademarks in WhenU.com's directory to identify relevant categories did not use the trademarks in commerce, because the marks were not used to identify the source of WhenU.com's goods or services. See Wells Fargo & Co. v. WhenU.com, Inc., 2003 U.S. Dist. LEXIS 20756, *65-*84 (E.D. Mich. Nov. 19, 2003); U-Haul Int'l, Inc. v. WhenU.com, Inc., 279 F. Supp. 2d 723, 727-29 (E.D. Va. 2003). By contrast, a third court disagreed and held that WhenU.com was using a plaintiff's trademark by causing ads to appear when users attempt to access the plaintiff's Web site and by including a domain name containing that trademark in the directory of terms triggering such ads, which allowed the advertiser to profit from the good will and reputation that led users to access the plaintiff's Web site. See 1-800 Contacts, Inc. v. WhenU.com, 2003 U.S. Dist. LEXIS 22932, *53-*58 (S.D.N.Y. Dec. 22, 2003).

Likelihood of Confusion

In PEI's litigation with Netscape and Excite, the lower court found no likelihood of confusion and, in doing so, emphasized that there was no evidence of actual confusion and no survey evidence, despite the time and PEI's ability to obtain survey evidence. In addition, the court found most of the likelihood of confusion factors inapplicable, because Netscape and Excite were not using PEI's trademarks to identify goods and were not competing with PEI. The court found no evidence of intent to confuse consumers. See Playboy Enters., 55 F. Supp. 2d at 1079, 1083-84, 1085-86.

By contrast, in finding sufficient evidence for PEI to go to trial on likelihood of confusion, the Ninth Circuit found that the defendants were using terms identical to PEI's strong marks to provide goods of advertisers that are directly competitive with PEI through the overlapping Internet marketing channel. The court also emphasized that consumers were likely to exercise little care because of price and to be easily diverted by the graphic nature of the ads.

The court gave particular emphasis to a consumer survey finding that Internet users searching for the terms at issue would think that PEI or an affiliate sponsored the banner ads appearing on Netscape's and Excite's search results pages. Other factors may have had a more subtle but persuasive effect on the court. For example, although the court found that the intent factor favored PEI slightly, the court emphasized evidence that Netscape and Excite required adult advertisers to link to the “playboy” and “playmate” search terms. The court also emphasized that many of the ads are confusingly labeled or not labeled at all. See Playboy Enters., 2004 U.S. App. LEXIS 442, at *2-*3, *13-*22.

In the WhenU.com cases, courts reached opposite conclusions on the likelihood of confusion. The Wells Fargo court concluded that the plaintiff had failed to establish a likelihood of confusion based largely on the fact that initial interest confusion was not accepted in that jurisdiction. Otherwise, that court limited its analysis to the plaintiff's survey evidence, which the court found to be flawed. Some of the many flaws found by the court were that the surveys were taken from other cases and not tailored to the relevant facts or population, that people were asked merely to imagine pop-advertisements rather than view them, and that the questions did not accommodate the possibility that users might perceive only some but not all pop-up advertisements to originate from the same source as the Web site over which they appear. The court's factual findings also emphasized that there was no evidence of actual confusion and noted that confusion was unlikely given that users of the SaveNow program are accustomed to the ads, the ads bear a prominent disclaimer, and they appear in a distinct window. Wells Fargo, 2003 U.S. Dist. LEXIS 20756, at *40-41, *88-*99.

By contrast, the 1-800 Contacts court found initial interest confusion to be likely and actionable. Like the Ninth Circuit, this court found that the strength of the mark, use of a keyword identical to the mark, and the competitive nature of the advertiser's products all weighed in favor of likely confusion. With respect to actual confusion, this court relied on the very survey rejected in the Wells Fargo decision ' a survey that had been designed for the 1-800 Contacts litigation. Although the court noted that survey's flaws and found that the factor favored neither party, the court noted that the survey at least suggested the likelihood of initial interest confusion. The court also stressed bad faith, based on inclusion of the plaintiff's trademark in WhenU.com's directory to take advantage of the plaintiff's good will. As to sophistication of the customers, the court found that customers expected to complete their Internet transactions quickly and with little effort. Finally, unlike the Wells Fargo decision, which emphasized the disclaimer in the ads and license provisions, this court based its ruling on an earlier version of the program that lacked the disclaimer, and held that disclaimers buried in a legal agreement would not dispel confusion. See 1-800 Contacts, 2003 U.S. Dist. LEXIS 22932, at *71-*103.

Factors that Made a Difference

The issues in these cases turn largely on the commercial impressions created by the particular keyword advertising practices. The following factors appear to have been particularly influential: 1) whether the jurisdiction in question recognizes initial interest confusion; 2) whether there is evidence of actual confusion or survey evidence to help the court understand consumer perceptions in this new medium; 3) whether Internet consumers are likely to be easily diverted and frustrated; 4) whether the practices are meant to appropriate the trademark owner's good will or draw consumer attention to a legitimate comparison; and 5) the degree to which branding and disclaimers make it easy for consumers to distinguish ads from the Web sites or search results that surround them. Given the wide array of keyword advertising practices, litigation is likely to continue in this area.

Until recently, courts have had relatively little to say about the practice of keyword advertising ' ie, triggering Internet advertisements to appear when users search for a keyword identical to a competitor's trademark. Practitioners could look only to a single decision denying Playboy Enterprises, Inc.'s (“PEI”) motion for a preliminary injunction against Netscape Communications Corp. (“Netscape”) and Excite, Inc. (“Excite”). Now, four courts have recently issued decisions reaching starkly different results on keyword advertising practices, including a Ninth Circuit decision reversing summary judgment that had been entered against PEI in its litigation with Netscape and Excite. While much remains to be resolved, certain factors have been particularly influential.

Practices at Issue

In PEI's litigation, PEI sued Netscape and Excite for selling to online adult-oriented Web sites the right to have banner advertisements displayed near search results when users of those search engines searched for the keywords “playboy” and “playmate,” which are identical in spelling to PEI's trademarks. According to the appellate decision, Netscape and Excite did not merely give such adult advertisers the option of linking their advertisements to these keywords, but required them to do so. These advertisements were found to be graphic in nature and either confusingly labeled or not labeled at all.

The other cases all involve a program distributed by WhenU.com, Inc. (“WhenU.com”) called SaveNow. SaveNow was distributed to computer users in conjunction with free screen savers and other software applications. SaveNow caused pop-up advertisements to appear over a Web site being viewed. SaveNow targeted advertisements toward particular categories that the program determined to be of interest to the computer user. The program would determine that the user was interested in a particular category, based on algorithms that took into account, among other things, particular keyword search terms and Internet addresses. Thus, if a user searched for the Web site of Wells Fargo or typed in the domain name www.wellsfargo.com, upon arriving at Wells Fargo's Web site, the user would be presented with a pop-up advertisement of one of WhenU.com's clients from the category of “finance.mortgage” in WhenU.com's directory. Thus, the advertisement appearing over the Web site of the trademark owner was likely to be competitive with the trademark owner. One decision noted that WhenU.com did not actually sell the right to have advertisements appear in conjunction with searches for particular trademarked terms, but only sold the right to have one's advertisements triggered by categories of interest, which, in turn, are triggered by the algorithms.

In the WhenU.com cases, one decision focused on an earlier version of the SaveNow program, whereby advertisements were branded with a green “$” mark and the text “SaveNow!” at the top of the pop-up window, as well as with a link at the bottom right of the window with the text “A WhenU offer – click? for info.” Other decisions emphasized a disclaimer that later versions of the program placed in the advertisements, which included the following text with a link: “This is a WhenU offer and is not sponsored or displayed by the websites you are visiting. …” One decision highlighted a license agreement whereby users must assent to a service described as one that would generate contextually relevant advertisements and coupons through pop-up advertising and other formats.

Use in Commerce and Fair Use

One point of conflict in these decisions is the question of whether the defendants' practices constituted a “use” of the plaintiffs' trademarks in commerce. In the first decision in PEI's litigation, a federal district court held that Netscape and Excite did not use PEI's trademarks in displaying banner advertisements of competing adult entertainment products and services in response to searches for the terms “playboy” and “playmate.” Although the court did not elaborate on the reason for this holding, other parts of the court's decision emphasized that these terms are generic words that are used by others and that PEI did not have the right to all commercial utility associated with these terms. Alternatively, the court deemed Netscape's and Excite's use of PEI's marks to be a fair use on the grounds that these words were used as words in the English language rather than as trademarks. P layboy Enters., Inc. v. Netscape Communications Corp., 55 F. Supp. 2d 1070, 1073, 1079, 1080-82, 1086-87 (C.D. Cal. 1999), aff'd , 202 F.3d 278 (9th Cir. 1999).

More recently, however, the Ninth Circuit reversed a subsequent summary judgment in favor of Netscape and Excite. The court found that the use of the marks was not functional because nothing about the marks had any impact on the functionality of the products with which they were used. The court found that the fair use defense for comparative advertising was not available where a likelihood of confusion is established. Finally, the court found that the use was not defensible as a “nominative” fair use, because the defendants did not limit themselves to using no more of the marks than necessary to identify products that could not otherwise be identified without such marks. Playboy Enters., Inc. v. Netscape Communications Corp., 2004 U.S. App. LEXIS 442, *22-*27 (9th Cir. Jan. 14, 2004).

In the WhenU.com cases, two courts each concluded that there was no use in commerce. These courts agreed that merely positioning advertisements near the plaintiffs' trademarks on the plaintiffs' Web sites did not constitute use in commerce. These courts distinguished cases involving the “framing” of a Web site inside branding supplied from a different source on the Internet than the source of the Web site being framed, where the boundaries between framing and framed sites are blurred. By contrast, these courts found that a clearly branded advertisement in a window over part of a Web site is distinguishable from the site over which the ad appears. These courts also found that WhenU.com's practices constituted legitimate comparative advertising. They further held that WhenU.com's practices were distinguishable from practices such as using a trademark in a domain name, which may hinder access to the trademark owner's Web site. Finally, these courts concluded that using the plaintiffs' domain names containing the plaintiffs' trademarks in WhenU.com's directory to identify relevant categories did not use the trademarks in commerce, because the marks were not used to identify the source of WhenU.com's goods or services. See Wells Fargo & Co. v. WhenU.com, Inc., 2003 U.S. Dist. LEXIS 20756, *65-*84 (E.D. Mich. Nov. 19, 2003); U-Haul Int'l, Inc. v. WhenU.com, Inc. , 279 F. Supp. 2d 723, 727-29 (E.D. Va. 2003). By contrast, a third court disagreed and held that WhenU.com was using a plaintiff's trademark by causing ads to appear when users attempt to access the plaintiff's Web site and by including a domain name containing that trademark in the directory of terms triggering such ads, which allowed the advertiser to profit from the good will and reputation that led users to access the plaintiff's Web site. See 1-800 Contacts, Inc. v. WhenU.com, 2003 U.S. Dist. LEXIS 22932, *53-*58 (S.D.N.Y. Dec. 22, 2003).

Likelihood of Confusion

In PEI's litigation with Netscape and Excite, the lower court found no likelihood of confusion and, in doing so, emphasized that there was no evidence of actual confusion and no survey evidence, despite the time and PEI's ability to obtain survey evidence. In addition, the court found most of the likelihood of confusion factors inapplicable, because Netscape and Excite were not using PEI's trademarks to identify goods and were not competing with PEI. The court found no evidence of intent to confuse consumers. See Playboy Enters., 55 F. Supp. 2d at 1079, 1083-84, 1085-86.

By contrast, in finding sufficient evidence for PEI to go to trial on likelihood of confusion, the Ninth Circuit found that the defendants were using terms identical to PEI's strong marks to provide goods of advertisers that are directly competitive with PEI through the overlapping Internet marketing channel. The court also emphasized that consumers were likely to exercise little care because of price and to be easily diverted by the graphic nature of the ads.

The court gave particular emphasis to a consumer survey finding that Internet users searching for the terms at issue would think that PEI or an affiliate sponsored the banner ads appearing on Netscape's and Excite's search results pages. Other factors may have had a more subtle but persuasive effect on the court. For example, although the court found that the intent factor favored PEI slightly, the court emphasized evidence that Netscape and Excite required adult advertisers to link to the “playboy” and “playmate” search terms. The court also emphasized that many of the ads are confusingly labeled or not labeled at all. See Playboy Enters., 2004 U.S. App. LEXIS 442, at *2-*3, *13-*22.

In the WhenU.com cases, courts reached opposite conclusions on the likelihood of confusion. The Wells Fargo court concluded that the plaintiff had failed to establish a likelihood of confusion based largely on the fact that initial interest confusion was not accepted in that jurisdiction. Otherwise, that court limited its analysis to the plaintiff's survey evidence, which the court found to be flawed. Some of the many flaws found by the court were that the surveys were taken from other cases and not tailored to the relevant facts or population, that people were asked merely to imagine pop-advertisements rather than view them, and that the questions did not accommodate the possibility that users might perceive only some but not all pop-up advertisements to originate from the same source as the Web site over which they appear. The court's factual findings also emphasized that there was no evidence of actual confusion and noted that confusion was unlikely given that users of the SaveNow program are accustomed to the ads, the ads bear a prominent disclaimer, and they appear in a distinct window. Wells Fargo, 2003 U.S. Dist. LEXIS 20756, at *40-41, *88-*99.

By contrast, the 1-800 Contacts court found initial interest confusion to be likely and actionable. Like the Ninth Circuit, this court found that the strength of the mark, use of a keyword identical to the mark, and the competitive nature of the advertiser's products all weighed in favor of likely confusion. With respect to actual confusion, this court relied on the very survey rejected in the Wells Fargo decision ' a survey that had been designed for the 1-800 Contacts litigation. Although the court noted that survey's flaws and found that the factor favored neither party, the court noted that the survey at least suggested the likelihood of initial interest confusion. The court also stressed bad faith, based on inclusion of the plaintiff's trademark in WhenU.com's directory to take advantage of the plaintiff's good will. As to sophistication of the customers, the court found that customers expected to complete their Internet transactions quickly and with little effort. Finally, unlike the Wells Fargo decision, which emphasized the disclaimer in the ads and license provisions, this court based its ruling on an earlier version of the program that lacked the disclaimer, and held that disclaimers buried in a legal agreement would not dispel confusion. See 1-800 Contacts, 2003 U.S. Dist. LEXIS 22932, at *71-*103.

Factors that Made a Difference

The issues in these cases turn largely on the commercial impressions created by the particular keyword advertising practices. The following factors appear to have been particularly influential: 1) whether the jurisdiction in question recognizes initial interest confusion; 2) whether there is evidence of actual confusion or survey evidence to help the court understand consumer perceptions in this new medium; 3) whether Internet consumers are likely to be easily diverted and frustrated; 4) whether the practices are meant to appropriate the trademark owner's good will or draw consumer attention to a legitimate comparison; and 5) the degree to which branding and disclaimers make it easy for consumers to distinguish ads from the Web sites or search results that surround them. Given the wide array of keyword advertising practices, litigation is likely to continue in this area.