Opinion
23-16060
10-22-2024
Andrew Gaggin (argued), Lerner &Rowe PC, Tucson, Arizona, for Plaintiff-Appellant. Maria C. Speth (argued) and Aaron K. Haar, Jaburg Wilk PC, Phoenix, Arizona, for Defendant-Appellee.
Argued and Submitted May 14, 2024 Phoenix, Arizona.
Appeal from the United States District Court for the District of Arizona D.C. No. 2:21-cv-01540-DGC, David G. Campbell, District Judge, Presiding.
COUNSEL
Andrew Gaggin (argued), Lerner &Rowe PC, Tucson, Arizona, for Plaintiff-Appellant. Maria C. Speth (argued) and Aaron K. Haar, Jaburg Wilk PC, Phoenix, Arizona, for Defendant-Appellee.
Before: Roopali H. Desai and Ana de Alba, Circuit Judges, and Edward M. Chen, [*] District Judge.
SUMMARY [**]
Lanham Act
The panel affirmed the district court's grant of summary judgment in favor of defendants in a trademark infringement action under the Lanham Act.
Plaintiff Lerner &Rowe, PC, a personal injury law firm based in Arizona, had three registered trademarks, including the name "Lerner &Rowe." In a strategy known as "conquesting," defendant Brown, Engstrand & Shely, LLC, doing business as The Accident Law Group, or ALG, purchased the term "Lerner &Rowe" as a Google Ads keyword.
The panel affirmed the district court's grant of summary judgment on Lerner &Rowe's trademark infringement claim on the ground that Lerner &Rowe failed to establish that ALG's use of the mark was likely to cause consumer confusion. The panel concluded that the strength of the mark weighed in favor of Lerner &Rowe. But the de minimis evidence of actual confusion weighed in favor of ALG, as did the reasonably prudent consumer's degree of care and the labeling and appearance of ALG's advertisements. And other factors did nothing to change the panel's conclusion that Lerner &Rowe failed to establish a genuine dispute of material fact regarding the likelihood of confusion element of a claim for trademark infringement.
Concurring in the majority opinion in full, Judge Desai wrote separately to urge the court to reconsider en banc the holding of Network Automation, Inc. v. Advance Systems Concepts, Inc., 638 F.3d 1137 (9th Cir. 2011), that keyword bidding and purchasing constitutes a "use in commerce," which is required to show a likelihood of confusion under the Lanham Act.
OPINION
DE ALBA, Circuit Judge:
"What's in a name?" WILLIAM SHAKESPEARE, ROMEO AND JULIET act 2, sc. 2, l. 46. According to Juliet Capulet, not much. Romeo Montague's last name, though charged with meaning, does not confuse her about who he is. In this keyword advertising trademark dispute, the district court saw most consumers as discerning Juliets. Appellant, however, likens them to the larger Capulet clan, a group more prone to confusion. As explained below, we disagree and affirm the district court's grant of summary judgment.
I. Factual and Procedural Background
Appellant Lerner & Rowe, PC ("Lerner &Rowe"), and Appellee Brown, Engstrand &Shely, LLC-which does business as The Accident Law Group ("ALG")-are both personal injury law firms based in Arizona. Founded in 2005, Lerner &Rowe is the larger of the two firms with nineteen offices throughout the state. It has three registered trademarks: on June 14, 2011, it registered the phrase "Lerner &Rowe Gives Back;" on March 3, 2015, it registered the name "Glen Lerner;" and, on May 19, 2020, it registered the name "Lerner &Rowe." Lerner &Rowe has spent over $100 million promoting its brand and trademarks in Arizona.
Since its founding in 2015 until 2021, ALG purchased the term "Lerner &Rowe" as a Google Ads keyword, which prompted ALG's advertisements to appear near the top of Google's search results list whenever someone searched for "Lerner & Rowe." This strategy, known as "conquesting," is a common internet marketing tool by which companies promote their services to potential customers who might be searching for a competitor. In fact, Lerner &Rowe has engaged in conquesting in other contexts. Importantly, while the format and copy of ALG's advertisements varied from search to search, they never included or referenced the term "Lerner &Rowe."
On September 8, 2021, Lerner &Rowe filed a complaint alleging claims for (1) trademark infringement, unfair competition, false designation of origin, and false description under the Lanham Act; (2) state trademark infringement and unfair competition; and (3) unjust enrichment. In a May 18, 2023, order, the district court granted summary judgment in favor of ALG on the trademark infringement and unjust enrichment claims but denied summary judgment on the unfair competition claims. ALG moved for reconsideration, and the district court subsequently entered summary judgment as to all claims. Lerner &Rowe timely appealed that ruling. We have jurisdiction pursuant to 28 U.S.C. § 1291.
II. Legal Standard
We review grants of summary judgment de novo. Multi Time Mach., Inc. v. Amazon.com, Inc., 804 F.3d 930, 935 (9th Cir. 2015). "[O]n a defendant's motion for summary judgment, not only does the movant carry the burden of establishing that no genuine dispute of material fact exists, but the court also views the evidence in the light most favorable to the non-moving party." JL Beverage Co., LLC v. Jim Beam Brands Co., 828 F.3d 1098, 1105 (9th Cir. 2016). A genuine dispute of material fact exists "if the evidence is such that a reasonable jury could return a verdict for the nonmoving party." Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248 (1986). "If the evidence is merely colorable or is not significantly probative, summary judgment may be granted." Id. at 249-50 (citations omitted). When, as here, the moving party does not have the burden of proof on an issue at trial, it "can prevail merely by pointing out that there is an absence of evidence to support the nonmoving party's case." Soremekun v. Thrifty Payless, Inc., 509 F.3d 978, 984 (9th Cir. 2007). "If the moving party meets its initial burden, the non-moving party must set forth, by affidavit or as otherwise provided in Rule 56, 'specific facts showing that there is a genuine issue for trial.'" Id. (quoting Anderson, 477 U.S. at 250). Due to the factintensive nature of trademark infringement claims, we grant motions for summary judgment infrequently. See JL Beverage, 828 F.3d at 1105. Nevertheless, when no genuine issue of material fact exists, we have not hesitated to affirm a grant of summary judgment. See Surfvivor Media, Inc. v. Survivor Prods., 406 F.3d 625, 634 (9th Cir. 2005); M2 Software, Inc. v. Madacy Ent., 421 F.3d 1073, 1085 (9th Cir. 2005).
III. Discussion
"To prevail on a claim of trademark infringement under the Lanham Act, 15 U.S.C. § 1114, a party 'must prove: (1) that it has a protectible ownership interest in the mark; and (2) that the defendant's use of the mark is likely to cause consumer confusion.'" Network Automation, Inc. v. Advanced Sys. Concepts, Inc., 638 F.3d 1137, 1144 (9th Cir. 2011) (quoting Dep't of Parks &Recreation for the State of Cal. v. Bazaar Del Mundo Inc., 448 F.3d 1118, 1124 (9th Cir. 2006)). Because the parties do not dispute that Lerner &Rowe has a protectible interest in its mark, this case concerns only the likelihood of confusion element.
When assessing the likelihood of confusion in the keyword advertising context, we primarily consider the following non-exhaustive list of factors:
(1) the strength of the mark; (2) the evidence of actual confusion; (3) the type of goods and degree of care likely to be exercised by the purchaser; and (4) the labeling and appearance of the advertisements and the surrounding context on the screen displaying the results page.Id. at 1154. Other, less relevant factors include the "proximity of the goods, similarity of the marks, marketing channels used, defendant's intent in selecting the mark, and likelihood of expansion of the product lines." Id. at 1145 (quoting AMF Inc. v. Sleekcraft Boats, 599 F.2d 341, 34849 (9th Cir. 1979)) (cleaned up). These factors are "not a rote checklist," and we must be flexible when analyzing them. Id. Depending on the circumstances of a given case, certain factors may be more important than others. Id. at 1148; see also Multi Time Mach., 804 F.3d at 937, 939 (affirming grant of summary judgment based on two factors: "evaluation of the web page at issue and the relevant consumer").
This case primarily concerns "initial interest confusion," which occurs when an alleged infringer uses a competitor's mark to direct consumer attention to its product. See Playboy Enters., Inc. v. Netscape Commc'ns Corp., 354 F.3d 1020, 1025 (9th Cir. 2004). "Although dispelled before an actual sale occurs, initial interest confusion impermissibly capitalizes on the goodwill associated with a mark and is therefore actionable trademark infringement." Id. Such a claim applies, however, only to "misleading and deceptive" uses of a mark, not to "legitimate comparative and contextual advertising." Network Automation, 638 F.3d at 1148. Therefore, in the keyword advertising context, we have emphasized that, "the owner of the mark must demonstrate likely confusion, not mere diversion." Id. at 1149; see also Playboy Enters., 354 F.3d at 1035 (Berzon, J., concurring) ("There is a big difference between hijacking a customer to another website by making the customer think he or she is visiting the trademark holder's website (even if only briefly) . . . and just distracting a potential customer with another choice, when it is clear that it is a choice.").
Lerner & Rowe also advanced a theory of source confusion, which occurs when consumers purchase services from an alleged infringer due to confusion about the actual provider of those services. See Brookfield Commc'ns, Inc. v. W. Coast Ent. Corp., 174 F.3d 1036, 1062 (9th Cir. 1999) (citing Dr. Seuss Enters., L.P. v. Penguin Books USA, Inc., 109 F.3d 1394, 1405 (9th Cir. 1997)). This does not, however, affect our analysis, because both theories turn on the same likelihood of confusion test.
A. Strength of the Mark
Strong trademarks receive greater protection because "a user searching for a distinctive term is more likely to be looking for a particular product, and therefore, could be more susceptible to confusion when sponsored links appear that advertise a similar product from a different source." Network Automation, 638 F.3d at 1149. Courts measure a mark's strength both conceptually-by its "inherent distinctiveness"- and commercially-by its "actual marketplace recognition." Id. (quoting Brookfield Commc'ns, 174 F.3d at 1058). Even when a mark is not inherently distinctive, commercial strength-"extensive advertising, length of exclusive use, public recognition"- can compensate for its conceptual weakness. Am. Int'l Grp., Inc. v. Am. Int'l Bank, 926 F.2d 829, 832 (9th Cir. 1991) (quoting Accuride Int'l, Inc. v. Accuride Corp., 871 F.2d 1531, 1536 (9th Cir. 1989)).
The district court correctly found, and ALG does not dispute, that Lerner &Rowe's mark is strong. Not only is the mark federally registered, but Lerner &Rowe has spent millions of dollars advertising it, garnering the business of over 100,000 clients. This factor weighs in favor of Lerner &Rowe.
B. Evidence of Actual Confusion "[A] showing of actual confusion among significant numbers of consumers provides strong support for the likelihood of confusion." Playboy Enters., 354 F.3d at 1026. In fact, if a plaintiff can demonstrate "that an 'appreciable number' of people are confused," that fact, alone, might entitle the plaintiff to a trial on the likelihood of confusion. Thane Int'l, Inc. v. Trek Bicycle Corp., 305 F.3d 894, 902 (9th Cir. 2002), superseded on other grounds by statute, Trademark Dilution Revision Act of 2006, Pub. L. No. 109312, 120 Stat. 1730-33, as recognized in Blumenthal Distrib., Inc. v. Herman Miller, Inc., 963 F.3d 859, 870 (9th Cir. 2020) (quoting Entrepreneur Media, Inc. v. Smith, 279 F.3d 1135, 1151 (9th Cir. 2002)). Nevertheless, because actual confusion evidence is difficult to gather, "the absence of such evidence is not dispositive." Off. Airline Guides, Inc. v. Goss, 6 F.3d 1385, 1393 (9th Cir. 1993).
Here, Lerner &Rowe's proffer of actual confusion consists of 236 phone calls that ALG's intake department received during which the caller mentioned Lerner &Rowe by name when responding to a question about how the caller found ALG's phone number. Data from Google shows that, between 2017 and 2021, searches for "Lerner &Rowe" returned results featuring ALG's advertisement 109,322 times. Evidence of 236 instances of actual confusion, therefore, constitutes only 0.216% of the total number of users exposed to the challenged advertisements. Moreover, users clicked on ALG's advertisements 7,452 times, or just 6.82% of the time Google displayed them. ALG separately commissioned an expert survey concluding that ALG's advertisements confused between 0% and 3% of consumers. The district court dismissed this evidence of actual confusion as de minimis and concluded that this factor favored ALG.
The district court concluded that most of these call log entries were too ambiguous to constitute reliable evidence of actual confusion. The entries are indeed terse, and many do not convey any apparent impression of customer confusion. For example, some callers mentioned Lerner & Rowe because the firm had referred them to ALG. This is not evidence of confusion at all. Other entries-like one that states, "Google. Thought we were L&R"-more likely express confusion. Most of the entries fall somewhere between these two poles in terms of the clarity with which they convey customer confusion. Nevertheless, for the sake of brevity, we will treat all 236 call log entries as evidence of actual confusion because, as discussed below, even that total, under the particular facts of this case, represents only de minimis evidence of actual confusion.
In the district court, the parties acknowledged that, because the call logs did not include entries from 2017, it would be more accurate to compare the 236 calls to the 102,382 results featuring ALG's advertisements that occurred between 2018 and 2021. Doing so results in a purported actual confusion rate of 0.231%, which does not meaningfully change our analysis.
Lerner &Rowe does not dispute these statistics. Nor did it commission its own survey. Rather, it relies on cases like Ironhawk Technologies, Inc. v. Dropbox, Inc., 2 F.4th 1150 (9th Cir. 2021), for the proposition that even one or two instances of actual confusion should weigh in the plaintiff's favor on summary judgment. In Ironhawk, we weighed two instances of actual confusion in favor of the plaintiff, concluding that "it is evidence a reasonable jury could rely on to support a finding of actual confusion or when assessing a likelihood of confusion under the totality of the circumstances." 2 F.4th at 1166; see also Entrepreneur Media, 279 F.3d at 1151 (holding that, while a jury could disregard as de minimis a single incident of actual confusion, such evidence still weighed slightly in favor of plaintiff's infringement claim for purposes of summary judgment). In Lerner &Rowe's view, its proffer of 236 instances of actual confusion easily meets Ironhawk's standard regardless of the number of times consumers viewed ALG's advertisements.
Typically, instances of actual confusion present a numerator with no denominator, saying little or nothing about the actual proportion of the consumer population that is confused. In such cases, we see the tip of an iceberg and have no ability to speculate about how much lies below the surface. Here, however, no speculation is necessary-we can see the entire iceberg. Because we have both the numerator-the 236 calls representing actual confusion- and the denominator-the 109,322 consumers who saw the advertisements-we can discern with a high degree of precision the proportion of all consumers who were actually confused. See 3 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 23:14 (5th ed.) ("Evidence of the number of instances of actual confusion must be placed against the background of the number of opportunities for confusion before one can make an informed decision as to the weight to be given the evidence."). The resulting 0.216% confusion rate is direct evidence of the likelihood of confusion comparable to, but more complete than, survey evidence. No reasonable jury would conclude that this percentage is anything but de minimis and fails to support a finding of likelihood of confusion. See Nutri/Sys., Inc. v. Con-Stan Indus., Inc., 809 F.2d 601, 606-07 (9th Cir. 1987) (holding that after a bench trial, the trial court properly discounted instances of confusion that "at best, were thin, and at worst, were trivial"); Entrepreneur Media, 279 F.3d at 1151 (holding that "a reasonable juror could find de minimis, and thus unpersuasive, one instance of actual confusion"); see also Henri's Food Prods. Co., Inc. v. Kraft, Inc., 717 F.2d 352, 358 (7th Cir. 1983) (holding that a survey confusion rate of 7.6% weighed against infringement).
Our conclusion does not conflict with cases like Ironhawk, where we weighed individual instances of confusion without the benefit of knowing the total number of opportunities consumers had for confusion. See 2 F.4th at 1165-66. We surmised that a reasonable jury would likely find the proffered evidence of actual confusion in Ironhawk de minimis, but we could not make that determination ourselves without more data. See id. at 1166. Here, on the other hand, we know how many times consumers searched for "Lerner &Rowe" on Google and saw an ALG advertisement. We also know how many of those consumers called ALG and, in a potential expression of confusion, referenced "Lerner &Rowe." The resulting calculation is simple and telling: unlike in Ironhawk, the evidence of actual confusion here is demonstrably de minimis.
While evidence showing the actual proportion of confused consumers is important, we do not suggest that courts should automatically discount de minimis instances of actual confusion when the record contains additional evidence of consumer confusion. The Fourth Circuit's decision in Rosetta Stone Ltd. v. Google, Inc., 676 F.3d 144 (4th Cir. 2012), is instructive. There, the district court disregarded five depositions from confused consumers because there had been more than 100,000 opportunities for confusion over a period of six years. Id. at 157-58. The Fourth Circuit noted that, if the depositions had been the only evidence of actual confusion before the district court, disregarding them would not have been improper. Id. at 158. But the plaintiff had presented other evidence, including records of 262 customer complaints, in-house studies from Google about the likelihood that the defendant's advertising strategy could confuse consumers, testimony from Google's in-house trademark attorneys who were themselves unable to distinguish between the links at issue in the case, and an expert survey demonstrating a net confusion rate among consumers of 17%. Id. at 158-59. Here, by contrast, Lerner &Rowe's de minimis actual confusion evidence stands alone. In fact, ALG presented the only other evidence of confusion-an expert survey showing a customer confusion rate of 0% to 3% and evidence of a 6.82% click-thru rate- which bolsters the de minimis nature of Lerner &Rowe's actual confusion evidence. See 5 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 32:189 (5th ed.) ("When the percentage results of a confusion survey dip below 10%, they can become evidence which will indicate that confusion is not likely.").
As one circuit has recognized, a click-thru rate represents the upper limit of initial interest confusion. See 1-800 Contacts, Inc. v. Lens.com, Inc., 722 F.3d 1229, 1244 (10th Cir. 2013). But we cannot know how many, if any, consumers clicked on ALG's advertisements out of confusion rather than mere diversion.
Having determined that Lerner &Rowe's evidence of actual confusion is de minimis, we must now decide how to weigh it. In one sense, the evidence Lerner &Rowe has presented is so slight it may as well have presented none at all. Due to the difficulties in gathering evidence of actual confusion, we have noted that "its absence [is] generally unnoteworthy." Brookfield Commc'ns, 174 F.3d at 1050; see also LaQuinta Worldwide LLC v. Q.R.T.M., S.A. de C.V., 762 F.3d 867 (9th Cir. 2014). But see M2 Software, 421 F.3d at 1083 (weighing plaintiff's failure to proffer evidence of actual confusion in favor of defendant); One Indus., LLC v. Jim O'Neal Distrib., Inc., 578 F.3d 1154, 1163 (9th Cir. 2009) (same). Here, however, the nature of the actual confusion evidence paints a picture that affirmatively contradicts Lerner &Rowe's assertions that ALG's advertisements were likely to confuse an appreciable number of consumers, compelling us to conclude that this factor should weigh substantially in favor of ALG. See Surfvivor Media, 406 F.3d at 633 (weighing de minimis actual confusion evidence against plaintiff when defendant presented consumer survey showing "an absence of significant confusion"); see also Brookfield Commc'ns, 174 F.3d at 1050 (noting "a crucial difference" between a plaintiff's concession of no actual confusion and a mere failure to present such evidence); Cohn v. Petsmart, Inc., 281 F.3d 837, 842-43 (9th Cir. 2002) (per curiam) (weighing factor against plaintiff where, under the circumstances, "some evidence of actual confusion should have become available").
C. The Reasonably Prudent Consumer's Degree of Care
Sophisticated consumers and those shopping for high-value products are likely to exercise a higher degree of care while shopping and are, therefore, less likely to be confused by similar marks. See Network Automation, 638 F.3d at 1152. Additionally, when it comes to online shopping, "the default degree of consumer care is becoming more heightened as the novelty of the Internet evaporates and online commerce becomes commonplace." Id. The district court weighed this factor in favor of ALG because acquiring legal services can be expensive and important and because those accustomed to online shopping are typically savvy enough to differentiate between search engine results.
We agree that this factor weighs in ALG's favor. Since at least 2010, we have recognized that "[c]onsumers who use the internet for shopping are generally quite sophisticated about" how the internet functions. Toyota Motor Sales, U.S.A., Inc. v. Tabari, 610 F.3d 1171, 1178 (9th Cir. 2010). For example, regular internet users can readily distinguish domain names associated with the companies they are searching for from those they are not. See id. Additionally, Google's search engine is so ubiquitous that we can be confident that the reasonably prudent online shopper is familiar with its layout and function, knows that it orders results based on relevance to the search term, and understands that it produces sponsored links along with organic search results. Moreover, in this case, the relevant consumers specifically typed in "Lerner &Rowe" as a search term, suggesting that they would be even more discerning of the results they received. Therefore, because this case involves shopping on Google by using the precise trademark at issue, this factor weighs in favor of ALG.
It is unnecessary for us to address the district court's assumption that the value of personal injury legal services heightens the degree of consumer care.
D. Labeling and Appearance of Advertisements
"[C]lear labeling can eliminate the likelihood of initial interest confusion in cases involving Internet search terms." Multi Time Mach., 804 F.3d at 937; see also Network Automation, 638 F.3d at 1153 ("In the keyword advertising context the 'likelihood of confusion will ultimately turn on what the consumer saw on the screen and reasonably believed, given the context.'" (quoting Hearts on Fire Co. v. Blue Nile, Inc., 603 F.Supp.2d 274, 289 (D. Mass. 2009))). The district court, after analyzing three screenshots depicting ALG's advertisements, concluded that the advertisements would not confuse a reasonably prudent consumer searching online for personal injury legal services. We agree.
Lerner & Rowe provided 28 screenshots for the district court's review, but 25 of those images were gathered after May 2021, when ALG stopped paying Google for "Lerner & Rowe" as an advertising keyword. Accordingly, the district court looked only to the three screenshots that pre-dated May 2021; we will do the same.
To frame the following discussion, the relevant screenshots depicting ALG's advertisements are reprinted below:
First screenshot:
(Image Omitted)
Second screenshot:
(Image Omitted)
Third screenshot:
(Image Omitted)
The most significant feature of the second and third screenshots is the clearly labeled result for Lerner &Rowe's website. Though the first screenshot does not display a result for Lerner &Rowe, we think it reasonable that, based on the other two screenshots, such a result likely appeared immediately after the ALG advertisement. But even if the list of search results did not include an entry for Lerner &Rowe after the ALG advertisement, our conclusion would remain the same. Indeed, we find it difficult to believe that consumers searching for the phrase "Lerner &Rowe" would not choose to click on the link that matches their search query word for word.
Nor do we think that ALG's advertisements are so confusing as to lure reasonably prudent online shoppers into unwittingly clicking on them in search of Lerner &Rowe's website. Lerner &Rowe attempts to demonstrate confusion by distinguishing Multi Time Machine v. Amazon.com, where we held that Amazon's search results page was so clearly labeled that no reasonable consumer would find it confusing. See 804 F.3d at 937-38. That case involved Amazon searches for the MTM Special Ops watch, a product that the manufacturer did not sell on Amazon. Id. at 933. When someone searched for "mtm special ops" on Amazon, the results page listed the search query twice above a "Related Searches" field that contained alternative search queries that might help the consumer find a related product. Id. Below the "Related Searches" field, separated by a gray bar, was a list of products available on Amazon that were similar to the MTM Special Ops watch. Id. at 934. The entry for each of these products included a photograph and listed the name of the product and the manufacturer in "large, bright, bold letters." Id. at 938.
Lerner &Rowe notes that, unlike in Multi Time Machine, Google's search results do not contain a "Related Results" field and do not separate advertisements from organic results with "borders, bars, or shading." First, it is not surprising that Google styles its search results differently from Amazon; they are distinct search engines with distinct functions. Second, Multi Time Machine did not elucidate a list of features that a search engine must incorporate in order for their results to be clearly labeled. Analyzing the search results in the context of the Google results at issue here, we conclude that the bolded "Ad" designation next to each of ALG's advertisements sufficiently distinguishes ALG's advertisements from the search's organic results. Moreover, the fact that ALG's advertisements sometimes appear above organic results for Lerner &Rowe does not change this analysis. We think that reasonably prudent consumers shopping on Google would be accustomed to scrolling past advertisements at the top of a list of search results to find the organic result relevant to their query.
We acknowledge that some of ALG's advertisements are not models of clarity. As Lerner &Rowe points out, sometimes the content of an advertisement contains generic statements that could apply to any personal injury law firm-for example, "Your Personal Injury Attorney-We Don't Win-You Don't Pay." In such cases, the only feature identifying ALG as the source of the advertisement is the URL, which is in a smaller, lighter font. While these features could possibly cause confusion in isolation, our job is to analyze the advertisements within the context of the entire search results page. That page invariably contains a result for Lerner &Rowe that includes the precise search term at issue, dispelling any confusion ALG's advertisements might cause. The parties' presentation of de minimis evidence of actual confusion only bolsters our conclusion that it is only the "[u]nreasonable, imprudent and inexperienced webshoppers" who might find the search results pages confusing. Tabari, 610 F.3d at 1176.
E. Other Factors
While the factors above are the most relevant to trademark infringement claims based on keyword advertising, other factors can also be helpful. See Network Automation, 638 F.3d at 1149-54 (weighing nine factors and finding four to be the most relevant to the court's analysis). Here, however, our assessment of these other factors does nothing to change our conclusion that Lerner &Rowe has failed to establish a genuine dispute of material fact regarding the likelihood of confusion element.
1. Proximity of the Goods
When companies provide similar services, consumers are more likely to confuse them. See Network Automation, 638 F.3d at 1150. Nevertheless, "the proximity of the goods . . . become[s] less important if advertisements are clearly labeled or consumers exercise a high degree of care, because rather than being misled, the consumer would merely be confronted with choices among similar products." Id. The district court correctly noted that, even though ALG and Lerner &Rowe are direct competitors offering similar services, savvy online shoppers would be able to differentiate between the parties' links on Google. If it has any weight at all, this factor falls in favor of ALG.
2. Marketing Channels
This factor might be relevant if ALG's advertisements appeared on a lesser-known or product-specific search engine, but "[t]oday, it would be the rare commercial retailer that did not advertise online, and the shared use of a ubiquitous marketing channel does not shed much light on the likelihood of consumer confusion." Network Automation, 638 F.3d at 1151. Lerner &Rowe cites a case from the year 2000 to argue that online marketing increases the likelihood of confusion. While that may have been true over twenty years ago when internet advertising was new, our precedent acknowledges that advertising on Google is commonplace today. The district court properly accorded this factor little to no weight.
3. Similarity of Marks
"Where the two marks are entirely dissimilar, there is no likelihood of confusion." Brookfield Commc'ns, 174 F.3d at 1054. Lerner &Rowe argues that this factor favors it because ALG's use of Lerner &Rowe's mark as a keyword means that ALG uses a mark identical to Lerner &Rowe's. Network Automation rejected this exact reasoning, holding that this factor should reflect "what consumers 'encountered in the marketplace,'" not what Google's algorithm uses to churn out search results. 638 F.3d at 1151. In this case, ALG does not display Lerner &Rowe's mark in its advertisements. In fact, the URL above each advertisement displays ALG's own mark, albeit in a lower-case, condensed form. These two marks-"Lerner &Rowe" and "Accident Law Group"-are in no way similar. This factor favors ALG.
4. Intent
"When the alleged infringer knowingly adopts a mark similar to another's, reviewing courts presume that the defendant can accomplish his purpose: that is, that the public will be deceived." Network Automation, 638 F.3d at 1153 (quoting Sleekcraft, 599 F.2d at 354). Apart from an affirmative intent to confuse, an alleged infringer's failure to take remedial steps when faced with evidence of confusion can cause a likelihood of confusion. See Playboy Enters., 354 F.3d at 1028-29. We agree with the district court that, because Lerner &Rowe's evidence of intent is identical to the evidence it offered to support its likelihood of confusion argument generally, it has failed to distinguish between an intent to deceive and an intent to compete on the part of ALG. Accordingly, this factor bears little to no weight.
5. Likelihood of Expansion of Product Lines
"The likelihood of expansion of product lines factor is relatively unimportant where two companies already compete to a significant extent." Brookfield Commc'ns, 174 F.3d at 1060. Lerner &Rowe acknowledges that this factor is unimportant to the likelihood of confusion analysis because it competes directly with ALG. The district court correctly acknowledged the same.
IV. Conclusion
The district court was correct to conclude that this is one of the rare trademark infringement cases susceptible to summary judgment. The generally sophisticated nature of online shoppers, the evidence demonstrating that there is not an appreciable number of consumers who would find ALG's use of the mark confusing, and the clarity of Google's search results pages, convince us that ALG's use of the "Lerner &Rowe" mark is not likely to cause consumer confusion. The district court's judgment is affirmed.
ALG alternatively asks us to affirm the district court's grant of summary judgment on the ground that ALG never used Lerner & Rowe's trademark in commerce. Network Automation, however, explicitly held that "the use of a trademark as a search engine keyword that triggers the display of a competitor's advertisement is a 'use in commerce' under the Lanham Act." 638 F.3d at 1145-46. Because no intervening Supreme Court decision is "clearly irreconcilable" with this holding, we have no power to overrule it. Miller v. Gammie, 335 F.3d 889, 900 (9th Cir. 2003) (en banc).
DESAI, Circuit Judge, concurring:
I concur in the majority opinion in full. But I write separately to urge our court to reconsider whether keyword bidding and purchasing constitutes a "use in commerce" under the Lanham Act. Our binding precedent says it does, Network Automation, Inc. v. Advance Systems Concepts, Inc., 638 F.3d 1137, 1144-45 (9th Cir. 2011), but I am not convinced that we got it right or that our holding withstands the test of time and recent advancements in technology.
To prevail on a trademark infringement claim, a plaintiff "must prove: (1) that it has a protectible ownership interest in the mark; and (2) that the defendant's use of the mark is likely to cause consumer confusion." Dep't of Parks &Recreation v. Bazaar Del Mundo Inc., 448 F.3d 1118, 1124 (9th Cir. 2006). Subsumed in the second element of this test is the requirement that a defendant uses the mark in commerce. 15 U.S.C. § 1114(1)(a). But we have not seriously grappled with whether bidding on keywords constitutes a "use in commerce." That is partly because, ordinarily, the bulk of our focus in trademark infringement cases is devoted to whether the defendant's conduct created a likelihood of consumer confusion. With the growing reliance by businesses on keyword advertising, it is time to revisit what "use in commerce" means in this context.
Under the Lanham Act, a mark is "used in commerce" when it is "used or displayed in the sale or advertising of services." 15 U.S.C. § 1127. This definition is easily satisfied when a defendant displays a mark. But what about when a defendant does not display a mark? Is it enough that a defendant merely bid on a mark, even if the defendant never displayed the mark themselves?
This definition relates to the requirements for registering a mark, but courts routinely use it in the infringement context as well. See Rescuecom Corp. v. Google Inc., 562 F.3d 123, 139-41 (2d Cir. 2009)
We have previously suggested that a defendant can "use" a mark in commerce even if the mark is not visibly displayed. See Brookfield Commc'ns, Inc. v. W. Coast Ent. Corp., 174 F.3d 1036, 1064-65 (9th Cir. 1999) (holding that use of competitor's trademark in metatags, which are not visible on a website, is actionable under the Lanham Act). Other circuits suggest the same. See, e.g., 1-800 Contacts, Inc. v. WhenU.Com, Inc., 414 F.3d 400, 411 (2d Cir. 2005) (recognizing that the use of metatags may involve conduct that constitutes a "use" under the Lanham Act). But this case presents a different question: Whether an action, like bidding on keywords, that involves no display or presentation of a mark whatsoever satisfies the "use in commerce" definition. In other words, does a buyer of advertising keywords who bids on certain terms and phrases "use" its competitor's mark when bidding on it?
In Network Automation, we answered, yes. 638 F.3d at 1144-45. But we provided no analysis to support this holding, id. at 1145, and we relied on cases with meaningfully different facts. Given that the cases on which Network Automation relied are readily distinguishable, the purpose of trademark infringement actions and modern practice on the internet suggest we may have gotten it wrong. (explaining how § 1127 evolved to apply to the infringement context, despite Congress's apparent intention that it apply to registration of trademarks).
I. Network Automation relied on factually distinguishable cases.
A. Rescuecom did not consider purchasers of advertising keywords.
Network Automation relied almost exclusively on the Second Circuit's decision in Rescuecom Corp. v. Google, Inc., 562 F.3d 123, 129 (2d Cir. 2009), for its conclusion that purchasing advertising keywords satisfies the "use in commerce" definition. 638 F.3d at 1145 (citing Rescuecom and concluding, "[w]e now agree with the Second Circuit that such use is a 'use in commerce' under the Lanham Act"). But the plaintiff in Rescuecom sued Google, the seller of the keywords, not the buyer of the keywords. 562 F.3d at 129. Specifically, the plaintiff alleged that Google's "Adwords" program and Keyword Suggestion Tool used the plaintiff's marks to cause consumer confusion. Id. at 12526. The district court granted Google's motion to dismiss, holding that Google did not use Rescuecom's mark in commerce. Id. at 127. The Second Circuit reversed. Id. at 131. It explained that Google satisfied § 1127's "use or display" definition because Google "displays, offers, and sells Rescuecom's mark to [its] advertising customers when selling its advertising services." Id. at 129. By "recommending and selling [Rescuecom's mark] to its advertisers," Google necessarily displayed Rescuecom's trademark in the sale of services. Id. The Second Circuit's decision in Rescuecom is based on the display of a trademark, a fact that does not exist here.
Purchasers of keywords do not display the mark. Here, Lerner &Rowe alleges that ALG bid on certain search terms-including "Lerner &Rowe"-and having been the highest bidder, paid Google to place its own advertisement near the top of the list when users use that search term. This process does not involve ALG displaying Lerner &Rowe's mark. Google-not ALG-displayed, offered, and sold the advertising term consisting of Lerner &Rowe's mark. While Google or other search engine providers may "use" trademarks by displaying and selling them as advertising words, it does not necessarily follow that bidding on those advertising words involves a "use." And, to be sure, the buyer of keywords does not in any way display a trademark to sell or advertise services.
B. Purchasing adwords is not comparable to using metatags.
Network Automation also pointed to a separate line of cases involving metatags to support its holding. Metatags are snippets of HTML code that describe the contents of the website. Brookfield, 174 F.3d at 1045. During the earlier days of the internet, many search engines relied on metatags in code to rank their search results. 4 J. Thomas McCarthy, McCarthy on Trademarks and Unfair Competition § 25A:3 (5th ed. 2024). "The more often a term appear[ed] in the metatags and in the text of the web page, the more likely it [wa]s that the web page [would] be 'hit' in a search for that keyword and the higher on the list of 'hits' the web page [would] appear." Brookfield, 174 F.3d at 1045. Internet users took advantage of this system, incorporating their competitors' trademarks into their website codes to improve the likelihood of appearing in a search for their competitor's mark.
We have previously assumed without expressly deciding that this type of conduct with metatags constitutes a "use in commerce." Id. at 1062-63. In Brookfield, we held that the use of metatags was actionable because it could cause initial interest confusion. Although the parties did not expressly raise the "use in commerce" issue, our conclusion implied that metatags constituted such a use.
But incorporating metatags consisting of a competitor's trademark into a website code is comparable to displaying or presenting a mark. Rescuecom explained, and we appear to have endorsed the view that such "internal" displays still constitute a "use in commerce." See, e.g., 562 F.3d at 129 (explaining that "use of a trademark in a software program's internal directory [does not] preclude[] a finding of trademark use"). Even if metatags do not involve an external display, they are functionally equivalent to "affixing" the competitor's mark to the product-a defendant affixes the competitor's mark to its website through its code to gain the benefits of the mark. This is precisely what the "use in commerce" requirement aims at. McCarthy, supra, § 23:11.50 (explaining that the "use in commerce" definition in § 1127 is a "relaxed remnant" of trademark law's requirement that a user "affix" a trademark to goods to obtain trademark protection).
A defendant bidding on keywords may not be the same as a defendant incorporating its competitor's trademarks into its own website. Although metatags and bidding on keywords are similar because neither involve a visible display of the competitor's mark on the defendant's website, the visibility of the mark or lack thereof is not what constitutes "use." Metatags constitute a "use" because the defendant affixes the competitor's mark to its website via its code. In contrast, keyword bidding does not require the defendant to display or affix a mark-internally or externally-in the advertising of its services.
Here, Google, not ALG, displayed Lerner &Rowe's mark on its website. ALG merely bid on keywords. Even if bidding on keywords resulted in the display of ALG's advertisements when consumers searched for Lerner &Rowe's mark, Google and not ALG is responsible for displaying the mark. Whether the defendant used a mark thus requires us to look at the defendant's conduct. Purchasing keywords may not be the same as using metatags for purposes of "use in commerce."
II. We should reconsider our holding in Network Automation en banc.
Because purchasing keywords is different than selling them or using metatags, Network Automation's holding is unsupported by existing case law. When considering whether ALG used or displayed Lerner &Rowe's mark in the sale or advertising of its services, 15 U.S.C. § 1127, the more reasoned conclusion may be that it did not. As noted above, ALG did not affix, display, offer, or present Lerner &Rowe's mark to any consumers. And while "use in commerce" is a relatively permissive standard, Network Automation, 638 F.3d at 1145, it is not boundless. Multiple considerations support the conclusion that the boundary could be drawn at ALG's conduct in this case.
First, trademark infringement typically requires presenting the mark to the allegedly confused consumers. In an ordinary infringement case, the defendant's presentation of a similar mark causes consumer confusion about the source of the goods or services. See, e.g., Surfvivor Media, Inc. v. Survivor Prods., 406 F.3d 625, 629 (9th Cir. 2005). ALG's actions look nothing like the ordinary case. Indeed, ALG never presented Lerner &Rowe's marks to the consumer on the other end of the search engine-or to any consumer at all. Google users entered their chosen search terms, and Google arranged the results, including sponsored advertisements, for the user. To the extent ALG displayed or presented anything to the consumer, it presented its own mark, which both parties acknowledge is not similar to "Lerner &Rowe." An action based only on one's own placement of their own product appears outside the realm of what the Lanham Act seeks to protect. 15 U.S.C. § 1114(a)(1).
Second, the traditional likelihood of confusion factors are not well-suited to address these circumstances. As Network Automation noted, even the Sleekcraft factors that typically apply in the internet context are "a particularly poor fit for the question presented here." 638 F.3d at 1148. We noted, for example, that an inquiry into the similarity of the marks "is impossible here where the consumer does not confront two distinct trademarks." Id. at 1151. Ultimately, Network Automation devised an entirely new factor to deal with competitive keyword advertising: "labeling and appearance." Id. at 1153-54. We give this factor great weight in our analysis. Id. (explaining that "likelihood of confusion will ultimately turn on what the consumer saw on the screen and reasonably believed, given the context"). Rather than continue relying on a nearly dispositive factor created exclusively for this context with little guidance, we should consider correcting our precedent and holding that purchasers of keywords do not "use" their competitors' trademarks in commerce.
And third, given the predominance of the internet in our lives, this type of advertising has become commonplace. Scrolling through sponsored ads at the top of a results page is often the rule-not the exception-when using a search engine. The familiarity of sponsored ads to those navigating internet platforms makes the likelihood of confusion inquiry difficult, if not impossible, to satisfy. McCarthy, supra, § 25A:7 ("Courts almost always find no likelihood of confusion if all that [a] defendant has done is use another's mark as a keyword to trigger an ad for defendant in which the other's trademark does not appear."). Consumers likely understand that, even when they search for a trademarked term, the sponsored results may not be associated with that trademark. This is not because the keyword purchaser has displayed or incorporated the trademark into its own page, but because sophisticated internet consumers understand the general norms and context in which internet advertisements appear. See Toyota Motor Sales, U.S.A., Inc. v. Tabari, 610 F.3d 1171, 1178 (9th Cir. 2010) (explaining that "[c]onsumers who use the internet for shopping are generally quite sophisticated" about how the internet works).
* * *
Twenty-five years ago, we recognized that "emerging technologies require a flexible approach" in the internet context. Brookfield, 174 F.3d at 1054. But that flexible approach is limited by the plain text and purpose of the Lanham Act. At bottom, trademark law is designed to protect parties against infringing uses of their marks. Bidding on and purchasing keyword search terms may not constitute such a use. We should take the opportunity to directly address this issue en banc rather than relying on our holding in Network Automation.
[*] The Honorable Edward M. Chen, United States District Judge for the Northern District of California, sitting by designation.
[**] This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.