1-800-Contacts, Inc. v. Federal Trade Comission

18-3848 1-800-Contacts, Inc. v. Federal Trade Comission

In the United States Court of Appeals for the Second Circuit August Term, 2019

Argued: March 5, 2020 Decided: June 11, 2021

Docket No. 18-3848

1-800 CONTACTS, INC.,

Petitioner,

v.

FEDERAL TRADE COMMISSION,

Respondent.

ON PETITION FOR REVIEW OF AN ORDER OF THE FEDERAL TRADE COMMISSION (DOCKET NO. 9372)

Before: LYNCH and MENASHI, Circuit Judges.*

• • Judge Peter W. Hall, originally a member of the panel, died on March 11, 2021. The two remaining members of the panel, who are in agreement, have determined the matter. See 28 U.S.C. § 46(d); 2d Cir. IOP E(b); United States v. Desimone, 140 F.3d 457, 458-59 (2d Cir. 1998).

1-800 Contacts, Inc., petitions from a Final Order of the Federal Trade Commission (FTC) finding that agreements between Petitioner 1-800 Contacts, Inc. and various competitors to, among other things, refrain from bidding on “keyword” search terms for internet advertisements, violate Section 5 of the FTC Act, 15 U.S.C. § 45. We hold that although trademark settlement agreements are not immune from antitrust scrutiny, the FTC (1) improperly considered the agreements to be “inherently suspect” and (2) incorrectly concluded that the challenged agreements are a violation of the FTC Act under the “rule of reason.”

PETITION FOR REVIEW GRANTED, FINAL ORDER VACATED AND REMANDED.

STEPHEN FISHBEIN, Shearman & Sterling LLP, New York, NY (Ryan A. Shores, Todd M. Stenerson, Brian C. Hauser, Shearman & Sterling LLP, Washington, D.C., on the brief) for Petitioner.

IMAD ABYAD, Federal Trade Commission (Gail F. Levine, Deputy Director, Geoffrey M. Green, Assistant Director, Joel Marcus, Deputy General Counsel, Barbara Blank, Daniel J. Matheson, Mariel Goetz, Attorneys, on the brief), for Alden F. Abbott, General Counsel, Federal Trade Commission, Washington, D.C., for Respondent.

Corbin K. Barthold and Cory L. Andrews, Washington Legal Foundation, Washington, D.C. for Amici Curiae Richard A. Epstein, Keith N. Hylton, Thomas A. Lambert, Geoffrey A. Manne, Hal Singer and Washington Legal Foundation in Support of Petitioner.

Theodore H. Davis Jr., Kilpatrick Townsend & Stockton LLP, Atlanta, GA and Sheldon H. Klein, President, American Intellectual Property Law Association, Arlington, VA, for Amicus Curiae American Intellectual Property Law Association in Support of Petitioner.

Bryan D. Gant, Seiji Niwa, White & Case LLP, New York, NY and Eileen M. Cole, White & Case LLP, Washington, D.C., for Amici Curiae United States Council for International Business in Support of Petitioner.

Mark A. Lemley, William H. Neukom Professor, Stanford Law School, Stanford, CA for Amici Curiae Intellectual Property, Internet Law and Antitrust Professors in Support of Respondent.

PER CURIAM:

Between 2004 and 2013, Petitioner 1-800 Contacts, Inc. (“1-800”) entered into thirteen trademark settlement agreements and one sourcing and services agreement with competitors (the “Challenged Agreements”). As explained below, the Challenged Agreements contained provisions restricting specific terms on which the parties could “bid” when participating in auctions held by companies that operate search engines. By restricting bidding on terms in these auctions, the competitors agreed not to advertise their products when consumers *4 used the search engines’ platforms to search the specific terms at issue. In August 2016, the Federal Trade Commission (“FTC” or the “Commission”) issued an administrative complaint against Petitioner, alleging that the Challenged Agreements and Petitioner’s enforcement of the agreements unreasonably restrain truthful, non-misleading advertising as well as price competition in search advertising auctions in violation of Section 5 of the FTC Act, 15 U.S.C. § 45. The claim was tried before an Administrative Law Judge (ALJ), who in 2017 issued an Initial Decision and Order finding that the agreements violate Section 5. Petitioner then appealed to the full Commission, which affirmed the ALJ’s conclusion in a three to one decision, with one Commissioner not participating. This timely petition for review followed the issuance of the Commission’s Final Order.

Although we hold that trademark settlement agreements are not automatically immune from antitrust scrutiny, the Commission’s analysis of the alleged restraints under the “inherently suspect” framework was improper. We further hold that the Commission incorrectly concluded that the agreements are an unfair method of competition under the FTC Act. We therefore GRANT the *5 petition for review, VACATE the Final Order of the Commission, and REMAND the case to the Commission with orders to DISMISS the administrative complaint.

BACKGROUND

Contact lenses, prescription eyewear designed to improve the user’s vision, can be sold only pursuant to a prescription. Such prescriptions specify both the characteristics of the lens, such as its strength, and the manufacturer brand. Thus, when consumers purchase contact lenses, they may not substitute one brand for another, but must purchase the brand listed on the prescription. Contact lenses are sold by four different types of retailers: independent eye care professionals; optical retail chains; mass merchants and club stores; and purely internet-based retailers, such as Petitioner. Internet-based retailers accounted for 17 percent of all contact lens sales in 2015, the year before these proceedings began. 1-800 accounts for a majority of all online sales of contact lenses. The price of contact lenses varies significantly based on retail channel; independent eye care professionals typically charge the most, followed by retail chains, mass merchants, and then online retailers. Petitioner, however, admits that it charges more than its rival online retailers. It prices its lenses somewhere below independent professionals and retail chains but above mass merchants and other club stores.

Petitioner and its competitors pay to advertise their sales of contact lenses on the internet. One way they do this is via “search advertising.” When an online shopper uses a search engine such as Google or Bing, the search engine’s program returns two types of results to the shopper: “sponsored” and “organic,” both of which provide links to web pages. Sponsored results are ads; they appear because the owner of the featured web page has paid for its page to appear in that space. Sponsored links are typically designated by a label like “Ad” or “Sponsored,” and by colored or shaded boxes around the link. Organic results, on the other hand, appear based exclusively on which results a search engine’s algorithm deems to be most relevant to the shopper’s search. Organic results are listed separately from the sponsored results.

Search engines determine which advertisements to display on a search results page based in part on the relevance or relation of the consumer’s search to various words or phrases called “keywords.” Advertisers bid on these keywords during auctions hosted by the search engines. The highest bidders’ ads are typically displayed most prominently on a page, though search engines consider other factors when determining where to place an ad on a results page, such as an ad’s quality and relevance to a consumer’s search. Search engines generally do *7 not limit the keywords available to advertisers at auction. As a result, competitors often bid on each other’s brand names so that their ad runs when a consumer searches for a competitor. Brand name terms are often trademarked.

Via bidding on “negative keywords,” an advertiser may also prevent its ad from being displayed when a consumer searches for a particular keyword. These negative keywords preclude ads from being displayed even when the search engine independently determined that the ad would be relevant to the consumer. The Commission suggests that this is useful when, for example, a retailer selling eyeglasses has bid on the advertising keyword “glasses” but wants to prevent its ad from appearing in response to the term “wine glasses.”

Many online retailers of contact lenses devote the majority of their advertising budgets to search advertising. The Commission found that these ads are presented to consumers “at a time when [they are] more likely looking to buy.” JA 279. Unlike its online retail competitors, Petitioner also uses other methods of advertising, including printed materials, radio, and television. Online search advertising, however, still represents a large portion of Petitioner’s advertising budget. Because Petitioner charges more than other online retailers, when its *8 competitors’ ads appear in response to a search for 1-800’s trademark terms, Petitioner’s sales tend to decrease.

In 2002, Petitioner began filing complaints and sending cease-and-desist letters to its competitors alleging trademark infringement related to its competitors’ online advertisements. 1 Between 2004 and 2013, Petitioner entered into thirteen settlement agreements to resolve most of these disputes. Each of these agreements includes language that prohibits the parties from using each other’s trademarks, URLs, and variations of trademarks as search advertising keywords. The agreements also require the parties to employ negative keywords so that a search including one party’s trademarks will not trigger a display of the other party’s ads. The agreements do not prohibit parties from bidding on generic keywords such as “contacts” or “contact lenses.” 2 Petitioner enforced the agreements when it perceived them to be breached.

1 Some of these trademark infringement allegations did not involve search advertising. 2 Advertisers may designate how closely the keyword they bid on must match the consumer’s search in order for their ads to be displayed. For example, Google, the leading search engine in the United States, offers different options for advertisers on its paid search platform. An advertiser may designate the keyword on which they have bid as “broad match,” “phrase match,” “exact match,” or, as mentioned, “negative match.” An ad tied to a keyword designated as “broad match” may appear when a consumer’s search on Google contains the specific keyword, any plural forms or synonyms of the keyword, or phrases similar to the keyword. An advertiser selecting “phrase match” will have its ad appear when a search contains additional words along with the keyword. An “exact match” limits the ad’s appearance to when the consumer searches

Apart from the settlement agreements, in 2013 Petitioner entered into a “sourcing and services agreement” with Luxottica, a company that sells and distributes contacts through its affiliates. JA 283. That agreement also contains reciprocal online search advertising restrictions prohibiting the use of trademark keywords and requiring both parties to employ negative keywords.

The FTC issued an administrative complaint against Petitioner in August 2016 alleging that the thirteen settlement agreements and the Luxottica agreement (the “Challenged Agreements”), along with subsequent actions to enforce them, unreasonably restrain truthful, non-misleading advertising as well as price competition in search advertising auctions, all of which constitute a violation of Section 5 of the FTC Act, 15 U.S.C. § 45. 3 The complaint alleges that the Challenged Agreements prevented Petitioner’s competitors from disseminating ads that would have informed consumers that the same contact lenses were available at a cheaper price from other online retailers, thereby reducing

the exact keyword. “Negative keywords” can also be designated by an advertiser for broad, phrase, or exact match. The Challenged Agreements do not specify whether the negative keywords must be employed using broad, phrase, or exact match. 3Section 5 of the FTC Act states that “[u]nfair methods of competition in or affecting commerce, and unfair or deceptive acts or practices in or affecting commerce, are hereby declared unlawful.” 15 U.S.C. § 45(a)(1).

competition and making it more difficult for consumers to compare online retail prices. The case was tried before an ALJ, who concluded that a violation had occurred.

As an initial matter, the ALJ rejected Petitioner’s assertion that trademark settlement agreements are not subject to antitrust scrutiny in light of FTC v. Actavis, 570 U.S. 136 (2013). Applying the “rule of reason” and principles of Section 1 of the Sherman Act, 15 U.S.C. § 1, the ALJ determined that “[o]nline sales of contact lenses constitute a relevant product market.” JA 120. He found that the agreements constituted a “contract, combination, or conspiracy” as required by the Sherman Act and held that the advertising restrictions in the agreements harmed consumers by reducing the availability of information, in turn making it costlier for consumers to find and compare contact lens prices. JA 184, 221-22.

Having found actual anticompetitive effects, as required under the rule of reason analysis, the ALJ rejected the procompetitive justifications for the agreements offered by Petitioner. He found that while trademark protection is procompetitive, it did not justify the advertising restrictions in the agreements and also that Petitioner failed to show that reduced litigation costs would benefit consumers. The ALJ issued an order that barred Petitioner from entering into an *11 agreement with any marketer or seller of contact lenses to limit participation in search advertising auctions or to prohibit or limit search advertising.

1-800 appealed the ALJ’s order to the Commission. In a split decision, a majority of the Commission agreed with the ALJ that the agreements violated Section 5 of the FTC Act. The majority, however, analyzed the settlement agreements differently from the ALJ. The majority classified the agreements as “inherently suspect” and alternatively found “direct evidence” of anticompetitive effects on consumers and search engines. The majority then analyzed the procompetitive justifications Petitioner offered for the agreements and rejected arguments that the benefits of protecting trademarks and reducing litigation costs outweighed any potential harm to consumers. Finally, the majority identified what it believed to be less anticompetitive alternatives to the advertising restrictions in the agreements. One Commissioner dissented, reasoning both that the majority should not have applied the “inherently suspect” framework and that it failed to give appropriate consideration to Petitioner’s proffered procompetitive justifications. This timely appeal followed.

JURISDICTION AND STANDARD OF REVIEW

We have jurisdiction over this appeal under 15 U.S.C. § 45(c). The majority opinion of the Commission “adopt[ed] the ALJ’s findings of fact to the extent that they [were] not inconsistent” with its opinion. JA 285. Factual findings of the Commission are binding “if they are supported by such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” FTC v. Ind. Fed’n of Dentists (IFD), 476 U.S. 447, 454 (1986) (internal quotation marks and citation omitted). The Commission’s legal conclusions are “for the courts to resolve, although even in considering such issues the courts are to give some deference to the Commission’s informed judgment that a particular commercial practice is to be condemned as ‘unfair.’” Id.

DISCUSSION

I. Actavis Considerations

Petitioner argues, as it did below, that trademark litigation settlements are generally immune from antitrust review. It contends that in Actavis, the Supreme Court “cabin[ed] its extension of antitrust scrutiny” to the “unusual” intellectual property settlements at issue there and did not intend to implicate “commonplace” *13 settlements. Petitioner’s Br. 43. Neither the ALJ nor any participating member of the Commission found this argument persuasive. 4 Nor do we.

In Actavis, the Supreme Court analyzed what are known as “reverse payment” patent settlements. 570 U.S. at 141. In short, manufacturers of brand name drugs paid manufacturers of generic drugs to keep the generic manufacturers from litigating the validity of the brand name manufacturers’ patents. See id. at 145. This effectively allowed the brand name manufacturers to maintain exclusive sales of certain drugs for longer than they would have if the applicable patent, through litigation, was found to be invalid. Id. at 153-54. In Actavis, the Court rejected the idea that the conduct at issue was immune from antitrust scrutiny just because it occurred within the context of a patent litigation settlement. Id. at 146-48. The Court explained that “it would be incongruous to determine antitrust legality by measuring the settlement’s anticompetitive effects solely against patent law policy, rather than by measuring them against procompetitive antitrust policies as well.” Id. at 148.

4 All four participating Commissioners agreed that Actavis does not immunize trademark settlement agreements from antitrust scrutiny. Commissioner Wilson did not participate in the appeal to the Commission.

Petitioner argues that Actavis represents an exception to the general rule against subjecting intellectual property (IP) settlement agreements to antitrust scrutiny because patents, unlike trademarks, for example, are inherently exclusionary and because the reverse payment scheme at issue in Actavis was “unusual.” Petitioner’s Br. 43 (citing Actavis, 570 U.S. at 147). To be sure, in Actavis the Court detailed how certain commonplace forms of settlement agreements did not, by the nature of their existence alone, create antitrust liability. 570 U.S. at 151-52. Contrary to Petitioner’s claim, however, the Court went on to say that the possibility that agreements may not always bring about anticompetitive consequences “does not justify dismissing the FTC’s complaint. An antitrust defendant may show in the antitrust proceeding that legitimate justifications are present[.]” Actavis, 570 U.S. at 156.

As in Actavis, Petitioner’s trademark, “if valid and infringed, might have permitted it to” preclude competitors from bidding on its trademarked terms in search advertising auctions or running advertisements on those terms. Id. at 147. We “take this fact as evidence that the agreement’s anticompetitive effects fall within the scope of” the trademark protections. Id. (internal quotation marks omitted). But the mere fact that an agreement implicates intellectual property *15 rights does not “immunize [an] agreement from antitrust attack.” Id.; see also In re Indep. Serv. Orgs. Antitrust Litig., 203 F.3d 1322, 1325 (Fed. Cir. 2000) (“Intellectual property rights do not confer a privilege to violate the antitrust laws.”); United States v. Microsoft Corp., 253 F.3d 34, 63 (D.C. Cir. 2001) (same). We have not shied away from considering antitrust claims that implicate trademark rights in the past, see, e.g., Clorox Co. v. Sterling Winthrop, Inc., 117 F.3d 50, 55-56 (2d Cir. 1997), and we decline to do so now. As in any antitrust case, we must “determine whether the restraints in the agreement[s] are reasonable in light of their actual effects on the market and their pro-competitive justifications.” Id. at 56.

II. Sherman Act Framework

Because “[t]he FTC Act's prohibition of unfair competition and deceptive acts or practices . . . overlaps the scope of § 1 of the Sherman Act . . . aimed at prohibiting restraint of trade,” California Dental Ass'n v. FTC (Cal. Dental), 526 U.S. 756, 762 n. 3 (1999), it was appropriate that the ALJ and the Commission consulted Sherman Act jurisprudence to determine whether the Challenged Agreements violated Section 5 of the FTC Act. See Realcomp II, Ltd. v. FTC, 635 F.3d 815, 824 (6th Cir. 2011); North Carolina Bd. of Dental Examiners v. FTC, 717 F.3d 359, 370-71 *16 (4th Cir. 2013) (recognizing a Section 1 violation as a “species” of unfair competition prohibited under the FTC Act).

To prove a Sherman Act violation – and by extension, a Section 5 violation – the FTC must establish (1) a contract, combination, or conspiracy exists that (2) unreasonably restrains trade. See Major League Baseball Props., Inc. v. Salvino, Inc. (MLB), 542 F.3d 290, 315-16 (2d Cir. 2008). In this case, the Challenged Agreements are undeniably contracts between Petitioner and its competitors. We “presumptively appl[y]” what is known as the “rule of reason” analysis to the Challenged Agreements to determine whether they restrain trade. Texaco Inc. v. Dagher, 547 U.S. 1, 5 (2006). Under that analysis an antitrust plaintiff “must demonstrate that a particular contract or combination is in fact unreasonable and anticompetitive before it will be found unlawful.” Id. As Justice Brandeis famously articulated:

The true test of legality is whether the restraint imposed is such as merely regulates and perhaps thereby promotes competition or whether it is such as may suppress or even destroy competition. To determine that question the court must ordinarily consider the facts peculiar to the business to which the restraint is applied; its condition before and after the restraint was imposed; the nature of the restraint and its effect, actual or probable. The history of the restraint, the evil believed to exist, the reason for adopting the particular remedy, the purpose or end sought to be attained, are all relevant facts. This is not because a good intention will save an otherwise objectionable

regulation or the reverse; but because knowledge of intent may help the court to interpret facts and to predict consequences.

Chicago Board of Trade v. United States, 246 U.S. 231, 238 (1918). A plaintiff bears the initial burden of showing that the challenged action has had an actual adverse effect on competition as a whole in the relevant market. North Am. Soccer League, LLC v. U.S. Soccer Fed’n, Inc., 883 F.3d 32, 42 (2d Cir. 2018). After a prima facie case of anticompetitive conduct has been established, the burden shifts to the defendant to proffer procompetitive justifications for the agreement. Id. “Assuming defendants can provide such proof, the burden shifts back to the plaintiffs to prove that any legitimate competitive benefits offered by defendants could have been achieved through less restrictive means.” Geneva Pharm. Tech. Corp. v. Barr Labs. Inc., 386 F.3d 485, 507 (2d Cir. 2004).

In some cases, however, “certain agreements or practices which because of their pernicious effect on competition and lack of any redeeming virtue are conclusively presumed to be unreasonable and therefore illegal without elaborate inquiry as to the precise harm they have caused or the business excuse for their use.” Northern Pacific Ry. Co. v. United States, 356 U.S. 1, 5 (1958). Such agreements are deemed per se illegal. See MLB, 542 F.3d at 315. This designation is saved for certain types of restraints, e.g., geographic division of markets or horizontal price *18 fixing, that have been established over time to “lack . . . any redeeming virtue.” Id.

The Supreme Court, however, has rejected fixed categories of analysis when considering the anticompetitive nature of a restraint. See Cal. Dental, 526 U.S. at 779. Some restraints, therefore, fall between the type of conduct typically labeled per se anticompetitive and that which is analyzed under a “full-blown” rule of reason analysis. MLB, 542 F.3d at 317. When “the great likelihood of anticompetitive effects can easily be ascertained[,]” courts apply an abbreviated rule of reason analysis sometimes known as the “quick-look” approach. Cal. Dental, 526 U.S. at 770. The Commission calls the standard it applies in these situations the “inherently suspect” framework. 5 JA 291.

Under the Commission’s “inherently suspect” framework, neither direct evidence of harm nor proof of market power is needed to show the anticompetitive effect of the restraint because the “likely tendency to suppress competition” posed by the challenged conduct makes it “inherently suspect.” Polygram Holding, Inc., 136 F.T.C. 310, 344-45 (2003), aff’d, 416 F.3d 29 (D.C. Cir. 2005). An “elaborate

5The “quick-look” and “inherently suspect” approaches are similar, and Petitioner does not take issue with the interchangeability of the two formulations of this abbreviated analysis.

market analysis” is unnecessary, Polygram, 416 F.3d at 35, and once the government has identified a “suspect” agreement, the burden shifts directly to the defendant to show any procompetitive justifications it might have for the restraint. See United States v. Apple, 791 F.3d 290, 330 (2d Cir. 2015).

This approach is only permissible when “an observer with even a rudimentary understanding of economics could conclude that the arrangements in question would have an anticompetitive effect on customers and markets.” Cal. Dental, 526 U.S. at 770; see also Dagher, 547 U.S. at 7 n.3 (rejecting a quick-look analysis because it applies only “to business activities that are so plainly anticompetitive that courts need undertake only a cursory examination before imposing antitrust liability”); Polygram, 416 F.3d at 37 (explaining that the inherently suspect framework is only applicable when “close family resemblance [exists] between the suspect practice and another practice that already stands convicted in the court of consumer welfare”).

Further, “[i]f an arrangement ‘might plausibly be thought to have a net procompetitive effect, or possibly no effect at all on competition,’ more than a ‘quick look’ is required.” MLB, 542 F.3d at 318 (quoting Cal. Dental, 526 U.S. at 771). In California Dental, the Supreme Court considered the California Dental *20 Association’s rule prohibiting price advertising, specifically discounted fees, and advertising relating to the quality of dental services. 526 U.S. at 761. There, the Court rejected the use of an abbreviated rule of reason analysis, holding that the existence of a plausible procompetitive justification – in that case, the prohibition of deceptive advertising in an asymmetrical information marketplace – effectively foreclosed the ability of courts to utilize the quick look approach. See id. at 771.

Here, the Commission viewed the advertising restrictions in the Challenged Agreements as inherently suspect; it also found that the agreements were a form of “bid rigging” that harmed search engines – i.e., an independent basis upon which it could apply the inherently suspect analytical framework. Petitioner and amici argue that the application of the inherently suspect framework was improper and that the Challenged Agreements should only be considered under a rule of reason analysis. We agree with Petitioner that the Challenged Agreements cannot be classified as inherently suspect.

Citing expert reports and economic theory, the government argues that the Commission was correct to employ the inherently suspect framework because restrictions on advertising are likely to cause consumers to pay more for contact lenses. But even if restraints on truthful advertising have a tendency to raise *21 prices, “[t]he fact that a practice may have a tangential relationship to the price of the commodity in question does not mean that a court should dispense with a full rule-of-reason analysis.” MLB, 542 F.3d at 317.

Crucially, the restraints at issue here could plausibly be thought to have a net procompetitive effect because they are derived from trademark settlement agreements. In Clorox, applying the rule of reason, we considered whether a trademark settlement agreement illegally restrained trade under the Sherman Act and we explained that “[t]rademarks are by their nature non-exclusionary.” 117 F.3d at 55-56. Agreements to protect trademarks, then, should not immediately be assumed to be anticompetitive – in fact, Clorox tells us instead to presume they are procompetitive. Id. at 60. As the Challenged Agreements restrict the parties from running advertisements on Petitioner’s trademarked terms, they directly implicate trademark policy.

The Commission acknowledged as much, finding Petitioner’s proffered procompetitive justifications to be “cognizable and, at least, facially plausible[.]” JA 296. Rather than take that fact as an indication that it should not apply an abbreviated rule of reason analysis, as the Supreme Court instructed in California Dental, the Commission instead set out to show (i) that there was a theoretical basis *22 for the alleged anticompetitive effect and that the restraints were likely, in this particular context, to harm competition and (ii) that Petitioner could have minimized the anticompetitive effects and accomplished its procompetitive justifications through less restrictive means. While this may be analytically acceptable in some situations, see Cal. Dental, 526 U.S. at 779 (noting to require a “more extended examination” does not always translate to a call for “plenary market examination”), it was not appropriate here.

Courts do not have sufficient experience with this type of conduct to permit the abbreviated analysis of the Challenged Agreements undertaken by the Commission. See Cal. Dental, 526 U.S. at 781 (explaining that the quick-look approach may be applicable if rule-of-reason analyses in case after case reach identical conclusions); Polygram, 416 F.3d at 36-37 (accepting the Commission’s definition of “inherently suspect” as describing restraints previously condemned by both “judicial experience and economic learning”). While both California Dental and Polygram consider advertising restraints, there are key differences between the restraints in those cases and the restraints here, 6 and our own

6In California Dental, as an initial matter, the court rejected the use of an abbreviated rule of reason analysis for restraints on price and quality advertising. 526 U.S. at 781. And in Polygram, the fact that the challenged conduct also restricted the parties from offering discounts on concert

precedent suggests that trademark agreements like those at issue here need to be examined using a fuller analysis. See Clorox, 117 F.3d at 55-56, 59 (applying a rule of reason analysis and rejecting the alleged anticompetitive harm of a trademark agreement); see also Actavis, 570 U.S. at 158-59 (rejecting the application of a quick-look analysis for intellectual property agreements). When, as here, not only are there cognizable procompetitive justifications but also the type of restraint has not been widely condemned in our “judicial experience,” see Polygram, 416 F.3d at 37, more is required. Cf. Bogan v. Hodgkins, 166 F.3d 509, 514 n.6 (2d Cir. 1999) (noting pre-California Dental that “[u]nder quick look, once the defendant has shown a procompetitive justification for the conduct, the court must proceed to weigh the overall reasonableness of the restraint using a full-scale rule of reason analysis” (internal quotation marks and citation omitted)). The Challenged Agreements, therefore, are not so obviously anticompetitive to consumers that someone with only a basic understanding of economics would immediately recognize them to be

albums was key to the D.C. Circuit’s affirmation of the inherently suspect analysis; the court thought that the agreement looked “suspiciously like a naked price fixing agreement between competitors[.]” See Polygram, 416 F.3d at 37.

so. 7 See Cal. Dental, 526 U.S. at 770. We are bound, then, to apply the rule of reason. 8

III. Application of the Rule of Reason

Under the rule of reason, the Commission bears the burden of establishing a prima facie case of anticompetitive effect. Direct evidence of anticompetitive effects establishes a prima facie case of a Sherman Act Section 1 violation and obviates the need for a detailed market analysis or showing of market power. 9

7 We do not discount the economic evidence cited by the Commission in coming to its conclusion, but the fact that it required the testimony of expert witnesses who provided empirical analyses in order to determine the net competitive effect of the Challenged Agreements underscores the point: these restraints are not obviously anticompetitive to someone with only a rudimentary understanding of economics. See MLB, 542 F.3d at 340, n.10 (Sotomayor, J., concurring). 8 We also reject the Commission and amici’s arguments that the restrictions constitute illegal bid rigging as support for their use of the inherently suspect framework. An absolute ban on competitive bidding, or bid rigging, would be anticompetitive on its face and may justify an abbreviated rule of reason analysis. Cal. Dental, 570 U.S. at 770 (citing Nat'l Soc'y of Prof'l Eng'rs v. FTC, 435 U.S. 679, 692-93 (1978)); see also United States v. Joyce, 895 F.3d 673, 679 (9th Cir. 2018) (finding bid rigging to be per se illegal). It is not clear to us, however, that the restrictions constitute such a ban. The Challenged Agreements do not prevent the parties from participating in keyword auctions, only from bidding on trademarked terms. Whether restrictions on advertisers’ use of particular terms leads to overall harm to the search engines is not obvious and therefore does not justify analyzing the agreements under the inherently suspect framework. Nor, as amici in support of the government argue, is it obvious that the restrictions constitute market division, another type of restraint that would justify an abbreviated analysis. See Palmer v. BRG of Georgia, Inc., 498 U.S. 46 (1990) (per curiam). 9 Though the government argued to the ALJ that the parties to the Challenged Agreements collectively have market power in the relevant market and that the nature of the restraints makes it likely that the Challenged Agreements will have an anticompetitive effect, the ALJ chose not to determine the prima facie case under that theory, nor did the Commission on appeal.

See IFD, 476 U.S. at 460. The Commission contends that it satisfied its burden by adducing evidence of increased contact lens prices and a reduction in the quantity of advertisements.

A. Anticompetitive Effects

Anticompetitive effects in a relevant market may be shown through direct evidence of output reductions, increased prices, or reduced quality in the relevant market. Ohio v. Am Express Co. (Am. Express), 138 S. Ct. 2274, 2284 (2018); see also North Am. Soccer League, 883 F.3d at 42. The Commission has also defined sufficient evidence of anticompetitive harm to include evidence of “retarded innovation, or other manifestations of harm to consumer welfare.” In re Realcomp II Ltd., No. 9320, 2007 WL 6936319 (F.T.C. Oct. 30, 2009), aff’d 635 F.3d 815. We reject the Commission’s argument that it has established direct evidence of anticompetitive effect in the form of increased prices. When an antitrust plaintiff advances an antitrust claim based on direct evidence in the form of increased prices, the question is whether it can show an actual anticompetitive change in prices after the restraint was implemented. See Brooke Grp. Ltd. v. Brown & Williamson Tobacco Corp., 509 U.S. 209, 236-37 (1993); MacDermid, 833 F.3d at 184. The government could not make that showing because it did not conduct an *26 empirical analysis of the Challenged Agreements’ effect on the price of contact lenses in the online market for contacts. The evidence offered by the government is theoretical and anecdotal; 10 it is not “direct.” Consequently, the Commission’s conclusion that differences between 1-800 Contacts’ prices and those of its competitors constitute direct evidence of the Challenged Agreements’ anticompetitive effects is not supported by substantial evidence. 11

10The government argues, inter alia, that (a) Petitioner admits that it charges more than their competitors in the relevant market; (b) economic theory strongly suggests that advertising restrictions tend to increase prices of any given product; and (c) Petitioner offered to meet or beat any price offered by other online retailers. Even accepting all this as true, it is not direct evidence that the Challenged Agreements caused the price of contact lenses in the relevant market to rise, as our precedent requires. MacDermid, 833 F.3d at 184. 11 A slightly different issue plagues the government’s argument that there is direct evidence of reduced revenues for search engines. To show this, the government did not show that Google or Microsoft, the companies who control the two most popular search engines, had lower revenues after the Challenged Agreements were put into place. Nor did the government introduce evidence that Petitioner spent less money on search advertising than it did before the Challenged Agreements came into effect. Instead, the government offered empirical evidence that the Challenged Agreements reduced the price paid by Petitioner for each click on one of its keywords. JA 1096-99. Empirical evidence is, as noted, required under our caselaw to find direct evidence of an anticompetitive effect. K.M.B., 61 F.3d at 127. But showing that a price for certain keywords dropped is not direct evidence of the effect on the market as a whole. See Clorox, 117 F.3d at 56 (quoting K.M.B., 61 F.3d at 127). This snapshot shows only that Petitioner paid less for certain keyword advertisements, no more and no less. While it is true that, when evaluating “whether horizontal restraints had an adverse effect on competition,” we do not always “need to precisely define the relevant market to conclude that these agreements were anticompetitive,” Am. Express, 138 S.Ct. at 2285 n.7, more was needed under our precedent if the Commission wished to show “direct” evidence in the market as a whole. See Clorox, 117 F.3d at 56. The “product” offered by search engines of which the Challenged Agreements allegedly restrain trade is, presumably, advertisers’ use of keywords, but that is not clear from the record. For the same reason, we also cannot say that the Commission’s

The government also argues that “disrupted information flow” is an anticompetitive effect and that a reduction in the quantity of advertisements is direct evidence of that effect. Respondent’s Br. 63. While, to our knowledge, no Court of Appeals has held that a reduction of truthful information is necessarily a manifestation of anticompetitive harm, our sister circuits have occasionally considered advertising restraints in different contexts and have found the conduct in question to have anticompetitive effects. See, e.g., California Dental Ass’n v. FTC (Cal. Dental II), 224 F.3d 942, 949 (9th Cir. 2000) (considering professional advertising restraints in an asymmetrical information marketplace); Polygram, 416 F.3d at 37 (holding that the FTC appropriately concluded an agreement to restrain price cutting and advertising violated the FTC Act); Realcomp II, 635 F.3d at 831-32, 832 n.9 (denying petition for review when petitioner’s policy limited access to internet marketing); Blackburn v. Sweeney, 53 F.3d 825, 827-29 (7th Cir. 1995) (identifying an agreement not to advertise in certain geographic areas as a per se illegal attempt to allocate markets). We need not decide whether the Commission’s theory of harm is viable, however, because we conclude that

conclusion that the Challenged Agreements harmed search engines by reducing quality in the market was supported by substantial evidence.

Petitioner has shown a procompetitive justification and the Commission fails to carry its burden at the third step.

B. Procompetitive Justifications

Petitioner asserts that the Challenged Agreements are justified by two procompetitive effects: reduced litigation costs and protecting Petitioner’s trademark rights. The Commission found that, while both of these justifications were “cognizable and facially plausible,” Petitioner did not show that they “have a basis in fact,” and therefore they were not “valid.” JA 309. We disagree. The protection of Petitioner’s trademark interests constitutes a valid procompetitive justification for the Challenged Agreements.

The Commission determined that, since “the [Challenged Agreements] restrict a type of competitive advertising that has never been found to violate the trademark laws, and the weight of authority overwhelmingly points to noninfringement[,]” trademark protection was not a valid procompetitive benefit that justified the Challenged Agreements. JA 313. This was incorrect. Trademarks are by their nature non-exclusionary, and agreements to protect trademark interests are “common, and favored, under the law.” Clorox, 117 F.3d at 55. As a result, “it is difficult to show that an unfavorable trademark agreement creates *29 antitrust concerns.” Id. at 57. This is true even though trademark agreements inherently prevent competitors “from competing as effectively as [they] otherwise might[.]” Id. at 59.

In Clorox, we found that the plaintiff had failed to show adverse effects on the market as a whole because the restrictions at issue did not restrict competitors’ ability to enter into the relevant market. Id. at 59. Although we held that the plaintiff in that case failed to present a prima facie case of anticompetitive harm, we also went on to detail how the procompetitive justifications of the agreement weighed against finding an antitrust violation. Id. at 60. We stated that “trademark agreements are favored in the law as a means by which parties agree to market products in a way that reduces the likelihood of consumer confusion and avoids time-consuming litigation.” Id. And again, Clorox counsels that we should “presume” that trademark settlement agreements are procompetitive. Id.

The Commission, however, decided that the trademark claims that led to the Challenged Agreements were likely meritless. While it claimed not to be determining the validity of Petitioner’s trademark claims, it did just that by weighing the potential validity of the trademark claims in order to show that *30 Petitioner’s procompetitive justification was invalid. 12 Even if the Commission’s analysis of the underlying trademark claims were correct, trademark agreements that “only marginally advance[] trademark policies” can be procompetitive. 13 See id. at 57. Under Clorox, “[e]fforts to protect trademarks, even aggressive ones, serve the competitive purpose of furthering trademark policies.” Id. at 61.

That does not mean that every trademark agreement has a legitimate procompetitive justification. If the “provisions relating to trademark protection are auxiliary to an underlying illegal agreement between competitors,” or if there were other exceptional circumstances, 14 we would think twice before concluding the challenged conduct has a procompetitive justification. See id. at 60. As in

12Though the Luxottica agreement was not the result of a trademark dispute, the agreement contains trademark protections similar to those in the other Challenged Agreements. And while the restrictions in the Luxottica Agreement may best be considered under the doctrine of ancillary restraints, see MLB, 542 F.3d at 334 (Sotomayor, J., concurring), the government does not attempt here to offer direct evidence differentiating the reductions in advertising stemming solely from that agreement. In any event, because we hold that protecting trademarks is a valid procompetitive justification for the restrictions, we see no further need to differentiate the Luxottica Agreement from the other agreements. 13 At the time the agreements were entered into, the law regarding the validity of Petitioner’s trademark claims was unsettled, and it remains so in this Circuit. The fact that the law was unsettled at the time is one reason a party might enter into a settlement agreement. 14It has not been argued that exceptional circumstances exist in this case. We, therefore, need not decide what circumstances might qualify as exceptional, von Hofe v. United States, 492 F.3d 175, 185 n.3 (2d Cir. 2007), such as, for example, agreements between parties with unequal bargaining power. See Clorox, 117 F.3d at 60 (“There is no evidence that [a party to the challenged agreement] entered the agreement under duress.”).

Clorox, however, there is a lack of evidence here that the Challenged Agreements are the “product of anything other than hard-nosed trademark negotiations.” 15 Id. Consequently, we find Petitioner met its burden at step two.

C. Less Restrictive Alternatives

Because Petitioner has carried its burden of identifying a procompetitive justification, the government must show that a less restrictive alternative exists that achieves the same legitimate competitive benefits. 16 Am. Express, 138 S. Ct. at 2284; North Am. Soccer League, 883 F.3d at 42. That is, the restraint “only survives a rule of reason analysis if it is reasonably necessary to achieve the legitimate objectives proffered by the defendant.” United States v. Brown Univ., 5 F.3d 658, 678-79 (3d Cir. 1993). “Less restrictive alternatives are those that would

15Intent plays a role in this conclusion. See Clorox, 117 F.3d at 60 (citing Chicago Bd. of Trade, 246 U.S. at 238). There is ample evidence that, if Petitioner’s competitors had not been precluded by the Challenged Agreements from running ads on Petitioner’s trademarks, they would have done so; the competitors’ use of the terms is what spawned the agreements in the first place. This is unlike a typical market division case, where the two parties agree not to compete in the same geographic areas because it would benefit both of their bottom lines. Cf. Palmer, 498 U.S. at 49- 50 (holding an agreement between bar review course providers dividing market territories for the purpose of raising prices was per se illegal). 16 The government argues that if we find that the Commission improperly weighed the procompetitive justification of Petitioner’s trademark protections, we should remand to allow the Commission to reconsider in the first instance. The Commission, however, already determined that, assuming the procompetitive justifications were legitimate, they were not reasonably necessary to achieve the proffered procompetitive benefits. We need not remand to allow them to rearticulate the same point.

be less prejudicial to competition as a whole.” North Am. Soccer League, 883 F.3d at 45 (internal quotation marks omitted). The Commission found that the government had shown a viable less restrictive alternative, namely that the parties to the Challenged Agreements could have agreed to require clear disclosure in each search advertisement of the identity of the rival seller rather than prohibit all advertising on trademarked terms. According to the government, therefore, the Challenged Agreements are overbroad.

In Clorox, however, we noted that “it is usually unwise for courts to secondguess” trademark agreements between competitors. 117 F.3d at 60. In this context, what is “reasonably necessary,” Brown Univ., 5 F.3d at 679, is likely to be determined by competitors during settlement negotiations, Clorox, 117 F.3d at 60. And, as articulated above, absent something that would negate the typically procompetitive nature of these agreements, “the parties’ determination of the scope of needed trademark protections is entitled to substantial weight.” Clorox, 117 F.3d at 60.

The government attempts to differentiate Clorox by arguing that the FTC is different than a private plaintiff, and when it brings an antitrust claim we should not give the settling parties as much latitude to negotiate a trademark agreement *33 as a court would in a private antitrust suit. Even if we were to accept the Commission’s argument that its presence in a case warrants less solicitude for trademark interests, the government still needs to show more than the mere possibility there could be crafted an alternative form of the trademark agreement. The alternative must be “substantially less restrictive.” Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law: An Analysis of Antitrust Principles and Their Application ¶ 1502 (3rd & 4th eds., 2019 Cum. Supp. 2010-2018). The alternative must also achieve the same legitimate competitive benefits outlined by the Petitioner. North Am. Soccer League, 883 F.3d at 42. And at the end of the day, our job is to “weigh[] the competing evidence to determine if the effects of the challenged restraint tend to promote or destroy competition.” Apple, 791 F.3d at 329 (internal quotation marks omitted).

The Commission majority thought that a disclosure requirement was enforceable because, inter alia, it has ordered similar requirements in the past. But the majority failed to consider the practical reasons for the parties entering into the Challenged Agreements. Under Clorox, this was insufficient. 117 F.3d at 60-61. The Commission did not consider, for example, how the parties might enforce such a requirement moving forward or give any weight to how onerous *34 such enforcement efforts would be for private parties. When the restraint at issue in an antitrust action implicates IP rights, Actavis directs us to consider the policy goals of the relevant IP law. See 570 U.S. at 149. Here, those considerations must include the practical implications of the government’s proffered alternatives on the parties’ ability to protect and enforce their trademarks.

While trademark agreements limit competitors from competing as effectively as they otherwise might, we owe significant deference to arm’s length use agreements negotiated by parties to those agreements. Clorox, 117 F.3d at 59-60. Doing so may give rise to collateral harm in a relevant market. But forcing companies to be less aggressive in enforcing their trademarks is antithetical to the procompetitive goals of trademark policy. 17 See id. at 61. And without considering the downstream effects of requiring less aggressive enforcement, the

17We acknowledge a concern that the Challenged Agreements require the parties to employ negative keywords, which prevent ads of competitors from appearing in a consumer search for each other’s trademarked terms – even absent purchase of a keyword (but rather due to a search engine’s independent determination that an ad is relevant to the consumer). Even if we were inclined to consider whether this aspect of the settlement agreement goes beyond any legitimate claim of trademark infringement, and therefore imposes a restraint on competition not justified by the procompetitive value of enforcing trademark rights, the Commission has neither made separate findings with respect to the specific anticompetitive effects of this narrow aspect of the settlement agreements, nor urged that we should evaluate this issue separately. Accordingly, we have no reason to consider that issue.

government has failed to show that the proffered alternatives achieve the same legitimate procompetitive benefits as those advanced by the Petitioner.

CONCLUSION

In this case, where the restrictions that arise are born of typical trademark settlement agreements, we cannot overlook the Challenged Agreements’ procompetitive goal of promoting trademark policy. In light of the strong procompetitive justification of protecting Petitioner’s trademarks, we conclude the Challenged Agreements “merely regulate[] and perhaps thereby promote[] competition.” Chicago Bd. of Trade, 246 U.S. at 238. They do not constitute a violation of the Sherman Act, and therefore an asserted violation of the FTC Act fails of necessity.

The petition for review is GRANTED, the Final Order of the Federal Trade Commission is VACATED, and the case is REMANDED with instructions to DISMISS the administrative complaint.