Viamedia, Incorporation v. Comcast Corporation

• Court: Court of Appeals for the Seventh Circuit
• Decided: February 24, 2020
• Docket: 18-2852
• Status: Published

Opinion

In the

United States Court of Appeals For the Seventh Circuit ____________________ No. 18-2852 VIAMEDIA, INC., Plaintiff-Appellant, v.

COMCAST CORPORATION and COMCAST CABLE COMMUNICATIONS MANAGEMENT, LLC, Defendants-Appellees. ____________________

Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 1:16-cv-05486 — Amy J. St. Eve, Judge. ____________________

ARGUED FEBRUARY 7, 2019 — DECIDED FEBRUARY 24, 2020 ____________________

Before BAUER, HAMILTON, and BRENNAN, Circuit Judges. Table of Contents I. The Markets and the Competitors ........................................7 A. Cable Television: History, Revenue Sources, and Competition .......................................................................8 1. Television Programming and Advertising ..............8 2 No. 18-2852

2. Revenue Sources: Competition and Cooperation 10 a. Competition for Advertising Dollars and Cooperation Through Interconnects ................10 b. Competition for Subscribers ..............................15 i. Growing MVPD Competition .....................16 ii. Incumbent Cable Companies’ Efforts to Stymie Competition for Subscribers ..........18 B. The Ad Rep Services Market .........................................20 1. The Role of Viamedia ...............................................20 2. Vertically Integrated MVPDs ..................................22 3. Back to the Interconnects .........................................23 C. Comcast Refuses Interconnect Access to Viamedia ...24 II. District Court Proceedings ..................................................35 III. Legal Standards and Analysis .............................................37 A. Sherman Act Section 2—Illegal Monopolization........38 B. Claims of Anticompetitive Conduct: Refusals to Deal and Tying ............................................................41 1. Refusals to Deal .........................................................44 a. Monopolists and Refusals to Deal ....................44 b. Aspen Skiing and Comcast ..................................46 c. Refusals to Deal and Motions to Dismiss ........54 i. Comcast’s Proposed Legal Standard..........57 ii. Inapposite Vertical Integration Cases ........63 No. 18-2852 3

2. Tying ...........................................................................67 a. Summary Judgment Standard ..........................69 b. Tying and Comcast’s Conduct ..........................70 i. Definition........................................................71 ii. Separate Products or Services .....................73 iii. Forced Purchase ............................................75 C. Section 2 Monopolization: Harms, Efficiencies, & Remedies ......................................................................84 1. Harm to Competition ...............................................84 a. Ad Rep Services...................................................85 b. The MVPD Market: MVPDs, Advertisers, Cable Subscribers ................................................85 c. Back to the Interconnects ...................................88 2. Procompetitive Justifications? ................................91 a. The Interconnects ................................................92 b. The Ad Rep Services Market .............................93 3. Remedies ....................................................................95 D. Antitrust Injury ...............................................................98 E. Role of Expert Witnesses ..............................................102 1. Standard ...................................................................103 2. Economic Expert Witness ......................................103 3. Lack of Expert Witness on Causation ..................104 Conclusion .................................................................................105 4 No. 18-2852

HAMILTON, Circuit Judge. Plaintiff Viamedia, Inc. has sued defendant Comcast Corporation for violating Section 2 of the Sherman Act, 15 U.S.C. § 2. Viamedia accuses Comcast of us- ing its monopoly power in one service market to exclude com- petition and gain monopoly power in another service market. The district court dismissed Viamedia’s case, in part on the pleadings and in part on summary judgment. We reverse. Vi- amedia’s allegations and evidence are sufficient to state and support claims that should be presented to a jury. Because the district court dismissed part of the case on the pleadings and the rest on summary judgment, we must treat as true Viamedia’s factual allegations and give it the benefit of factual disputes and favorable inferences from the evi- dence. To make sense of this case, we explain some basic busi- ness arrangements in the markets that put television pro- gramming in American homes, as well as market definitions necessary in evaluating the antitrust claims. The parties agree on the definitions of the relevant geo- graphic and service markets. Viamedia asserts claims against Comcast for monopolization in three geographic markets: the Chicago, Detroit, and Hartford metropolitan areas. In each of those three geographic markets Comcast now has monopoly power over two separate service markets: Interconnect ser- vices and advertising representation services. Interconnect services are cooperative selling arrangements for advertising through an “Interconnect” that enables providers of retail ca- ble television services to sell advertising targeted efficiently at regional audiences. Advertising representation services for retail cable television providers assist those providers with the sale and delivery of national, regional, and local advertis- ing slots. This market in advertising representation services is No. 18-2852 5

the one in which Viamedia competed with Comcast. In each geographic market, according to Viamedia’s evidence, Com- cast used its monopoly power over the cooperative Intercon- nects to force its smaller retail cable television competitors to stop doing business with Viamedia, thereby gaining monop- oly power over the market for advertising representation ser- vices. Viamedia has presented evidence that Comcast’s elimina- tion of its only competitor in the advertising representation services market has harmed competition in violation of Sec- tion 2. According to Viamedia’s evidence, its customers for advertising representation services (i.e., Comcast’s retail cable competitors) did not switch to Comcast because it offered a better-quality or lower-priced service. They switched because Comcast used its monopoly power over the Interconnects to present its cable competitors with a Hobson’s choice: either start buying advertising representation services from us and regain access to the Interconnects, or keep buying those ser- vices from Viamedia and stay cut off from the Interconnects they needed to compete effectively. According to Viamedia’s evidence, Comcast deliberately adopted a strategy it knew would cost Comcast itself millions of dollars in the short run, but the strategy eventually gave it monopoly power in these local markets for advertising representation services. Giving Viamedia the benefit of its allegations and evidence, this is not a case in which Section 2 is being misused to protect weaker competitors rather than competition more generally. See Lee- gin Creative Leather Products, Inc. v. PSKS, Inc., 551 U.S. 877, 906 (2007), quoting Atlantic Richfield Co. v. USA Petroleum Co., 495 U.S. 328, 338 (1990) (purpose of the antitrust laws is to protect “competition, not competitors”). 6 No. 18-2852

As now the sole provider of advertising representation services to its cable competitors, Comcast can also damage competition beyond the relatively narrow markets for adver- tising representation services in Chicago, Detroit, and Hart- ford. This control allows it to undercut competition in two more markets: cable television services to retail customers, and the sale of advertising spots to local retailers. By estab- lishing itself as the gatekeeper for its cable competitors’ ad- vertising, Comcast has gained access to their sensitive mar- keting and promotional pricing information. And because Comcast took control of its rival cable companies’ inventory of local ads, local retailers no longer have a choice of cable companies from whom they buy ad time.1 Viamedia has thus offered evidence to defeat summary judgment on its claim that Comcast unlawfully used its mo- nopoly power over the Interconnects to tie those services to its advertising representation services. Viamedia has also ad- equately stated a claim that Comcast has unlawfully refused to deal with Viamedia and any cable competitor that bought advertising representation services from Viamedia. On the pleadings and the summary judgment record, Viamedia’s prima facie claims of monopolization are similar to but stronger than the successful plaintiff’s Section 2 claim in As- pen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585

1 Comcast sells advertising representation services through an entity called Comcast Cable Communications Management, LLC, which was formerly called Comcast Spotlight. The district court and the parties have referred to Comcast’s ad-related services division as both Comcast and Comcast Spotlight. We use “Comcast” to refer to both together but make clear when we refer to Comcast Spotlight in particular. No. 18-2852 7

(1985). We remand this case for any further necessary discov- ery and for trial. In Part I, we lay out the key facts: in Part I-A, the structure of the cable television markets; in Part I-B, the specifics of Comcast’s and Viamedia’s businesses, including the advertis- ing representation services they both offer and the critical role that Interconnects play for providers of cable television pro- gramming; and in Part I-C, Comcast’s refusal to continue providing Interconnect access to Viamedia or any of its cus- tomers in Chicago, Detroit, and Hartford. In Part II, we review the district court proceedings. Then, in Part III-A, we lay out the legal standards under Section 2 that apply to Viamedia’s claims. In Part III-B-1, we apply that law to Comcast’s decision to refuse to allow Viamedia or its customers access to the In- terconnects. In Part III-B-2, we apply that law to Viamedia’s claim that Comcast illegally tied Interconnect services to ad- vertising representation services. In Part III-C, we evaluate in greater detail the harm to competition alleged by Viamedia and the procompetitive justifications offered by Comcast, highlighting considerations that will be relevant on remand. Finally, in Parts III-D and III-E, we address issues of antitrust injury and the district court’s exclusion of expert witnesses. I. The Markets and the Competitors Because the district court dismissed one claim on the pleadings and the other on summary judgment, we present the relevant allegations and evidence in the light reasonably most favorable to plaintiff Viamedia, the non-moving party. The parties agree on the definition of the relevant geographic markets, and the relevant service-product markets are not dis- puted on appeal. The relevant geographic markets are the Chicago, Detroit, and Hartford metropolitan areas, called 8 No. 18-2852

Direct Marketing Areas or DMAs. The monopolized service market in each metropolitan area is that for the sale of adver- tising representation services (“ad rep services” in industry terms) related to so-called spot advertising on cable systems. To assess the harm to competition that can result from mo- nopolization of the market for ad rep services, we must ex- plain the related markets for retail cable television services to consumers, as well as access to the cable companies’ coopera- tive advertising distribution platforms called Interconnects. A. Cable Television: History, Revenue Sources, and Competi- tion Understanding these markets’ competitive dynamics re- quires a bit of history about the evolution of television in the United States, including the challenges that cable companies have faced in competing with over-the-air broadcast pro- gramming. 1. Television Programming and Advertising An awkward acronym, MVPDs, stands for “multichannel video programming distributors.” That umbrella term in- cludes cable companies like Comcast and Cox, as well as “overbuilders” like RCN and Wide Open West, known as WOW!. Beyond cable companies, MVPDs also include direct- broadcast satellite companies (AT&T’s DirecTV and Dish Net- work), as well as companies formerly associated only with tel- ephone service (e.g., Verizon’s FiOS and AT&T’s U-verse). The No. 18-2852 9

two largest MVPDs are Comcast and AT&T-DirecTV, which together have more than 20 million television subscribers.2 Focusing on the advertising-supported network program- ming carried by these MVPDs, we explain the special obsta- cles cable companies face in taking advantage of advertising revenue. The cable companies’ solution—jointly developing the Interconnects—created a later opening for a dominant ca- ble company like Comcast to use its power over several Inter- connects to gain a monopoly in a related market and to gain some measure of oversight and control of its smaller cable competitors. For decades, television programming was dominated by three broadcast networks and was funded largely by the ad- vertisements that ran in the programming. To help advertis- ers know how many and which viewers they were reaching, the industry adopted various audience measurement metrics, most importantly “Designated Market Areas” or “DMAs.” DMAs are meant to capture regional audiences that are likely to view the same programming. They often encompass more than a single county and can also cross state lines. As cable television companies got started, they typically won exclusive franchise areas granted by local governments. Their further expansion was then “subsidized by monopoly profits” from these exclusive territories. U.S. Dep’t of Justice, Voice, Video and Broadband: The Changing Competitive Landscape and Its Impact on Consumers 71 (Nov. 2008) (DOJ Report). Ca- ble companies then grew and consolidated by sewing

2 For statutory definitions of “multichannel video programming dis- tributor” (MVPD), “cable service,” and “video programming,” see 47 U.S.C. § 522(13), (6), and (20). 10 No. 18-2852

together such local franchises. Critically for this case, the patchwork combinations of local franchise areas did not align with DMAs. 2. Revenue Sources: Competition and Cooperation Most revenue for MVPDs comes from (1) the sale of adver- tising and (2) customer subscription fees. In geographic areas where MVPDs overlap, they compete on both fronts. The con- duct at issue in this litigation affects both fronts and millions of households in the key metropolitan areas, and it potentially affects tens of millions more in other metropolitan areas. a. Competition for Advertising Dollars and Coopera- tion Through Interconnects The mechanics of advertising are central to this lawsuit, accounting for the existence of the market in which Comcast and Viamedia competed. For every hour of programming, networks allot a certain number of minutes for advertise- ments. Contracts between an MVPD and a network (e.g., CNN or ESPN) typically make two or three of those minutes per hour available for the MVPDs to sell themselves, with the networks selling the remainder. MVPDs can sell these time slots to advertisers in various increments of time, such as 15, 30, or 60 seconds. Each increment is typically referred to as a “spot cable availability,” or “spot avail.” Approximately 75% of the spot avails are sold to advertisers. The MVPDs use the remaining 25% to advertise their own products and services. This brings us to the source of the problem here. In the early days of cable, advertisers who wanted to reach an entire DMA such as Chicago faced an obstacle. Cable systems had grown organically, with each cable service obtaining fran- chises “through the simple addition of new systems as No. 18-2852 11

opportunities arose,” leaving cable “company holdings … typically scattered across the country.” Patrick R. Parsons, Horizontal Integration in the Cable Television Industry: History and Context, 16 The Journal of Media Economics, no. 1 (2003) at 23, 37. Most important, the DMAs that helped broadcast television advertising reach entire marketing areas did not align with cable companies’ franchise areas. And yet, for a substantial percentage of spot avails, it would be most profit- able to sell them on a regional, DMA-wide level. As a result, cable companies had a weaker competitive po- sition for advertising dollars vis-à-vis the broadcast networks and satellite providers, who could easily deliver advertising to an entire DMA. Cable companies could not offer DMA- wide coverage, so advertisers would pay less for spot avails. To ensure DMA-wide coverage, an advertiser had to contract separately with each cable provider whose footprint included any part of that DMA. This was inefficient. The cable companies came up with a solution. They banded together to create a platform called an Interconnect that could bring together all cable providers within a given DMA. The cable companies could contribute their DMA-wide spot avails to the Interconnect, which would provide a single point of contact for advertisers. An advertiser could then pur- chase a particular time slot and be assured that its advertise- ment would appear in cable subscribers’ programming throughout the DMA. Thus, Interconnect services are pro- vided DMA by DMA. As described by Comcast, “intercon- nects were formed voluntarily by MVPDs in markets to pool their resources and offer DMA-wide selling of cable/MVPD advertising inventory[.] … Otherwise, advertisers trying to 12 No. 18-2852

cobble together a wide-footprint, MVPD-based advertising campaign would have to go MVPD-by-MVPD.” To cable subscribers, the national, regional, and local ad- vertisements appear seamlessly within television shows and live sports events. But the hidden seam of the Interconnects— and the ways its spot avails are paid for and delivered—is the locus of Comcast’s allegedly anticompetitive conduct. For purposes of this suit, the services provided by the In- terconnects must be distinguished from advertising represen- tation services. An Interconnect operator will:  Pool inventory of spot avails from multiple MVPDs on a DMA-wide basis;  Employ sales personnel to sell and/or coordinate sales of DMA-wide spot avails;  Distribute schedules of participating MVPDs’ spot avails to facilitate coordinated merging of local ad- vertising schedules;  Coordinate insertion of ads (although MVPDs themselves generally provide the technical equip- ment for ad insertion into programming); and  Collect money from Interconnect advertisers and coordinate payment to participating MVPDs or their ad representatives. Thus, the Interconnects allow the participating MVPDs to sell their DMA-wide advertising in a way that mimics the No. 18-2852 13

broadcast networks’ and satellite providers’ comprehensive coverage of a DMA.3 All participating MVPDs were intended to benefit from the Interconnects, and all were encouraged to participate to maximize the value of the DMA-wide spot avails. To quote Comcast again: “The value of an interconnect increases as more MVPDs in an area participate, so our incentive is to have as many MVPDs participate as possible.” First Am. Cplt. ¶ 39. Because MVPDs will contribute only about one-third of their spot avails to the Interconnects, and compete with one an- other for local ad sales, Interconnect participants took steps to avoid giving preferential treatment to any single MVPD par- ticipant. To ensure fair administration of the Interconnects, they were initially conceived as being operated by non- MVPD, neutral third parties. At the time of their formation, Interconnects were overseen by boards of directors elected by all MVPD members. Interconnects thus became valuable bridges to advertisers, translating into millions of dollars of advertising revenue each year in each market. Interconnects are especially valua- ble to smaller MVPDs. Once an Interconnect gains a critical mass of subscribers, regional or national advertisers are less likely to bother dealing with standalone MVPDs, especially those with small shares of DMA subscribers. And selling spot avails only to local (as opposed to DMA-wide or national) ad- vertisers will not compensate for the lost revenue if an MVPD

3 This type of cooperative arrangement is also available at the national

level, with National Cable Communications (NCC) able to place ads across multiple DMAs or nationwide, replicating broadcast networks’ and satellite providers’ nationwide coverage. As the country’s largest cable provider, Comcast now controls 60% of the NCC. 14 No. 18-2852

is shut out of the Interconnect. Purely local spot avails are sources of revenue and local business relationships, but they have lower profit margins. An Interconnect is what economists call a “two-sided plat- form.” It serves as a clearinghouse, offering “different prod- ucts or services to two different groups who both depend on the platform to intermediate between them.” Ohio v. American Express Co., 138 S. Ct. 2274, 2280 (2018). Such connectivity gives an Interconnect its value but can also be misused to harm competition. On one side of the Interconnect are the advertisers, who are interested in reaching the greatest number of viewers, es- pecially within a targeted DMA. The more subscribers an MVPD can bring to the table, the more advertisers will pay to reach that expanded audience. On the other side of the Inter- connect are the MVPDs and their retail customers. The more advertisers that participate, the more valuable the Intercon- nect is to the MVPDs and their customers. Cable customers watching a ballgame or their favorite comedy may not think about the value of the advertisements they see, but MVPDs can use advertising revenue to keep monthly subscription prices lower and to run promotional discounts to bring in even more subscribers. Those new subscribers will in turn make the MVPD a more valuable and attractive advertising venue. The Interconnect can thus produce a competitively vir- tuous feedback loop. “[T]he value of the services that [an In- terconnect] provides increases as the number of participants on both sides of the [Interconnect] increases.” See American Express, 138 S. Ct. at 2280–81. Or, as Comcast puts it: “The value of an interconnect increases as more MVPDs in an area No. 18-2852 15

participate, so our incentive is to have as many MVPDs par- ticipate as possible.” An Interconnect is not necessarily, however, a one-way ratchet to increased demand. Decreased participation on ei- ther side of the Interconnect can also reduce its value. Thus, adapting language from American Express, an Interconnect “losing participation on [the cable provider side] decreases the value” of the advertiser side, and if advertisers “leave due to this loss in value, then the [Interconnect] has even less value to [the cable providers]—risking a feedback loop of declining demand.” 138 S. Ct. at 2281. Whether the Interconnects are procompetitive or not de- pends on the competitive dynamics among its participants. In a competitive market, for example, the risk of negative feed- back may serve as a check on the ability of any one participant to raise prices or otherwise exert market power. See 138 S. Ct. at 2281 n.1. Conversely, in a less competitive market, access to the crucial Interconnects can be used to exclude competitors and harm competition. The Interconnects are so important that exclusionary conduct can become a weapon to injure competitors.4 b. Competition for Subscribers We have just outlined the ways in which MVPDs compete and cooperate in the pursuit of advertising revenue, which is

4 There is no challenge here to the legality of the Interconnects them- selves, at least as originally conceived, which seem to fit the model of cer- tain procompetitive cooperative arrangements among competitors. See generally Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U.S. 1 (1979). Whether that remains the case when one MVPD controls an Interconnect is a question not presented here. 16 No. 18-2852

the focus of Viamedia’s claims. Yet to see the full potential harm to competition caused by Comcast’s alleged conduct, we must also describe the MVPDs’ competition for subscrib- ers. Comcast’s alleged conduct is all the more dangerous to competition because it was made possible by accelerating in- dustry consolidation and has the potential to interfere with MVPDs’ competition with one another. The industry dynam- ics provide important context to understand the exclusion of Viamedia from a handful of DMA Interconnects, at least ini- tially, and the broader potential impact on MVPD markets in general. i. Growing MVPD Competition Until the mid-1990s, cable companies typically operated as monopolists with exclusive local cable franchises in their re- spective areas. They showed little interest in building into one another’s franchise areas and forcing competition. A combi- nation of legal, regulatory, and practical barriers limited com- petitive entry by new MVPDs, and those limits were often supported by incumbent cable providers. DOJ Report at 32. Thus, only satellite companies DirecTV and Dish Network, with their nationwide coverage, could compete with cable companies for subscribers. Satellite companies were able to take some market share, particularly in rural areas, but their competitive threat to cable companies proved to be limited. DOJ Report at 5, 10, 22 & n.88, 59. The 1990s saw major changes in the MVPD landscape. The cable industry shifted “toward regional consolidation, with specific companies carving out large parts of the country within which to group their systems.” Patrick R. Parsons, Hor- izontal Integration in the Cable Television Industry: History and Context, 16 Journal of Media Economics, no. 1 (2003) at 23, 37. No. 18-2852 17

The larger companies “bought and traded individual sys- tems,” with “the various systems in a given city” increasingly “fall[ing] into the hands of a single cable company.” Id. With changes in technology and the regulatory environ- ment, however, cable companies were about to face new com- petitors. The Telecommunications Act of 1996 was intended to break down barriers among cable, telephone, satellite, and in- ternet businesses to galvanize competition—and it did. Larry Satkowiak, The Cable Industry: A Short History Through Three Generations 47–48 (The Cable Center 2015). The lines between MVPDs, traditional telephone companies, and new broad- band internet service providers became increasingly blurred as these companies started offering multiple services to con- sumers. DOJ Report at 1, 17, 19. Cable companies introduced telephone voice services, which had previously been a legal monopoly in many states, and started selling bundles of tele- phone, video, and broadband Internet access. Id. at 9, 11. Meanwhile, broadband internet service providers like RCN and WOW!, known as “overbuilders,” built their own infra- structure in areas already served by incumbent cable compa- nies and rolled out multiple services. Id. at 8 & n.33, 21 & n.78, 47.5 And traditional telephone service providers responded in

5 “The term ‘overbuild’ describes the situation in which a second cable

operator enters a local market in direct competition with an incumbent cable operator. In these markets, the second operator, or ‘overbuilder,’ lays wires in the same area as the incumbent, ‘overbuilding’ the incum- bent’s plant, thereby giving consumers a choice between cable service pro- viders.” Report and Order and Further Notice of Proposed Rulemaking at 15 n.97, In the Matter of Implementation of Section 621(a)(1) of the Cable Com- munications Policy Act of 1984 as amended by the Cable Television Consumer Protection and Competition Act of 1992, FCC MB Docket 05-311 (Mar. 5, 2007) (FCC 2007 Report and Order). 18 No. 18-2852

kind. Verizon introduced its FiOS service in 2005, and AT&T followed with its MVPD service dubbed “U-Verse.” Id. at 6–7. Thus, cable providers—formerly the beneficiaries of cable franchise monopolies—suddenly faced a new array of com- petitors. Today, many DMAs are served by an incumbent ca- ble provider (e.g., Comcast), one or more overbuilder cable providers (e.g., RCN and WOW!), one or more telephone companies offering video services (e.g., Verizon FIOS), and two satellite dish providers (DISH and AT&T-DirecTV). ii. Incumbent Cable Companies’ Efforts to Stymie Competition for Subscribers This new competition led to credible reports of lower prices and falling cable subscription rates in areas with new MVPD entrants—exactly what one would hope to see in com- petitive markets. DOJ Report at 38–39 & nn.180–83. Incum- bent cable companies were forced to “respond[] to new entry by improving customer service, increasing bandwidth speeds …, adding more programming channels and services, and rolling out enhanced products (such as HD).” Id. at 48; see also id. at 45–46. But new MVPD competitors continued to encoun- ter obstacles, including some put in place by the incumbent cable providers. The incumbents had strong incentives to try to stymie these new competitors. As the FCC noted, competi- tion from new cable companies reduces rates far more than competition from satellite companies. FCC 2007 Report and Order at 26 ¶ 50. Among the obstacles for new competitors relevant to this case, incumbent cable providers entered into exclusive con- tracts with apartment buildings dense with potential sub- scribers, which new entrants could not reach, and exclusive No. 18-2852 19

and discriminatory contracts with programmers, whose con- tent new entrants could not carry. FCC 2007 Report and Order at 18 ¶ 35. Overbuilders and the FCC reported in 2008 that “[e]xclusivity and discrimination in access to programming are the most powerful tactics that incumbent operators use in an effort to block or otherwise constrain [new] competition.” DOJ Report at 74; see also id. at 73, 75, 89; Petition of RCN Telecom Services, Inc., to Deny Applications or Condition Consent at 24, 27, In the Matter of Applications for Consent to the Transfer of Control of Licenses of Comcast Corporation and AT&T Corporation to AT&T Comcast Corp., FCC MB Docket 02-70 (Apr. 29, 2002) (RCN 2002 FCC Petition) (RCN recounting “the difficulties it has encountered in gaining, and keeping, access to critical, non-substitutable local programming con- trolled by Comcast” and the “numerous instances in which the incumbents (Comcast and its predecessors) have received exclusive building rights covering a period of years”). Incumbent providers also created barriers to signing up individual customers by locking existing subscribers into long-term contracts. Because of these long-term contracts “there is only a small window when a customer is able to move.” DOJ Report at 52. “The incumbent [cable provider] knows when that window is, but the new entrant does not.” Id. “The new entrant must spend resources marketing to cus- tomers during periods when they cannot switch or will have disincentives to doing so,” while “an incumbent can target discounts and other incentives to subscribers immediately prior to the expiration of their contracts.” Id. at 52–53. This competitive dynamic helps explain why smaller cable compa- nies would hesitate to turn over their promotional advertising plans to their dominant cable competitor in advance of the 20 No. 18-2852

actual promotions—which is the likely result of Comcast’s ac- tions challenged in this case. B. The Ad Rep Services Market 1. The Role of Viamedia As MVPDs were trying to establish themselves in new markets, the sale of spot avails provided a key source of rev- enue that helped subsidize offers to attract subscribers. In- cumbent cable companies had been selling their spot avails to advertisers for decades, with the scale, internal structures, and sales and operational personnel to support those activi- ties. The new overbuilders and telephone service providers had no such experience or infrastructure. Enter Viamedia. The new MVPD competitors could have all spent money to hire their own advertising sales staffs, to buy and implement billing systems, to set up monitoring pro- tocols, and to deal with the necessary equipment to insert those ads seamlessly and accurately into programming. Many, including MVPDs in Chicago, Detroit, and Hartford, chose instead to contract for these spot advertising services with Viamedia. RCN, for example, could focus on competing with incumbent MVPDs through attracting subscribers and building out its footprint, with an assured ad revenue stream managed by Viamedia. With an Interconnect already in place, the new MVPDs (or Viamedia on their behalf) could sign an agreement with that Interconnect so that advertisers could place DMA-wide ads that reached the new entrants’ subscrib- ers along with the those of the incumbents. These ad rep services are at the core of this lawsuit. The ad rep services that Viamedia provides its customer MVPDs in- clude: No. 18-2852 21

 Allocating the MVPD’s inventory of spot avails among different sales channels—i.e., local ads, sold in competition with other MVPDs; DMA-wide ads; or multi-DMA/national ads;  Researching, marketing, pricing, and selling an MVPD’s inventory of spot avails to advertisers, in- cluding the approximately one-third of spot avails sold to local retailers in competition with other MVPDs;  Interfacing with the relevant Interconnect for spot avails allocated to regional, DMA-wide ads;  Providing technical services such as encoding video files and operating and maintaining the soft- ware needed to run, insert, traffic, monitor, and ar- chive ads;  Organizing the MVPD’s inventory of spot avails into schedules and ensuring that each ad runs cor- rectly during those schedules; and  Performing financial services, such as accounting, billing, and collection. Viamedia employs the personnel needed for these functions, spreading these costs among all of its MVPD customers. If an MVPD retained Viamedia to provide this full range of ser- vices for all of its inventory of spot avails, including the com- petitive selling of local spot avails, it was said that the MVPD had secured “full turnkey” representation. As overbuilders and telephone companies continued their build out, Viamedia was able to expand the areas and MVPD customers to which it could supply services. 22 No. 18-2852

2. Vertically Integrated MVPDs By contrast, Comcast does not need an independent ad rep services provider like Viamedia. Instead, Comcast is vertically integrated and has its own wholly-owned subsidiary that pro- vides ad rep services both in-house and to other competing MVPDs. In markets where Comcast does not operate the In- terconnects, its in-house ad rep services arm secures Intercon- nect access for its own MVPD service and its customer/com- petitors’ MVPD services, just as Viamedia used to do in Chi- cago, Detroit, and Hartford. Several other MVPDs have simi- lar internal divisions that provide spot cable ad rep services.6 In fact, Viamedia is unique in that it is the only ad rep services firm of any size that is independent—i.e., not owned by an MVPD. MVPDs that have their own ad rep services divisions or subsidiaries, such as Comcast, compete with Viamedia to pro- vide these services to other MVPDs. And just as MVPDs com- pete for subscribers wherever their service footprints overlap, the providers of ad rep services compete DMA by DMA. The ad rep services providers organize their sales forces around the boundaries of DMAs and provide services only to the MVPDs who have subscribers within those DMAs. Hybrid ar- rangements also exist. Some MVPDs do not contract for “full turnkey services,” but instead seek ad rep services for only a portion of their spot avails and sell the remaining spot avails themselves.

6 These include Charter’s Spectrum Reach, Cox’s Cox Media, Altice’s Suddenlink Media and Altice Media Solutions, and Mediacom’s OnMe- dia. No. 18-2852 23

This unusual market structure thus involves three levels of competition: (1) MVPDs compete against one another for subscribers; (2) some vertically integrated MVPDs’ ad rep ser- vices arms compete against Viamedia (and potentially against each other) for clients; and (3) MVPDs compete with one an- other for some sales of their spot avails to advertisers. We need to keep all three levels in mind. 3. Back to the Interconnects As part of the continuing industry consolidation in the 2000s, Comcast moved into many new DMAs. It also ex- panded from being one of several cable companies that par- ticipated in some DMA Interconnects to being the largest par- ticipant. For example, Comcast was able to acquire over 3,300 local cable franchising areas through its purchase of AT&T Broadband’s and Adelphia’s cable properties. FCC 2007 Re- port and Order at 15 & n.95. As Comcast repeatedly acquired other cable systems, it grew to be the largest participant in dozens of DMAs and became the sole “operator” or “man- ager” of those DMAs’ Interconnects, including in Chicago, Detroit, and Hartford. Yet the Interconnects continued to function as they had be- fore industry consolidation. They provided a single point of contact for distributing DMA-wide ads, as well as access to and collecting fees from all MVPDs (or their ad rep service providers) that participated in the Interconnects. If an MVPD did not participate in an Interconnect, an advertiser could not reach its subscribers, making an ad buy within the DMA less valuable for any remaining MVPD Interconnect participants. During this period, in 2003, Viamedia entered into agree- ments with Comcast for Interconnect access in the Chicago 24 No. 18-2852

and Detroit DMAs, which ran until May 2012. Viamedia sought this access because it provided ad rep services to cable overbuilders RCN (in Chicago) and WOW! (in Chicago and Detroit) under contracts that ran until 2014. As noted above, typical industry practice is for approximately one-third of an MVPD’s spot avail inventory to be sold on a DMA-wide basis. In line with that practice, Viamedia agreed to sell a portion of RCN’s and WOW!’s spot avail inventory on a DMA-wide ba- sis through the Comcast-controlled Interconnects. Viamedia sold the remaining portion of RCN’s and WOW!’s spot avails both nationally and—in competition with Comcast—locally. Comcast also agreed not to solicit Viamedia’s MVPD clients until four months before the Viamedia/MVPD contracts ex- pired, although the MVPDs remained free to contact Comcast. C. Comcast Refuses Interconnect Access to Viamedia This was the competitive landscape for Comcast’s conduct challenged in this lawsuit. Internal Comcast PowerPoint presentations explained that Comcast viewed its “Next phase” as “consolidat[ing the] core business” of ad rep ser- vices, and then “look[ing] at other businesses we can leverage (our technologies or platforms).” A212 n.68. There is evidence that Comcast saw the Interconnects as one such point of “lev- erage.” As noted, Viamedia’s Interconnect access agreements with Comcast for the Chicago and Detroit DMAs were due to ex- pire in 2012. Viamedia’s contracts for ad rep services with RCN and WOW! were extended until 2015 (RCN) and 2014 (WOW!). As 2012 neared, Comcast faced a choice. It could compete for RCN’s and WOW!’s ad rep services business the following year, as it already competed for RCN’s, WOW!’s, and other MVPDs’ business in many other DMAs. Or it could No. 18-2852 25

try to use its control over the Interconnects to shut out the competition for ad rep services. At first, it appeared that Com- cast would take the route of competition, contacting RCN and WOW! to express interest in selling them ad rep services. But Comcast then changed its strategy. It tried instead to take ad- vantage of its control over the Interconnects. Comcast notified Viamedia in December 2011 that it would refuse to permit Vi- amedia any further access to the Interconnects. In June 2012 Comcast executed on that notice. For the first time in any DMA since the Interconnects had been created, an Intercon- nect operator—Comcast—had cut off Interconnect access to an MVPD or an MVPD representative. Comcast executed this strategy in other DMAs, as well, similarly denying Viamedia access to Interconnects on behalf of Viamedia’s customer MVPDs (Comcast’s competitor MVPDs). For example, in the Hartford DMA, Comcast had previously provided full-turnkey service to AT&T’s MVPD. When Frontier acquired AT&T’s Hartford network in 2014, it had the option of assuming the Comcast contract. Frontier, however, had been unhappy with Comcast’s customer service when it used Comcast in other DMAs, so it switched from Comcast to Viamedia. Comcast then excluded Frontier’s spot avails from the Hartford Interconnect, resulting in millions of dollars in lost ad revenues for Frontier and Viamedia, as well as Comcast itself, and degrading the value of the Hartford In- terconnect.7 By contrast, in DMAs where the Interconnects

7 On these points, the record contains more precise numbers in docu- ments that have been under seal. Here and elsewhere in this opinion, we have used verbal descriptions rather than specific numbers for important information that has been submitted under seal. We are skeptical, how- ever, about the grounds for sealing much, if not all, of the evidence under seal. Simultaneously with this decision, we are issuing an order that 26 No. 18-2852

were controlled not by Comcast but by other large, incumbent cable companies such as TimeWarner Cable, access to the In- terconnects had not yet been pulled. Comcast then returned to Viamedia with a series of offers that would have required Viamedia to “assign” 100% of its customers’ spot avails to Comcast in exchange for a one-time “finder’s fee.” That was essentially an offer to pay Viamedia to exit the marketplace. At the end of July 2014, Comcast pro- vided a more detailed offer. It would have had the same effect as the first offer—a payment to Viamedia to stop providing spot cable ad rep services. In August 2014, Viamedia received the third iteration of Comcast’s offer, which at first appeared promising. When Viamedia received a detailed offer in writ- ing, however, it discovered that Comcast had added a provi- sion that would permit Comcast unilaterally, and on just four hours’ notice, to take any ad inventory from Viamedia and contribute it to the Interconnect. That uncertainty would have rendered those spot avails virtually worthless to advertisers. Along with these onerous terms, the revenue-share pro- posals appeared to be below market rate for Interconnect- only access, compared to both Viamedia’s prior agreement with Comcast and other Interconnect-only access agreements in any other comparable DMAs. In short, the agreements did not offer access to the Interconnect in a way that would allow Viamedia to provide ad rep services to its MVPD customers. Nor were these terms to be found in any other Interconnect-

unseals evidence we identified in an earlier order to show cause, and we are ordering the parties to show cause why any of the remaining sealed evidence, including that obtained from non-parties, should remain under seal at this time. No. 18-2852 27

only agreement employed by any Interconnect operator in any DMA. Moreover, during these “negotiations” in July 2014, the head of Comcast’s cable spot ad rep services division and his colleague, Hank Oster, expressed concern that Comcast COO Dave Watson was “wavering on why we won’t let Viamedia in the Interconnects.” Whatever second thoughts some within Comcast might have had, however, Comcast’s approach did not change. As Oster later candidly explained, by July 2014 Comcast already “had made the decision” to exclude Viame- dia from the Interconnects. Comcast was also telling WOW! that it would not allow it to return to the Interconnect with Viamedia as its ad rep, and that Comcast was taking that po- sition as part of its “strategic plan.” A230. Comcast urges us to infer—as a matter of law—that it was acting for procompetitive reasons. The evidence, though, can easily support the inference that Comcast was instead choos- ing to inflict financial pain on both its competitors and itself to gain monopoly power in the ad rep services market, which would also produce a new advantage over its retail cable com- petitors. By cutting Viamedia off from the Chicago and De- troit Interconnects in 2012, Comcast ensured that its competi- tor MVPDs’ spot avails could not be distributed through the Interconnects while they were represented by Viamedia un- der their existing contracts. This was an expensive decision for Comcast. As operator of the Interconnect, Comcast’s internal analysis of the “Reve- nue Impact” of its decision predicted that Comcast itself would lose $10.6 million in just the first six months after cut- ting off Viamedia’s (and thus RCN’s and WOW!’s) access, in- cluding $2.3 million in lost cash flow. A838, A787–88. The 28 No. 18-2852

evidence of actual effects is consistent with that prediction. In the years that RCN and WOW! were unable to access the In- terconnects (June 2012 through December 2015), they lost ap- proximately $27 million in ad revenue. Comcast itself lost $7 million in commissions. A248, A637, A648 (figs. 35, 46). More- over, Comcast’s own spot avails would have decreased in value because an advertiser could no longer reach all cable subscribers within the DMA through the Interconnect. But as an amicus supporting Comcast points out, Comcast could easily afford to sacrifice millions in Interconnect fees and lower ad revenue in order to inflict this harm on its MVPD competitors, advertisers, and Viamedia. As the domi- nant MVPD provider in markets across the country, this “temporary and localized lost revenue is small potatoes,” a mere “rounding error.” Brief for Washington Legal Founda- tion’s as Amicus Curiae Supporting Appellees at 22. Just so. With Comcast and Viamedia as the only two providers of ad rep services in the Chicago and Detroit DMAs, Comcast’s denial of Interconnect access to Viamedia left Comcast with an effective monopoly over both Interconnect services and ad rep services. The window of time between Viamedia’s fore- closure from Interconnect access and Comcast’s competitor MVPDs’ return to the market to seek bids for their ad rep ser- vices would be the time for Comcast “to overpower ViaMe- dia,” as a Comcast employee in the Detroit DMA explained. A217 (budget presentation). So the evidence supports an in- ference that Comcast willingly chose to inflict short-term fi- nancial losses on itself. Why? A reasonable explanation is that it did so because it could survive those losses (the “small po- tatoes” and “rounding error”) to obtain and use monopoly power in the ad rep services market. No. 18-2852 29

In this lawsuit Comcast has argued that RCN and WOW! chose it over Viamedia on the merits of its offered services. Comcast highlights, for example, testimony from an RCN representative that—after comparing Comcast’s offer with In- terconnect access to Viamedia’s offer without—“It was not in the end a very difficult decision to make.” DA688. But this an- swer presupposes that Comcast shutting its competitors out of the Interconnects could be a reasonable basis to treat RCN’s decision as uncoerced. There is evidence that Comcast did just that. Contrary to the assertions of the district court and our colleague who dissents in part, for example, a Comcast em- ployee working in the Chicago and Detroit DMAs explained that Comcast had adopted “a business practice” that “if an MVPD wants to get access to a Comcast controlled Intercon- nect, it has to hire Comcast as its sale representative.” A215. Viamedia’s evidence also supports a finding that WOW! and RCN did not go willingly into Comcast’s arms. Both pushed back against Comcast’s demands (or threats) that they either use it for their advertising services or face exclusion from the Interconnects if they stayed with Viamedia. A WOW! employee communicating with Comcast reported back to WOW! colleagues that Comcast was “maintaining their posi- tion that [WOW!] can be in the IC [Interconnect] but only if they [Comcast] rep us directly.” A215 n.81, A230 n.129. It is a factual question whether it was reasonable at the time for the smaller MVPDs to “understand [that] to be part of the inter- connect [they] would need to be with Comcast Spotlight,” Comcast’s ad rep services arm. DA687. Comcast’s competitor MVPDs immediately began losing money after Comcast excluded them from the Interconnects. They lamented that their reductions in cash flow were 30 No. 18-2852

“primarily due to the loss of the Comcast Interconnect reve- nues in Chicago and Detroit.” A233. Despite that pressure, though, the MVPDs continued to resist Comcast’s demands. As discussed above, at the time, Comcast was trying to buy TimeWarner Cable, a proposed deal that was under review by federal agencies. With a forum to share their ongoing experi- ences with Comcast, RCN (futilely) filed comments with the FCC, alerting regulators that “Comcast was not being truth- ful” when it said “RCN is free to join the Comcast-managed interconnects at any time,” because “Comcast will only allow RCN to join the interconnects if RCN employs Comcast Spot- light instead of Viamedia.” A215 n.81, A886–87. Viamedia also was not going quietly. Even though Com- cast had barred it from Interconnect access, Viamedia contin- ued to compete for RCN’s and WOW!’s business. Without In- terconnect access, their MVPD customers’ spot avails would not bring in nearly as much revenue, which left Viamedia’s bids’ proposed revenue shares at a substantial disadvantage compared to Comcast’s bids. A231–32. Nonetheless, Viame- dia’s bids caused consternation for Comcast. Internal Com- cast emails reflect executives’ disbelief. They called it “abso- lutely unbelievable” that Viamedia could make a remotely competitive bid without Interconnect access. A232. In contrast to Viamedia’s bids, Comcast bids touted the “exclusive” ben- efit of Interconnect access that WOW! would receive if it se- lected Comcast for ad rep services. Comcast said that the “generous” financial terms it offered included the “sizable an- nual guarantee” that would be attributable to “the oppor- tunity to add WOW! subscribers [back] to the important De- troit and Chicago Interconnects.” In other words, sign up with Comcast for ad rep services, and we will stop your bleeding— No. 18-2852 31

the bleeding that we have inflicted by barring you from the DMA Interconnects. Substantial evidence thus shows that Comcast’s MVPD competitors did not want to buy ad rep services from Com- cast. Their reluctance was not based on a short-sighted inabil- ity to see the procompetitive benefits of Comcast’s vertical in- tegration or what Comcast touts as “one-stop shopping.” Ra- ther, these MVPDs had economically rational reasons for seeking to avoid this entanglement with their dominant com- petitor, which would naturally have divided loyalties. In ad- dition, WOW! considered Viamedia to be “by far the best ad partner from a technical team to work with.” A560, DA 685. RCN testified that it would prefer to obtain ad rep services from an independent company like Viamedia rather than Comcast because, “all things being equal, even close to being equal,” it “had concerns about being a partner with a com- pany associated with our competitor.” A236 n.150. Thus, as Comcast’s MVPD competitors assessed the situa- tion, the possible outcomes all amounted to unfair wins for Comcast. Its actions could have resulted in three different out- comes, each of which would work to its benefit and harm its competitors. First, if Comcast succeeded in having its compet- itor MVPDs buy Comcast ad rep services, Comcast would gain the following benefits:  Comcast’s smaller MVPD rivals would now be con- tributing additional revenue toward their domi- nant competitor, Comcast;  The majority of spot avails that MVPDs had for- merly kept out of the Interconnects to allocate to, among other outlets, local ads (for which the 32 No. 18-2852

MVPDs compete against each other for sales) would now come under the control of their compet- itor Comcast—whose contracts required that Com- cast have “sole and exclusive control” over all spot avails;  Comcast’s competitor MVPDs would just have to trust that Comcast would make the best business decisions on behalf of its competitors when allocat- ing adds to the national, regional, and local sales markets. For example, the smaller MVPDs prefer to weight some of their ad sales to non-Interconnect local sales, which help the MVPDs with local busi- ness relationships that can lead to additional sales of services, such as providing business internet connectivity (sales for which they compete against Comcast); the Interconnect operator, on the other hand, prefers DMA-wide Interconnect ad sales for which it gets higher margins—an ad mix choice that Comcast would be free to make for its smaller MVPD competitors;  Comcast would be a single seller of advertisements in the local market, eliminating competition;  Comcast would not only have access to its compet- itor MVPDs’ ad sales information, but the MVPDs would have to provide Comcast with all of their own promotional ad materials to current and po- tential subscribers that they are attempting to retain or win away from Comcast, giving Comcast a chance to undercut them. That would be in addi- tion to other competitively sensitive information No. 18-2852 33

(e.g., number and location of its subscribers) that would need to be disclosed. Viamedia has offered evidence that what drove Comcast’s actions was this close relationship with competing MVPDs— not hypothesized economic efficiencies from ordinary vertical integration. For example, there are some DMAs where Com- cast controls the Interconnects, but the participating MVPDs do not have overlapping footprints with Comcast’s service ar- eas. In those DMAs, Comcast still offers Interconnect-only agreements on terms similar to the terms of the former Com- cast-Viamedia agreements for the Chicago and Detroit DMAs. Viamedia offered evidence on summary judgment (de- scribed above) of a second outcome in which Comcast’s MVPD competitors would forgo Interconnect access entirely and renew with Viamedia rather than switch to Comcast. If its MVPD competitors made that choice, those MVPDs would be cut off from a large percentage of ad revenue, which in turn would hinder them from funding promotional offers to their subscribers, potentially leading their subscribers to switch to Comcast. In the meantime, Comcast’s lost millions from Inter- connect fees and reduced advertising revenue within the DMA would continue to be a mere “rounding error.” In the third potential outcome—which only Comcast con- tends was actually a possibility—Comcast would not bar competing MVPDs completely from Interconnect access but instead would permit them to have Interconnect-only access if they took care of their own ad services, without using either Viamedia’s or Comcast’s ad rep services. In this scenario, Comcast would lose the revenue it would have gained from providing full-turnkey service to the MVPDs, but it would still earn Interconnect access fees and the Interconnect’s value 34 No. 18-2852

would not be degraded. The result would also raise rivals’ costs by forcing them to provide internally the staff, technol- ogy, and services that Viamedia had previously provided at lower cost. Those fixed costs would be difficult for those MVPDs to afford and would shift revenue away from sub- scriber promotions and further infrastructure build-out. See FCC 2007 Report and Order at 8 ¶ 13 (“Revenues from cable services are, in fact, a driver for broadband deployment,” i.e., the build-out of additional cable infrastructure).8 Faced with this Hobson’s choice, Comcast’s competitor MVPDs chose to sign with Comcast in 2015. WOW! noted that “a key decision point” in this “choice” was its understanding that “in order to remain competitive, we need to be in the In- terconnect.” A233. WOW! signed with Comcast for ad rep ser- vices in Chicago and Detroit in 2015. WOW! continued, how- ever, to use Viamedia as its ad rep in some non-Comcast DMAs. Similarly, although RCN had planned to renew its contract with Viamedia, it too ultimately decided to sign with Comcast for Chicago and Detroit. By 2016, a Comcast em- ployee congratulated a colleague regarding its new monopoly in ad rep services in the Chicago DMA: “THE WOW AND RCN DEALS PROVIDE [COMCAST] WITH COMPLETE

8 Viamedia has presented evidence that RCN and WOW! did not view

bringing ad rep services in-house as a viable option. When confronted with Comcast’s refusal to deal with Viamedia, both said they had no choice but to enter into ad rep agreements with Comcast. A215, A887. The in-house option was always available in theory. But RCN and WOW! are presumed to be economically rational actors. They had always chosen to buy these services from outside companies, suggesting that in-house was not an economically viable option. No. 18-2852 35

REPRESENTATION OF THE CHICAGO MARKET.” A214 n.77. II. District Court Proceedings In 2016, Viamedia sued Comcast for violating Section 2 of the Sherman Antitrust Act, as well as various state antitrust statutes, and for tortious interference. The parties agree that analysis under the state antitrust statutes tracks federal anti- trust law, so the federal antitrust analysis controls whether the state antitrust claims survive. Viamedia is no longer press- ing its tortious interference claim, so our only focus is Section 2 of the Sherman Act. Comcast moved to dismiss the complaint for failing to state a claim. The district court construed Viamedia’s com- plaint as alleging that Comcast engaged in three types of mo- nopolistic conduct recognized by the antitrust laws: (1) Com- cast’s refusal to deal with Viamedia by cutting off access to the Interconnects, (2) Comcast’s exclusive dealing, and (3) Comcast’s tying of Interconnect access to the purchase of Comcast’s ad rep services. The district court granted the motion to dismiss with re- spect to Viamedia’s refusal-to-deal claim, faulting Viamedia for failing to demonstrate through its allegations that Com- cast’s conduct was “irrational but for its anticompetitive ef- fects.” Viamedia, Inc. v. Comcast Corp., 218 F. Supp. 3d 674, 698 (N.D. Ill. 2016). The court hypothesized that Comcast’s com- plete foreclosure of Viamedia from the market potentially serves a procompetitive purpose and “offers potentially im- proved efficiency.” 218 F. Supp. 3d at 699. And because “ver- tical integration is usually procompetitive,” Comcast likely had “a rational procompetitive purpose: it has become ‘a one- 36 No. 18-2852

stop shop’ in certain DMAs for MVPDs wishing to sell adver- tisements on a regional basis.” Id. at 698–99. Comcast’s “short- term losses” in excluding Viamedia and Comcast’s competi- tor MVPDs from the Interconnects were not “necessarily in- dicative of anticompetitive conduct,” the court reasoned in dismissing a later version of the complaint on identical grounds, because a “monopolist might wish to withdraw from a prior course of dealing … in order to pursue perfectly competitive ends.” Viamedia, Inc. v. Comcast Corp., No. 16-CV- 5486, 2017 WL 698681, at *4 (N.D. Ill. Feb. 22, 2017), quoting Novell, Inc. v. Microsoft Corp., 731 F.3d 1064, 1075 (10th Cir. 2013). After discovery on Viamedia’s exclusive dealing and tying claims, the district court granted Comcast’s motion for sum- mary judgment because Viamedia had failed to “present evi- dence that tends to exclude the possibility that [defendant’s] conduct was as consistent with competition as with illegal conduct.” Viamedia, Inc. v. Comcast Corp., 335 F. Supp. 3d 1036, 1054 (N.D. Ill. 2018). In the district court’s view, there was no evidence that Comcast conditioned its sale of Interconnect services to MVPDs on their purchase of ad rep services be- cause in DMAs outside the relevant geographic markets, Comcast did offer Interconnect-only access to other MVPDs. Id. at 1058–59. Further, because “both RCN and WOW! wanted full-turnkey representation,” the purchase of the two prod- ucts together could not be considered tying. Id. at 1059. In fact, the district court concluded (in tension with the observation that Comcast offered Interconnect-only access in other DMAs), Comcast had “no reason to offer” Interconnect-only access to RCN and WOW! because, the district court again hy- pothesized, an Interconnect-only deal would be less No. 18-2852 37

substantial, less profitable, and less efficient for Comcast than a full-turnkey deal. Id. at 1059. In any event, even if Comcast had refused to deal with its competitor MVPDs unless they met Comcast’s condition of purchasing ad rep services, the court viewed this as simply a reformulation of Viamedia’s already-dismissed refusal-to- deal claim. 335 F. Supp. 3d at 1062, 1070, 1072. To support this conclusion, the district court noted that Viamedia sought in- junctive relief that would restore its access to the Intercon- nects and give RCN and WOW! the option to decline pur- chase of Comcast’s ad rep services. Id. at 1074. In addition, the district court found that Viamedia’s two experts’ testimony was based upon an incorrect understand- ing of the law—i.e., that Comcast had engaged in anticompet- itive conduct for which it could be held responsible under the antitrust laws. Id. at 1064–74. The district court concluded that this justified excluding the damages expert’s testimony in its entirety, as well as a portion of the economic expert’s testi- mony. Viamedia has appealed the final judgment dismissing its claims. III. Legal Standards and Analysis With the facts and competitive dynamics set out, we turn to the legal standards and analysis. In Part III-A, we describe the standards for an antitrust violation under Section 2, the monopolization provision of the Sherman Act. Undisputed by the parties, we explain that Comcast is a monopolist in the relevant geographic markets (here: Chicago, Detroit, and Hartford) for both Interconnect and ad rep services, and that it is the dominant MVPD retail cable provider. 38 No. 18-2852

In Part III-B, we address Viamedia’s two claims. In Part III- B-1, we set out the legal test for refusals to deal and assess Comcast’s conduct, explaining why this claim should not have been dismissed on the pleadings. In Part III-B-2, we turn to Viamedia’s tying claim. These related claims are both based on the same course of conduct, resulted in the same anticom- petitive harms, and would be subject to the same procompet- itive justifications or defenses. The decision to dismiss one claim on the pleadings while allowing the other, closely re- lated claim to go as far as summary judgment offered poten- tial for confusion, but in the end, both claims need to be tried. In Part III-C, we evaluate in greater detail the harm to com- petition alleged by Viamedia and the procompetitive justifi- cations offered by Comcast, highlighting considerations that will be relevant on remand. In Part III-D, we explain that Vi- amedia has presented evidence of a cognizable antitrust in- jury as a rival driven from the market by a tying arrangement. Finally, in Part III-E, we address the district court’s rulings excluding expert witness evidence. Our resolution of Viame- dia’s refusal-to-deal and tying claims largely resolves its chal- lenge to the testimony’s exclusion. The district court’s deci- sion on this score was based almost entirely upon its errone- ous legal analysis. On remand, the district court will need to take a fresh look at the expert reports in light of this opinion. A. Sherman Act Section 2—Illegal Monopolization Section 2 of the Sherman Antitrust Act imposes liability on “Every person who shall monopolize … any part of the trade or commerce among the several States.” 15 U.S.C. § 2. A pri- vate plaintiff like Viamedia may bring a civil claim as a person No. 18-2852 39

who was “injured in his business or property by reason of an- ything forbidden in the antitrust laws.” 15 U.S.C. § 15(a). Judicial decisions interpreting Section 2 have long held that simple possession of monopoly power, or the pursuit of it, is not in itself illegal. United States v. U.S. Steel Corp., 251 U.S. 417, 451 (1920) (“[T]he law does not make mere size an offence, or the existence of unexerted power an offense. It … requires overt acts.”); United States v. Aluminum Co. of America, 148 F.2d 416, 429–30 (2d Cir. 1945) (“size does not determine guilt” as the monopolist may have gained market power “by force of accident,” or “by virtue of his superior skill, foresight and industry”; therefore, “there must be some ‘exclusion’ of competitors”). Thus, a firm violates the monopoly provision in Section 2 only when it both (1) possesses “monopoly power in the relevant market” and (2) engages in “the willful acqui- sition or maintenance of that power as distinguished from growth or development as a consequence of a superior prod- uct, business acumen, or historic accident.” Verizon Communi- cations, Inc. v. Law Offices of Curtis V. Trinko, LLP, 540 U.S. 398, 407 (2004), quoting United States v. Grinnell Corp., 384 U.S. 563, 570–71 (1966); see also Phillip E. Areeda & Herbert Hovenkamp, Antitrust Law: An Analysis of Antitrust Principles and Their Application ¶ 600a, at 3, ¶ 650a, at 91 (4th ed. 2015) (Areeda & Hovenkamp).9 On appeal, the parties do not dispute several often-conten- tious issues in antitrust cases: the relevant geographic and product markets, and market power. See Goldwasser v. Ameritech Corp., 222 F.3d 390, 397 (7th Cir. 2000) (“Few would

9 This appeal does not present any issues under the Section 2 language barring attempts and conspiracies to monopolize. 40 No. 18-2852

say that the first element is easily proved: it is exceedingly dif- ficult to prove market power, or monopoly power … .”). The relevant geographic markets are the specific DMAs in which Viamedia asserts Comcast’s conduct harmed competition: Chicago, Detroit, and Hartford. See Tampa Elec. Co. v. Nashville Coal Co., 365 U.S. 320, 332–33 (1961) (a relevant antitrust geo- graphic market is the area in which sellers operate and where purchasers can predictably turn for supplies). Comcast’s con- duct in other DMAs may be relevant for comparison pur- poses. The relevant product market allegedly monopolized is ad rep services for MVPDs. The immediate effect of Comcast’s conduct was to force out its only competitor in that market to gain monopoly power in the relevant geographic markets for those services. This market is inextricably connected to access to the cooperative mechanism of the Interconnects, as well as to the related markets for MVPD retail cable services and the sale of MVPD spot avails. Understanding the harm to compe- tition in these related markets helps in assessing Comcast’s alleged conduct. “Antitrust analysis must always be attuned to the particular structure and circumstances of the industry at issue.” Trinko, 540 U.S. at 411; see also Areeda & Hovenkamp ¶ 1802d, at 79–80 (When assessing exclusionary conduct, it is “necessary to examine market power or share at both of the two market levels involved.”). As for market power, in the Chicago, Detroit, and Hart- ford markets, Comcast started with monopoly control over Interconnect access and services. Comcast has acquired a pure monopoly in the market for ad rep services in these No. 18-2852 41

metropolitan areas, where it is also by far the dominant MVPD retail cable provider.10 A firm’s market power is important because, without it, a firm will have little to no ability to distort or harm competi- tion, no matter how great its desire to do so, even when en- gaging in conduct that in different circumstances might be perceived as anticompetitive. See Novell, Inc. v. Microsoft Corp., 731 F.3d 1064, 1071 (10th Cir. 2013) (Gorsuch, J.) (“Not infre- quently, the initial question of market power proves deci- sive.”). Even “[m]ildly reprehensible behavior might be enough to challenge a firm whose power is significant.” Areeda & Hovenkamp ¶ 600b, at 4. B. Claims of Anticompetitive Conduct: Refusals to Deal and Tying The dispute here focuses on the second prong of the Sec- tion 2 test: did Comcast “willfully acquire” or “maintain” its new monopoly power in the ad rep services market, or is its new market dominance “a consequence of a superior product or business acumen” or the result of an “historic accident”? Baked into this inquiry is an assessment of what types of anticompetitive conduct are prohibited as illegally acquiring or maintaining monopoly power, rather than the kind of pro- competitive conduct the antitrust laws do not impede. The lat- ter includes innovation resulting in superior products, the in- troduction of efficiencies reflecting superior business acumen, or even the luck of a firm that unwittingly stumbles into a

10 For antitrust purposes “monopoly power and market power typi- cally are used interchangeably” and simply mean that “a firm can influ- ence the price it receives for its product.” Dennis W. Carlton and Jeffery M. Perloff, Modern Industrial Organization 137 (2d ed. 1994). 42 No. 18-2852

monopoly position. See Areeda & Hovenkamp ¶ 600a, at 3 (setting out first two prongs identified above and articulating the two “subsidiary questions”: (3) “given that § 2 requires some element of conduct in addition to substantial market power, what kinds of conduct or intent transform power into unlawful monopolization; and (4) what defenses, if any, save monopoly power from condemnation?”). The statutory text does not provide the answers, but case law over more than a century provides extensive guidance. Courts recognize various types of conduct that have the po- tential to harm competition. The types of conduct alleged in this case are “exclusionary” in nature, impairing rivals’ op- portunity to compete in a way that is inconsistent with “com- petition on the merits.” Areeda & Hovenkamp ¶ 650a, at 92; see also id. ¶ 651b, at 99–100; Covad Communications Co. v. Bell Atlantic Corp., 398 F.3d 666, 675–76 (D.C. Cir. 2005) (agreeing with plaintiff’s description of defendant’s refusal to deal as “‘predatory’ … because, in the vernacular of antitrust law, a ‘predatory’ practice is one in which a firm sacrifices short- term profits in order to drive out of the market or otherwise discipline a competitor”). In the present case, Viamedia alleges and has offered evi- dence that Comcast: (1) refused to deal with Viamedia by denying it Interconnect access, and (2) engaged in tying by denying MVPDs Interconnect access unless they purchased Comcast’s ad rep services.11 We set out below the legal tests for refusals to deal and tying, which help in assessing whether

11 To “simplify the issues” on appeal, Viamedia elected not to pursue exclusive dealing as a distinct theory of liability. No. 18-2852 43

such conduct is anticompetitive and illegal, or instead harm- less or even procompetitive. Conduct that can harm competition may fit into more than one of these court-devised categories. After all, the “means of illicit exclusion, like the means of legitimate competition, are myriad.” Trinko, 540 U.S. at 398, quoting United States v. Mi- crosoft Corp., 253 F.3d 34, 58 (D.C. Cir. 2001) (en banc). Al- though “the standard for a § 2 violation is significantly stricter in its power assessment [than for a § 1 claim], it is broader and less categorical in its definition of proscribed conduct.” Areeda & Hovenkamp ¶ 777a, at 324. This means that a dom- inant firm’s conduct may be susceptible to more than one court-defined category of anticompetitive conduct. A “simple refusal to deal” is conduct where one firm “refuses to deal no matter what,” whereas “[t]ying and exclusive dealing are two common examples” of “conditional refusals to deal”—i.e., one firm will refuse to deal with another firm unless “some condition is met.” Herbert Hovenkamp, FRAND and Anti- trust, Cornell L. Rev. (forthcoming 2020) (manuscript at 11), available at http://papers.ssrn.com/sol3/papers.cfm?ab- stract_id=3420925. Similarly, “[m]any of the practices that have been characterized as exclusive dealing could also be de- scribed as tying” because “[t]he economic distinction between the two is most often slight or nil.” Areeda & Hovenkamp ¶ 1800b, at 7–8, ¶ 1800a, at 4. The fact that the categories of conduct here are conceptu- ally related and may overlap should not cause confusion if we stay focused on the underlying inquiry: the conduct “must harm the competitive process and thereby harm consumers.” Microsoft, 253 F.3d at 58; see also Nynex Corp. v. Discon, Inc., 525 U.S. 128, 135 (1998) (plaintiffs “must allege and prove 44 No. 18-2852

harm, not just to a single competitor, but to the competitive process, i.e., to competition itself”). At bottom, the purpose of identifying these categories of conduct is to help determine “the presence or absence of harmful effects, which are both the reason for any antitrust concern and often the simplest el- ement to disprove.” Areeda & Hovenkamp ¶ 1701d, at 33. We therefore start by assessing how Comcast’s conduct fits into these categories under Section 2, mindful that we should stay focused on the effect Comcast’s conduct has on competition. 1. Refusals to Deal The district court dismissed on the pleadings the portion of Viamedia’s complaint focused on a refusal-to-deal theory. We review de novo a grant of a motion to dismiss, “con- stru[ing] the complaint in the light most favorable to the plaintiff, accepting as true all well-pleaded facts alleged, and drawing all possible inferences in [its] favor.” Tamayo v. Blago- jevich, 526 F.3d 1074, 1081 (7th Cir. 2008); see also Goldberg v. United States, 881 F.3d 529, 531 (7th Cir. 2018) (in reviewing dismissal for failure to state a claim, we accept facts alleged by plaintiff without vouching for their objective truth). We next set out the general principles underlying a refusal-to- deal claim and then explain how the leading case—Aspen Ski- ing—maps onto Comcast’s conduct. We then reject Comcast’s argument that Viamedia’s claim could properly be dismissed on the pleadings. a. Monopolists and Refusals to Deal Monopolists are both expected and permitted to compete like any other firm. Cargill, Inc. v. Monfort of Colorado, Inc., 479 U.S. 104, 116 (1986). A monopolist is not obliged to “watch[] the quality of its products deteriorate and its customers No. 18-2852 45

become disaffected” and “lie down and play dead” because “even a monopolist is entitled to compete.” Goldwasser v. Ameritech Corp., 222 F.3d 390, 397 (7th Cir. 2000). “Part of com- peting like everyone else is the ability to make decisions about with whom and on what terms one will deal.” Id.; see also Au- thenticom, Inc. v. CDK Global, LLC, 874 F.3d 1019, 1025 (7th Cir. 2017). And just because “a firm has monopoly power doesn’t mean that the law should prevent it from competing,” as “[i]t would be absurd to require the [monopolist] to hold a price umbrella over less efficient entrants.” Richard A. Posner, An- titrust Law 196 (2d ed. 2001). Thus, the general rule is that even monopolists “are free to choose the parties with whom they will deal, as well as the prices, terms, and conditions of that dealing.” Pacific Bell Telephone Co. v. Linkline Communications, Inc., 555 U.S. 438, 448 (2009), citing United States v. Colgate & Co., 250 U.S. 300, 307 (1919). Yet there are “limited circumstances” under which a mo- nopolist’s refusal to deal with another party will be illegal an- ticompetitive conduct. Id.; see also Areeda & Hovenkamp ¶ 1800c5, at 21 (“Section 2 of the Sherman Act reaches unilat- eral refusals to deal when the refusals constitute monopoliza- tion … .”). For example, in Lorain Journal Co. v. United States, 342 U.S. 143 (1951), a monopolist newspaper was “an indis- pensable medium of advertising for many” local businesses but refused to deal with any advertiser who placed any ad with a new radio competitor in an effort “to destroy and elim- inate” the new competitor. Id. at 152, 150. The Court was not persuaded by the newspaper’s argument that it had “a right as a private business concern to select its customers and to re- fuse to accept advertisements from whomever it pleases.” Id. at 155. 46 No. 18-2852

In a holding that resonates in this case, Lorain Journal ex- plained: “In the absence of any purpose to create or maintain a mo- nopoly, the act does not restrict the long recognized right of trader or manufacturer engaged in an entirely private busi- ness, freely to exercise his own independent discretion as to parties with whom he will deal.” Id., quoting Colgate, 250 U.S. at 307; see also Goldwasser, 222 F.3d at 397 (acknowledging cir- cularity of Colgate test). With the newspaper’s clear expecta- tion that it would “outlast” the new competition and regain its complete monopoly, and with “no apparent efficiency jus- tification for its conduct,” Lorain Journal has been described as “entirely correct.” Robert H. Bork, The Antitrust Paradox: A Policy at War with Itself 344–45 (2d ed. 1993). This theory of li- ability was endorsed again in Aspen Skiing Co. v. Aspen High- lands Skiing Corp., 472 U.S. 585 (1985), the leading case on this issue. b. Aspen Skiing and Comcast Comcast takes the position that after the Supreme Court’s 2003 Trinko decision, any “antitrust claims based on a duty to deal with rivals ‘bit the dust.’” In the face of both Aspen Skiing and the actual language of Trinko, we must reject that argu- ment about what the law should be. Trinko itself said just the opposite: “Under certain circumstances, a refusal to cooperate with rivals can constitute anticompetitive conduct and violate Section 2,” and the “leading case for § 2 liability based on re- fusal to cooperate with a rival … is Aspen Skiing.” 540 U.S. at 408; see also Linkline, 555 U.S. at 448 (“There are also limited circumstances in which a firm’s unilateral refusal to deal with its rivals can give rise to antitrust liability.”), citing Aspen Ski- ing, 472 U.S. at 608–11. No. 18-2852 47

What do those limited circumstances look like? In addi- tion to reiterating Aspen Skiing’s continued, albeit narrow, va- lidity, the Court has also provided useful guidance on pri- mary factors to consider when determining whether poten- tially anticompetitive conduct falls within Aspen Skiing’s bounds. To provide background on what role those factors play in a court’s analysis, we summarize the facts of Aspen Skiing, focusing on the primary factors the Supreme Court has continued to highlight. The case involved four ski mountains that were initially developed and operated under separate ownership. 472 U.S. at 587. For over a decade, the four mountains offered a variety of ski-lift tickets and packages, including a joint ticket that al- lowed skiers to gain convenient access to all four mountains. Id. at 588–89. Even as defendant Aspen Skiing Company (Ski Co.) came to control three of the four mountains, thus gaining market power over the Aspen ski area, the joint “interchange- able ticket” program continued to include the fourth moun- tain, which was independently owned by plaintiff Aspen Highlands Skiing Corporation (Highlands). Id. at 590–92. Revenues from this cooperative arrangement were distrib- uted according to mountain usage. The joint ticket was a pop- ular and profitable package for both parties. The four-moun- tain package outsold by a two-to-one margin the Ski. Co. packages that offered access to only its three mountains. Id. at 592. Ski Co. management concluded, however, that if the four- mountain ticket were not available at all, customers would de- fault to buying just Ski Co.’s three-mountain pass. Ski Co.’s president explained that “the 4-area ticket was siphoning off revenues that could be recaptured by Ski Co. if the ticket was 48 No. 18-2852

discontinued.” Id. In the following year’s negotiations, Ski Co. made a revenue share offer to Highlands on such unfavorable terms that Ski Co. correctly expected Highlands “could not accept” it. Id. The joint ticket was no longer offered. Id. In an attempt to stanch the flow of lost business, Highlands “tried a variety of increasingly desperate measures to re-create the [four-mountain] joint ticket,” including “offering to buy the defendant’s tickets at retail price.” Trinko, 540 U.S. at 408–09, citing Aspen Skiing, 472 U.S. at 593–94. Ski Co. refused to per- mit Highlands even to “pay full retail value for the daily lift tickets,” with a Ski Co. official explaining, “we will not sup- port our competition.” Aspen Skiing, 472 U.S. at 593–94 n.14. Highlands filed suit. At trial, defendant Ski Co. primarily relied on the testi- mony of its economic expert, which included the theory that Ski Co.’s conduct had such procompetitive justifications as eliminating “free-riding by Highlands.”12 Ski Co. offered evi- dence that its own product was being devalued by being as- sociated with “the inferior skiing services offered at High- lands.” Aspen Skiing, 472 U.S. at 609–10. Ski Co. also argued that it could save administrative expenses and other costs by eliminating the joint ticket, which Ski Co. found “administra- tively cumbersome.” Id. at 592. In short, defendant Ski Co. ar- gued that “the conduct at issue was pro-competitive conduct that a monopolist could lawfully engage in.” Id. at 599. Procompetitive justifications were also highlighted in the jury instructions. The jury was instructed that a monopolist

12See George L. Priest & Jonathan Lewinsohn, Aspen Skiing: Product Differentiation and Thwarting Free Riding as Monopolization, in Antitrust Sto- ries 248 (Eleanor M. Fox and Daniel A. Crane, eds., 2007). No. 18-2852 49

“is not barred from taking advantage of scale economies by constructing a large and efficient factory,” nor is it “under a duty to cooperate with its business rivals … if valid business reasons exist for that refusal.” Id. at 597. Ski Co. could be found liable only if it “gained, maintained, or used monopoly power in a relevant market by arrangements and policies which rather than being a consequence of a superior product, superior business sense, or historic element, were designed primarily to further any domination of the relevant market.” Id. Therefore, “if there were legitimate business reasons for the refusal [to deal], then the defendant, even if he is found to possess monopoly power in a relevant market, has not vio- lated the law,” because the law is not concerned with conduct which may “benefit consumers by making a better product or service available”—only conduct that “has the effect of im- pairing competition.” Id. at 597. The jury “resolved all con- tested questions of fact in Highlands’ favor,” id. at 599, includ- ing a finding “that there were no valid business reasons for the refusal.” Id. at 605. The Supreme Court upheld the jury verdict for the plain- tiff. The Court reiterated Lorain Journal’s rejection of the argu- ment that “the right to refuse to deal with other firms … is unqualified.” Id. at 601–02 & n.27, citing Lorain Journal, 342 U.S. at 155, and Colgate, 250 U.S. at 307. This conclusion was supported by three key factors. First, Ski Co. “elected to make an important change in a pattern of distribution that had originated in a competitive market and had persisted for several years,” including after “the character of the market was changed by Ski Co.’s acqui- sition of monopoly power.” Id. at 603. Such a pre-existing re- lationship supports a presumption that the joint arrangement 50 No. 18-2852

was efficient and profitable. Trinko, 540 U.S. at 408–09 (distin- guishing Aspen Skiing from situation where that presumption would not apply—e.g., a defendant who would never have “voluntarily engaged in a course of dealing with its rivals … absent statutory compulsion”); see also Linkline, 555 U.S. at 450 (refusing to impose a duty to deal on a defendant when “such duty arises only from FCC regulations, not from the Sherman Act”). The Court explained in Aspen Skiing: In any business, patterns of distribution develop over time; these may reasonably be thought to be more efficient than alternative patterns of distribution that do not develop. The patterns that do develop and persist we may call the op- timal patterns. By disturbing optimal distribu- tion patterns, one rival can impose costs upon another, that is, force the other to accept higher costs. 472 U.S. at 604 n.31, quoting Robert H. Bork, The Antitrust Par- adox 156 (1978). Second, the Court compared Ski Co.’s conduct in the As- pen market with Ski Co.’s arrangements in comparable mar- kets where it lacked such dominance, noting that cooperative joint tickets were “used in other multimountain areas which apparently are competitive.” Aspen Skiing, 472 U.S. at 603–04 & n.30. The Court could thus “infer that such tickets satisfy consumer demand in free competitive markets.” Id. Third, defendant Ski Co. decided to forgo profitable trans- actions by refusing to permit Highlands to purchase ski tick- ets at the retail price for the sake of harming Highlands. 472 U.S. at 608 (“The jury may well have concluded that Ski Co. No. 18-2852 51

elected to forgo these short-run benefits because it was more interested in reducing competition … over the long run by harming its smaller competitor.”). Ski Co. made this “decision to avoid providing any benefit to Highlands even though ac- cepting the coupons would have entailed no cost to Ski Co. itself, would have provided it with immediate benefits, and would have satisfied its potential customers.” Id. at 610. These factors all pointed to Ski Co.’s conduct causing an- ticompetitive harm. But whether its conduct “may properly be characterized as exclusionary” also required consideration of possible procompetitive justifications, including any bene- ficial or harmful impacts on consumers or competition itself. Id. at 605, citing Bork, Antitrust Paradox at 138. Critical to this case, the Court treated procompetitive justification as a fac- tual issue properly resolved by the jury. The Court focused on “the evidence relating to Ski Co. itself, for Ski Co. did not per- suade the jury that its conduct was justified by any normal business purpose.” Id. at 608 (emphasis added). Conflicting evidence presented at trial undermined Ski Co.’s arguments that the joint ticket was “administratively cumbersome” (no more so than the joint tickets Ski Co. used in other, competi- tive markets) and that Highlands’ “inferior skiing services” were free-riding on Ski Co.’s services (a joint ticket “allowed consumers to make their own choice on these matters of qual- ity”). Id. at 608–10. Highlands refuted Ski Co.’s procompetitive justifications with exactly the kind of evidence that is helpful to prove ex- clusionary conduct or “predation,” including “statements made by the officers or agents of the company, evidence that the conduct was used threateningly and did not continue when a rival capitulated, or evidence that the conduct was not 52 No. 18-2852

related to any apparent efficiency.” Id. at 608–09 n.39, quoting Bork, Antitrust Paradox at 157 (emphasis in Aspen Skiing). The Court concluded that “the evidence supports an inference that Ski Co. was not motivated by efficiency concerns and that it was willing to sacrifice short-run benefits and consumer goodwill in exchange for a perceived long-run impact on its smaller rival.” Id. at 611. The Aspen Skiing factors help case-by-case assessments of whether a challenged refusal to deal is indeed anticompeti- tive, even though no factor is always decisive by itself. For ex- ample, even “a monopolist might wish to withdraw from a prior course of dealing and suffer a short-term profit loss in order to pursue perfectly procompetitive ends—say, to pur- sue an innovative replacement product of its own.” Novell, Inc. v. Microsoft Corp., 731 F.3d 1064, 1075 (10th Cir. 2013) (Gorsuch, J.). Similarly, forgoing short-run profits may some- times reflect desirable, procompetitive behavior, such as ef- forts to offer “promotional discounts.” Id. And a defendant may have “procompetitive rationales for treating a rival dif- ferently,” such as if “it’s more costly to deal with distant rivals than other nearby customers.” Id. at 1078 n.4. But because the factors as a whole provide a window into likely harm to com- petition, a court should start with the Aspen Skiing factors in determining whether a refusal to deal is unlawful. The Supreme Court has described Aspen Skiing as “at or near the outer boundary of § 2 liability.” Trinko, 540 U.S. at 409. Given the facts we must assume here, Viamedia has pre- sented a case that is well within those bounds and appears stronger than Aspen Skiing. A comparison of Viamedia’s alle- gations to the facts found by the jury in Aspen Skiing (and No. 18-2852 53

which the Supreme Court considered significant to its analy- sis) is instructive:

ASPEN SKIING VIAMEDIA ALLEGATIONS Long-term business rela- Same tionship that created joint offering Relationship existed absent Same any statutory obliga- tion/duty (Trinko) Can presume prior relation- Same ship was thus mutually ad- vantageous Sudden course reversal Same Course reversal came at a Same monetary loss for defendant Refused to sell ser- Same vice/product at retail price Sold product at retail price Same to others in the relevant market Unhappy customers Same Discouraged customers Same from doing business with its smaller rival 54 No. 18-2852

Defendant continued to Same deal with competitors in other competitive markets Procompetitive justifica- Same tions are a question for the factfinder Exclusionary conduct aimed Same at the only other competitor in the market Ski Mountain Passes Different: Ad Rep Services

In light of the similarities, unless the Court meant to limit Aspen Skiing to ski resorts, we see no sound basis to distin- guish Viamedia’s case as a matter of law. Comcast’s alleged conduct, absent compelling evidence to the contrary, indi- cates its “calculation that its future monopoly retail price would be higher” by foreclosing its ad rep services competi- tor. Trinko, 540 U.S. at 409. In addition, unlike in Aspen Skiing, where the ultimate customers were skiers who did not com- pete against the defendant ski resort, Comcast’s refusal to deal with Viamedia has left its MVPD customers in these mar- kets no practical choice but to turn over their ad sales busi- ness, along with their sensitive business information and a large percentage of their ad revenue, to their dominant MVPD competitor. c. Refusals to Deal and Motions to Dismiss Comcast nonetheless contends this case can be decided on the pleadings because “there is no liability under Aspen Skiing where, as here, the defendant’s alleged termination of a pre- existing course of dealing was not ‘irrational but for its anti- competitive effect.’” Comcast relies on the district court’s No. 18-2852 55

acceptance of Comcast’s thinly supported assertion that it had a “valid business purpose” in refusing to deal with Viamedia because Comcast’s replacement of Viamedia as WOW!’s and RCN’s ad representative is a course of conduct that “offers po- tentially improved efficiency.” See 218 F. Supp. 3d at 698–99 (emphasis added). Comcast contends this “valid business ob- jective” is what “distinguishes this case from Aspen Skiing, where the defendant ‘fail[ed] to offer any efficiency justifica- tion whatever for its pattern of conduct.’” Appellees’ Br at 27, quoting Aspen Skiing, 472 U.S. at 608. Comcast’s argument has the facts wrong. Its reading fails to comport with the actual language of the opinion, the jury instructions, and the evidence presented by both parties. In Aspen Skiing the Court was reviewing a jury verdict. Only af- ter a month-long trial had the jury “resolved all contested questions of fact in Highlands’ favor” and “concluded that there were no valid business reasons for the refusal.” Aspen Skiing, 472 U.S. at 599, 605. The Court concluded that “the ev- idence supports an inference that Ski. Co. was not motivated by efficiency concerns.” Id. at 610 (emphasis added). Comcast next cites Novell in support of its argument that a factual dispute regarding the existence of procompetitive jus- tifications is appropriate for resolution on the pleadings. Yet Novell was a decision based on an eight-week trial. 731 F.3d at 1066. And what about Olympia Equipment Leasing Co. v. West- ern Union Telegraph Co., 797 F.2d 370 (7th Cir. 1986)? That de- cision followed a “trial [that] lasted more than six weeks and produced the usual mountain of testimony and exhibits.” Id. at 372. Valid business justifications are relevant only to the re- buttal of a prima facie case of monopolization. 56 No. 18-2852

Thus, balancing anticompetitive effects against hypothe- sized justifications depends on evidence and is not amenable to resolution on the pleadings, at least where the plaintiff has alleged conduct similar to that in Aspen Skiing. See also, e.g., Illinois ex rel. Burris v. Panhandle Eastern Pipe Line Co., 935 F.2d 1469, 1482 (7th Cir. 1991) (“Whether valid business reasons motivated a monopolist’s conduct is a question of fact.”). Adapting language from our colleagues in the D.C. Circuit, the correct approach in this situation requires a district court to acknowledge that: [Comcast’s] defense—that its refusal to deal was economically justified—depends upon a ques- tion of fact and therefore is not cognizable in support of a motion to dismiss. It is, of course, entirely possible [Comcast] will be able to prove … [that] its refusal to deal was a reasonable business decision. On the other hand, it is also possible [Comcast’s] refusal to deal reflected its willingness to sacrifice immediate profits from the sale of [Interconnect access] in the hope of driving [Viamedia] out of the market and recov- ering monopoly profits in the long-run. … The district court cannot choose between these com- peting explanations without first resolving questions of fact not before it upon a motion to dismiss. Covad Communications Co. v. Bell Atlantic Corp., 398 F.3d 666, 676 (D.C. Cir. 2005) (reversing dismissal of refusal-to-deal claim on pleadings). This analysis must also include the harm from Comcast’s alleged tying conduct, which we turn to be- low. Viamedia has alleged—and offered evidence of— No. 18-2852 57

enough harm to competition from Comcast’s refusal-to-deal and tying conduct for its claim to go forward. Consideration of procompetitive justifications must wait for a comprehen- sive rule of reason analysis. i. Comcast’s Proposed Legal Standard Comcast both misunderstands the law and relies on inap- posite cases by conflating the vertical integration of its MVPD and ad rep services functions with its control over the coop- erative Interconnects and alleged misuse of that power. Com- cast proposes that if a defendant merely postulates “a valid business purpose”—apparently including any business pur- pose a defendant could dream up, regardless of feasibility or value—that “ends the inquiry.” “[T]here is no ‘balancing’ of benefits and harms,” Comcast declares. In support of that proposition, Comcast points to the United States’ Amicus Brief (in support of neither party) filed in this case, which of- fers a test dubbed the “no economic sense test.” Appellee Br at 27–28; see also United States Brief at 11–12 (relying on the formulation articulated in the United States’ amicus brief in Trinko, available at 2003 WL 21269559, and elaborated upon in Gregory J. Werden, Identifying Exclusionary Conduct Under Section 2: The “No Economic Sense” Test, 73 Antitrust L.J. 413, 422–25 (2006)). The proposed “no economic sense” test would condemn conduct as “exclusionary or predatory” only if it “would make no economic sense for the defendant but for its ten- dency to eliminate or lessen competition.” United States Brief at 11. A “gross benefit [or gain] for the defendant” is not enough, however: “Conduct fails the no economic sense test if it is expected to yield a negative payoff, net of the costs of undertaking the conduct, and not including any payoff from 58 No. 18-2852

eliminating competition.” Werden at 416 (emphasis added). Or—as explained by the government at oral argument here— it is an objective “balancing” test that requires more than just “a slight procompetitive benefit or efficiency gain.”13 This test is essentially the same one employed by the Tenth Circuit in Novell, which noted that “the monopolist’s conduct must be irrational but for its anticompetitive effect.” 731 F.3d

13Comcast’s confusion may stem from the terse proposed name of “no economic sense,” which does not appear to invite balancing. The test is actually more nuanced than the name suggests, and it is not meant to resolve every Section 2 challenge. As Werden sensibly notes, its “utility … ultimately is apt to vary,” and we “should not presume that a single test must resolve every exclusionary conduct case.” Werden at 421 & n.31. Furthermore, it has been observed that although the “no economic sense” test “offers good insights into when aggressive actions by a single firm go too far,” it “can lead to erroneous results unless” one also “seek[s] to ‘balance’ gains to the monopolist against losses to consumers, rivals, or others.” Areeda & Hovenkamp ¶ 651b3, at 106–07. Otherwise we could arrive at absurd outcomes: “Theoretically, an act might benefit the defend- ant very slightly while doing considerable harm to the rest of the econ- omy, and it would be lawful.” Id. It is possible the test could be adapted to meet these criticisms, given that a court should not consider any gain from eliminating competition, but—in any event—the “no economic sense test” was not intended to displace all other approaches. Rather, it “is likely to be most useful as one part of a sufficient condition: If challenged con- duct has a tendency to eliminate competition and would make no eco- nomic sense but for that tendency, the conduct is exclusionary.” Werden at 418. Areeda and Hovenkamp also suggest a broader approach, in which harm “wholly disproportionate” to the valid business justification can also support a refusal-to-deal-claim. ¶ 772c2, at 223 (“Condemnation would be appropriate only for conduct that (1) clearly injures an actual or prospec- tive rival either (2a) with no good business justification at all, or (2b) with a business justification that is poorly fitted to the result or wholly dispro- portionate to the harm that is inflicted.”). No. 18-2852 59

at 1075, citing Aspen Skiing, 472 U.S. at 597, Trinko, 540 U.S. at 407, and Werden at 422–25. However formulated, this test is aimed in part at the potential overweighting of the Aspen Ski- ing factor of a defendant forsaking short-term profits. Id. As noted, this factor is relevant but should not always be dispos- itive because “a short-term profit sacrifice is neither necessary nor sufficient for conduct to be exclusionary.” Werden at 424; see also id. (“short-run profit sacrifice also is not necessary for conduct to be exclusionary because the anticompetitive gains from exclusionary conduct sometimes can be reaped immedi- ately”); Areeda & Honvenkamp ¶ 651b3, at 107 (“monopoliz- ing conduct is not necessarily costly to the defendant”). Because the Aspen Skiing factors are helpful but not dis- positive, this more nuanced approach considering both pro- competitive benefits and anticompetitive harms is necessary to answer the ultimate question of whether competition was harmed. The plaintiff ultimately needs to prove “that the mo- nopolist’s refusal to deal was part of a larger anticompetitive enterprise, such as (again) seeking to drive a rival from the market or discipline it for daring to compete on price.” Novell, 731 F.3d at 1075. The result of such conduct is to harm com- petition by “entrench[ing] a dominant firm and enabl[ing] it to extract monopoly rents once the competitor is killed off or beaten down.” Id. As the above paragraphs suggest, and without our en- dorsing any particular catchy title for this analytical ap- proach, the calculation of procompetitive benefits net of anti- competitive harms does not easily lend itself to a pleading standard. Rule of reason cases “place[] a premium on objec- tive tests based on evidence that is typically not in the defend- ant’s exclusive control”—for example, Comcast’s cost savings 60 No. 18-2852

and other efficiencies it may have obtained due to its conduct. See Herbert Hovenkamp, The Rule of Reason, 70 Fla. L. Rev. 81, 86 (2018). This is why it is typically considered an “adequate pleading in a rule of reason antitrust case” for a plaintiff to allege (1) “evidence of market structure” (i.e., market power and relevant markets, which are not in dispute in this case) and (2) “exclusionary effect” (i.e., foreclosure of a competitor from a market, which is also not in dispute in this case)— “both of which can ordinarily be obtained without access to the defendant’s own records—[and] indicate that an antitrust violation is plausible.” Id at 90. To the extent that refusal-to-deal claims require more at the pleading stage, it is enough to allege plausibly that the re- fusal to deal has some of the key anticompetitive characteris- tics identified in Aspen Skiing. The Supreme Court said as much in Trinko, in which it affirmed dismissal of a complaint, distinguished Aspen Skiing, and emphasized that “the defend- ant’s prior conduct sheds no light upon the motivation of its refusal to deal—upon whether its regulatory lapses were prompted not by competitive zeal but competitive malice.” 540 U.S. at 409. Trinko specifically identified the absence of two factors—a prior and voluntary course of dealing, and re- fusal to sell at retail price—in distinguishing Aspen Skiing. The former factor was important, because it “suggested a willing- ness to forsake short-term profits to achieve an anticompeti- tive end.” Id., citing Aspen Skiing, 472 U.S. at 608, 610–611. In Covad, the D.C. Circuit adopted a similar, if slightly more ex- plicit, holding that a plaintiff must eventually show a sacrifice of short-term profits to prevail on a refusal-to-deal claim and that alleging that a refusal to deal was “predatory” was suffi- cient at the pleadings stage. 398 F.3d at 675–76. No. 18-2852 61

Viamedia’s pleading adequately alleges an anticompeti- tive refusal to deal. As described above, Viamedia’s claim closely tracks Aspen Skiing and contains the key elements that were missing in Trinko: a prior course of voluntary conduct, sacrifice of short-term profits, and refusal to sell to rivals on the same terms as other potential buyers. Certainly, no more is required. We leave open the question whether allegations of short-term losses are necessary to state a refusal-to-deal claim. A case might present itself in which other factors—such as a prior course of conduct, exploitation of power over a co- operative network, refusal to sell at retail price, and discrimi- natory treatment of rivals—could plausibly support the infer- ence that a refusal to deal is “prompted … by anticompetitive malice.” Trinko, 540 U.S. at 880. But this case is easier and does not require precise delineation of the requirements of a re- fusal-to-deal pleading. Even if an allegation that a defendant’s conduct was irra- tional but for its anticompetitive effect were necessary, Vi- amedia has plausibly alleged just that. In a section of the First Amended Complaint entitled “Comcast’s Refusal to Deal with Viamedia is Irrational But for its Anticompetitive Ef- fects,” Viamedia walked through the long-term course of dealing prior to Comcast’s conduct; the subsequent degrada- tion of the value of the cooperative Interconnects; the financial losses suffered by Comcast itself, as well as by Viamedia and Comcast’s competitor MVPDs; Comcast’s willingness to offer Interconnect-only access in other markets where it did face competition; and the fact that “[t]here are no procompetitive justifications” to be achieved by the conduct given that there were “no material administrability problems in allowing Vi- amedia to participate in Interconnects” on behalf of its MVPD customers. First Am. Cplt. ¶¶ 154–68. Viamedia’s allega- 62 No. 18-2852

tions—regardless of the standard applied—are more than suf- ficient to pass muster under Bell Atlantic Corp. v. Twombly, 550 U.S. 544 (2007), and Ashcroft v. Iqbal, 556 U.S. 662 (2009). With Viamedia meeting the appropriate pleading stand- ard and presenting evidence of harm to competition, the re- mainder of the case should settle into the traditional analysis followed in rule of reason cases, with the burden shifting to Comcast, which “may, of course, introduce its own proof of inevitability, superior skill, or business justification.” Areeda & Hovenkamp ¶ 650c, at 94. To be more specific, under this burden-shifting framework, once Viamedia has “successfully establish[ed] a prima facie case under § 2 by demonstrating an- ticompetitive effect, then [Comcast] may proffer a ‘procom- petitive justification’ … a nonpretextual claim that its conduct is indeed a form of competition on the merits because it in- volves, for example, greater efficiency or enhanced consumer appeal.” Microsoft, 253 F.3d at 59. If such a justification is of- fered, “the burden shifts back to [Viamedia] to rebut that claim.” Id. If Viamedia cannot rebut the evidence of Comcast’s procompetitive justifications, “then [Viamedia] must demon- strate that the anticompetitive harm of the conduct outweighs the procompetitive benefit.” Id. This burden-shifting has evolved based on which party has access to the various cate- gories of evidence and information, with any evidence of pro- competitive justifications likely to be under the defendant’s control. Cf. United States Department of Justice and Federal Trade Commission, Horizontal Merger Guidelines § 10, at 30 (Aug. 19, 2010) (Merger Guidelines) (“much of the No. 18-2852 63

information relating to efficiencies is uniquely in the posses- sion” of the firms seeking to justify a transaction).14 ii. Inapposite Vertical Integration Cases Comcast also relies on cases involving vertically inte- grated defendants with facts that, in crucial ways, do not map onto this case. In seeking to shoehorn this case into this cate- gory, Comcast caused confusion in the district court—and continues in this effort on appeal—by glossing over the unu- sual market structures in this case and portraying itself as just a “prototypical” vertically integrated firm. This misconcep- tion is accomplished by conflating (1) Comcast’s actual verti- cal integration of its MVPD cable services with its ad rep ser- vices functions with (2) its control over the cooperative Inter- connects in the relevant geographic markets. Comcast’s

14 Courts apply a “similar balancing approach” in rule of reason cases,

whether alleged under § 1 or § 2. Microsoft, 253 F.3d at 59, citing Standard Oil Co. v. United States, 221 U.S. 1 (1911). The Supreme Court has recently reiterated this balancing test for a rule of reason § 1 case: “To determine whether a restraint violates the rule of reason … a three-step, burden-shift- ing framework applies. Under this framework, the plaintiff has the initial burden to prove that the challenged restraint has a substantial anticom- petitive effect that harms consumers in the relevant market. If the plaintiff carries its burden, then the burden shifts to the defendant to show a pro- competitive rationale for the restraint. If the defendant makes this show- ing, then the burden shifts back to the plaintiff to demonstrate that the procompetitive efficiencies could be reasonably achieved through less an- ticompetitive means.” Ohio v. American Express Co., 138 S. Ct. 2274, 2284 (2018) (omitting internal citations); see also Areeda & Hovenkamp ¶ 1820a, at 188, ¶ 1821, at 207 (observing that courts require “stronger proof of offsetting efficiencies” when defendants possess greater ability to foreclose rivals from the market). 64 No. 18-2852

argument for dismissal on the pleadings depends upon this confusion. To start, “[e]ven a monopolist … is free to integrate, espe- cially when integration creates no new monopoly in any sec- ond area.” Areeda & Hovenkamp ¶ 1700j1, at 14–15. Such an integration allows the defendant to achieve cost-savings by “elimination of double marginalization.” United States v. AT&T, Inc., 916 F.3d 1029, 1036 (D.C. Cir. 2019). In other words, prior to vertical integration, the firms providing com- plementary products would “earn[] margins over cost before their products reached consumers.” Id. at 1044. After integra- tion, goes the theory, there is no need for two entities to earn margins over cost, and “the merged entity would eliminate that cost and … pass on some of those cost savings to consum- ers in order to attract additional” customers. Id. Thus, it would rarely be an antitrust violation for a firm to supply it- self through vertical integration, and a plaintiff would not generally have a right under antitrust law to demand that a defendant forgo supplying itself from an in-house source. Areeda & Hovenkamp ¶ 1700j1, at 14–16 & n.35. This principle has been illustrated in some Section 2 cases, in which a company claimed antitrust injury when a larger company vertically integrated and provided in-house what it formerly purchased from the smaller company. For example, in Port Dock & Stone Corp. v. Oldcastle Northeast, Inc., 507 F.3d 117 (2d Cir. 2007), a company with a local monopoly in crushed stone (aggregate) sold the aggregate through two dis- tributors. The aggregate monopolist decided to vertically in- tegrate by purchasing one of the distributors and bringing all of its distribution in-house. Id. at 119. The remaining distrib- utor alleged a Section 2 violation, claiming that “its injury No. 18-2852 65

resulted from [defendant’s] vertical integration into the distri- bution market.” Id. at 120. The facts of Port Dock do not map onto the conduct of Comcast, which was already vertically in- tegrated and was instead exploiting its control over the coop- erative Interconnects. Similarly, the defendant in Christy Sports, LLC v. Deer Val- ley Resort Co., 555 F.3d 1188 (10th Cir. 2009), owned the moun- tain on which it leased property to plaintiff, and “the creator of a resort has no obligation under the antitrust laws to allow competitive suppliers of ancillary services on its property.” Id. at 1193. Notably, the mountain was not a cooperative enter- prise, and the customers (skiers) did not compete against the defendant mountain resort. And in Novell (which, recall, was a case that went to trial), defendant Microsoft was the sole owner of the intellectual property it had made available to in- dependent software vendors. 731 F.3d at 1067. And again, the customers (computer users) did not compete against Mi- crosoft. These opinions about “prototypical” vertically integrated firms recognize, nevertheless, that different circumstances could support a cognizable antitrust claim in cases like this one. See Port Dock, 507 F.3d at 124–25 (“Vertical expansion by a monopolist, without more, does not violate section 2,” un- less there is an allegation of an “anticompetitive incentive to create a downstream monopoly,” or other “special circum- stances in which a monopolist’s vertical expansion could be anticompetitive.”); Christy Sports, 555 F.3d at 1196 (“We would not even preclude the theoretical possibility that such a change [by refusing to deal] could give rise to an antitrust claim, for example, if by first inviting an investment and then disallowing the use of the investment the [defendant] 66 No. 18-2852

imposed costs on a competitor that had the effect of injuring competition in a relevant market.”); Novell, 731 F.3d at 1076 (plaintiff could have proven refusal-to-deal case against Mi- crosoft, but at trial “presented no evidence from which a rea- sonable jury could infer that Microsoft’s discontinuation of this arrangement suggested a willingness to sacrifice short- term profits, let alone in a manner that was irrational but for its tendency to harm competition”). Even if this were a vertical integration case, Viamedia ad- equately alleged that Comcast presented just such a “special circumstance.” Its conduct eliminated its only competitor in the ad rep services market and increased control over its MVPD competitors in the retail cable market. But again, this is not a case of simple vertical integration. Comcast is verti- cally integrated and has been at all relevant times. No one ob- jects to a vertically integrated Comcast offering both Intercon- nect services and ad rep services. Viamedia does not seek to force Comcast to buy ad rep services from Viamedia; nor does Viamedia seek to force Comcast to allow Viamedia to re-sell or distribute Comcast’s ad rep services. Viamedia simply wants to ensure that MVPDs can freely choose Viamedia as their supplier of ad rep services if that is their preferred choice. Another distinguishing fact is that the Interconnects are joint, cooperative efforts among competing MVPDs. That dis- tinguishes this case from cases involving vertically integrated defendants, as in Port Dock. Viamedia seeks to regain access to the Interconnects to operate on behalf of its MVPD custom- ers. It is true that by virtue of acquiring numerous other cable companies, Comcast now controls the Interconnects at issue. But they are cooperative ventures that jointly set prices for No. 18-2852 67

competitor MVPDs’ spot avails. Comcast itself has described the Interconnects as a “collection of two or more cable TV sys- tems that work together to distribute commercials to a wider geographic area than a single system would otherwise reach, giving advertisers the option to reach all cable households within a market with one buy.” First Am. Cplt. ¶ 156. The Interconnects’ cooperative structure explains why Comcast describes itself not as an Interconnect “owner” but as an Interconnect “operator” and describes its function as a firm that “operates interconnects in DMAs including Chicago and Detroit.” See, e.g., Appellees’ Br. at 7. It also explains why Comcast describes Viamedia as having “participated in inter- connects,” and after being denied access, Comcast says, Vi- amedia sought “readmission.” Id. Typical vertically inte- grated firms do not refer to themselves as the “operators” of their assets, and they do not describe their buyers as “partici- pating” in the vertically integrated firms’ services, let alone say that buyers might seek “readmission” to those services. Taking control of and exploiting control of a previously coop- erative mechanism is not vertical integration.15 Accordingly, we reverse the Rule 12(b)(6) dismissal of Vi- amedia’s claim for monopolization through an unlawful re- fusal to deal. 2. Tying The “essential characteristic of an invalid tying arrange- ment lies in the seller’s exploitation of its control over the

15 Viamedia disputes the accuracy of describing Comcast as an Inter- connect “owner.” Deposition testimony characterizes an Interconnect simply as “an agreement between two or more MVPDs in a DMA to dis- tribute commercials … across all partners in the interconnect.” 68 No. 18-2852

tying product to force the buyer into the purchase of a tied product that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms.” Jef- ferson Parish Hospital Dist. No. 2 v. Hyde, 466 U.S. 2, 12 (1984). Viamedia contends that Comcast conditioned the sale of In- terconnect services (the tying product) on the purchase of ad rep services (the tied product). Viamedia alleges that Comcast engaged in anticompetitive conduct on two fronts. Comcast inserted itself between Viamedia and its competitor MVPDs by: (1) denying Viamedia access to the Comcast-controlled In- terconnects, and (2) then using its control over the Intercon- nects to demand that its smaller MVPD competitors turn over to Comcast 100% of their spot avails, including the sale of lo- cal spots, an area in which Comcast and the MVPDs had for- merly competed. Viamedia has offered evidence that this two- front strategy was successful. Comcast excluded its only com- petitor in the ad rep services market—gaining a pure monop- oly. It also gained new control over and insight into its MVPD competitors that it could not have achieved otherwise. The district court, however, granted summary judgment on Viamedia’s tying claim, a decision we also review de novo. Schlaf v. Safeguard Prop., LLC, 899 F.3d 459, 465 (7th Cir. 2018). “Summary judgment is appropriate only if there are no dis- puted questions of material fact and the moving party is enti- tled to judgment as a matter of law,” so we “examine the rec- ord in the light most favorable to the [non-movant], granting [it] the benefit of all reasonable inferences that may be drawn from the evidence and reversing if we find a genuine issue concerning any fact that might affect the outcome of the case.” Id. (citations omitted); see also Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255 (1986) (on summary judgment, courts must No. 18-2852 69

refrain from making credibility determinations or weighing evidence). a. Summary Judgment Standard The district court applied, and Comcast argues for, a sum- mary judgment standard that requires plaintiffs in Section 2 monopolization cases to present evidence that “tends to ex- clude the possibility” that a monopolist’s conduct is just “as consistent with competition as with illegal conduct.” 335 F. Supp. 3d at 1061, quoting Matsushita Electric Industrial Co. v. Zenith Radio Corp., 475 U.S. 574, 588 (1986). The proper ques- tion on summary judgment is whether Viamedia has pre- sented evidence to establish a genuine dispute of material fact as to whether Comcast engaged in exclusionary conduct for- bidden by Section 2. See Matsushita, 475 U.S. at 585; Eastman Kodak Co. v. Image Technical Services, 504 U.S. 451, 483 (1992); see also, generally, In re Text Messaging Antitrust Litig., 782 F.3d 867 (7th Cir. 2015). When determining whether there is a “genuine issue of material fact … the substantive law will identify which facts are material. Anderson v. Liberty Lobby, Inc., 477 U.S. at 247–48 (emphasis omitted). “The Court’s re- quirement in Matsushita that the plaintiffs’ claims make eco- nomic sense did not introduce a special burden on plaintiffs facing summary judgment in antitrust cases.” Eastman Kodak, 504 U.S. at 467–68. Viamedia’s tying theory is not economically implausible, unlike the alleged twenty-year-long conspiracy to charge predatorily low prices in Matsushita itself. A competitor’s claim that a rival used monopoly power in a tying product market to gain a monopoly in a tied product market is “facially anticompetitive and exactly the harm that antitrust laws aim to prevent.” Eastman Kodak, 504 U.S. at 479. The 70 No. 18-2852

sufficiency of the tying claim depends on whether Comcast forced RCN and WOW! to buy its ad rep services. Our summary judgment inquiry can be framed in the language of Matsushita—are the facts “as consistent with” forcing as with noncoerced action?—but this formulation does not get us anywhere beyond the general summary judgment standard. In the following analysis, we ask whether Viamedia has presented evidence of forcing sufficient to create a genuine dispute for trial. b. Tying and Comcast’s Conduct In granting summary judgment, the district court con- cluded that no reasonable jury could find as a matter of fact that Comcast tied Interconnect services to the purchase of its ad rep services. We respectfully disagree. Viewing the evi- dence in the light most favorable to Viamedia, and without making credibility determinations or weighing the parties’ competing evidence, we conclude that Viamedia has offered sufficient evidence that Comcast illegally tied purchase of its ad rep services to the Interconnect access it already controlled. First, it is undisputed that Comcast (a) has market power in the tying market for Interconnect services and (b) has now foreclosed all competition in the tied market for ad rep ser- vices. Second, there is substantial evidence that the coopera- tive Interconnects are a separate service from Comcast’s ad rep services. Third, Viamedia offered evidence that Comcast forced its competitor MVPDs to become its customers for ad rep services if they also wanted to keep their access to the In- terconnects. No. 18-2852 71

i. Definition Tying is conduct in which a firm will “sell one product [the tying product] but only on the condition that the buyer also purchases a different (or tied) product.” Northern Pacific Ry. Co. v. United States, 356 U.S. 1, 5–6 (1958); see also Sheridan v. Marathon Petroleum Co., 530 F.3d 590, 592 (7th Cir. 2008); Areeda & Hovenkamp ¶ 1700a, at 4. The seller will purchase the tied product “not because the party imposing the tying requirement has a better product or a lower price” but be- cause the seller has “power or leverage” in the market for the tying product. Northern Pacific Railway, 356 U.S. at 6. Tying is still nominally subject to a per se rule of illegality, but it is “a most peculiar per se rule.” Areeda & Hovenkamp ¶ 1701c, at 31; see Jefferson Parish Hospital Dist. No. 2 v. Hyde, 466 U.S. at 26–29; Eastman Kodak, 504 U.S. at 462. 16 The factual elements that must be proven for a tying claim capture much of what must be demonstrated in a rule of reason case. Show- ing that the purchase of the tied product was forced uses many of the same concepts used to analyze refusals to deal: some assessment of market power, rough predictions of anti- competitive harm, and consideration of procompetitive

16 We stay “still” because in recent years the Supreme Court has held that some categories of conduct that were formerly treated as per se illegal are now subject to a rule of reason analysis. See Leegin Creative Leather Products, Inc. v. PSKS, Inc., 551 U.S. 877, 907 (2007) (resale price mainte- nance no longer per se illegal, overruling Dr. Miles Medical Co. v. John D. Park & Sons, 220 U.S. 373 (1911)); Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36, 57 (1977) (non-price vertical restraints no longer per se illegal, overruling United States v. Arnold, Schwinn, & Co., 388 U.S. 365 (1967)); State Oil Co. v. Khan, 522 U.S. 3, 22 (1997) (vertical maximum price re- straints no longer per se illegal, overruling Albrecht v. Herald Co., 390 U.S. 145 (1968)). 72 No. 18-2852

justifications. See, e.g., Illinois Tool Works Inc. v. Independent Ink, Inc., 547 U.S. 28, 44 (2006) (“While some such [tying] ar- rangements are still unlawful, such as those that are the prod- uct of a true monopoly … that conclusion must be supported by proof of power in the relevant market.”) (citation omitted); Sheridan v. Marathon Petroleum Co., 530 F.3d 590, 593–94 (7th Cir. 2008) (describing the tying rule and its market-power re- quirement); Areeda & Hovenkamp ¶ 1760(b), at 379 (noting that, even when treated as per se illegal, “the Supreme Court has almost always been willing to consider a defendant’s of- fered justifications”). “When the defendant is a dominant firm” and meets “a much stricter power requirement,” however, the “special screening function” of the tying factors is “largely unneces- sary, and the more general standards of § 2 become relevant” because “the technical requirements … attach only to per se ties.” Areeda & Hovenkamp ¶ 777, at 324. Thus, “when the defendant is a monopolist in the ‘tying product,’” it may be superfluous to go through a detailed inquiry into whether there are “separate products.” Id. ¶ 617b2, at 52–53; see also id. ¶ 777, at 324–25 & n.9 (noting that in Eastman Kodak, the Supreme Court treated a conditional refusal to sell parts with- out service as a tying arrangement, although on remand the tying claim was dropped and a Section 2 violation was found without any “separate products” requirement); United States v. Microsoft Corp., 253 F.3d 34, 96–97 (D.C. Cir. 2001) (en banc). Similarly, “[w]hen a defendant’s market share and the under- lying market structure make monopolization or attempt plau- sible, then a tie that contributes significantly to the mainte- nance or creation of monopoly power violates § 2 even though it is unilaterally imposed.” Areeda & Hovenkamp ¶ 777, at 325. No. 18-2852 73

Here, Comcast’s monopoly power in the tying market of Interconnect services in the three metropolitan areas and its successful capture of a monopoly position in the tied market of ad rep services in the same areas are undisputed. And yet great effort has been made to parse whether Comcast’s con- duct satisfies some platonic ideal of tying conduct. We, too, walk through the tying factors at issue (separate product and forced purchase) and determine, taking the record as a whole, that Viamedia has provided sufficient evidence to create a question of fact as to each factor. Ultimately, the focus in this Section 2 case must remain on “whether, viewing the monop- olist’s conduct as a whole, it has unreasonably maintained or enhanced its monopoly position.” Id. ¶ 777, at 324. ii. Separate Products or Services “[W]hether one or two products are involved turns … on the character of the demand for the two items.” Jefferson Par- ish, 466 U.S. at 19; see also Eastman Kodak, 504 U.S. at 462 (“For service and parts to be considered two distinct products, there must be sufficient demand so that it is efficient for a firm to provide service separately from parts.”); Northern Pacific Rail- way, 356 U.S. at 5–6. Comcast disputes whether Interconnect access and ad rep services are separate services. The district court assumed they are. On this record, that was correct. The fact that buyers may wish to purchase and use two complementary products together does not, in and of itself, convert the two separate products into a single product. Ra- ther, the market must “be assessed at the pre-contract rather than post-contract stage.” Areeda & Hovenkamp ¶ 1802d6, at 89, citing Tampa Electric Co. v. Nashville Coal Co., 365 U.S. 320 (1961); see also id. ¶ 1802d6, at 88 (noting that a “grouping of sales covered by [a single] contract does not become a relevant 74 No. 18-2852

market for that reason”). In this case, RCN and WOW! viewed the services as separate prior to entering into their present contracts with Comcast. Interconnect services and ad rep services are different functionally, as already described at length. See above at pages 10–24. To summarize, a provider of Interconnect ser- vices bundles and re-sells ads from multiple MVPDs in a re- gional market. An ad rep has a more direct relationship with an MVPD, directly representing it in regional and/or local ad sales, and potentially acting as its representative with an In- terconnect. The “character of the demand” for the two ser- vices also differs, as demonstrated by how the market partic- ipants have sold and purchased the services. See Jefferson Par- ish, 466 U.S. at 19–21. Viamedia has offered only ad rep ser- vices for almost two decades. Comcast formerly offered the two services separately in the relevant DMA geographic mar- kets. It continues to offer them separately in other DMAs. MVPDs like Comcast that also operate Interconnects “often have separate salespeople selling local advertising and selling Interconnect advertising.” DA650. And the other alleged vic- tims of Comcast’s tying conduct—its smaller MVPD compet- itors—previously purchased these services separately and, as shown above, expressed strong interest in continuing to do so when Comcast was forcing them to buy the two together in Chicago, Detroit, and Hartford. Comcast’s claim that the two services are not distinct is also flatly inconsistent with one of Comcast’s primary argu- ments on the next factor. In arguing that it did not force the MVPDs to buy its ad rep services, Comcast points to other DMAs where it sells Interconnect services separately. Even if there might be a viable factual dispute on whether it is No. 18-2852 75

possible to consider these two services a single “unified mar- ket,” which we doubt, that would be at minimum a question to “be resolved by the trier of fact.” Eastman Kodak, 504 U.S. at 463. Comcast is not entitled to summary judgment on this ground. iii. Forced Purchase Ample evidence shows that Comcast conditioned MVPDs’ access to the Interconnects on hiring Comcast as their ad rep. Internal Comcast documents indicate that it ended its relationship with Viamedia specifically to obtain full-turnkey deals with the MVPDs. A216 n.84. Internal WOW! emails show that its executives understood that WOW! would have to hire Comcast for ad rep services if it wanted to regain access to the Interconnects. A215 n.81. RCN testified to the FCC that “Comcast will only allow RCN to join the interconnects if RCN employs Comcast spotlight instead of Viamedia.” Id. A Comcast employee working in the Chi- cago and Detroit markets testified to the Department of Jus- tice that Comcast had a business practice that “if an MVPD wants to get access to a Comcast [Spotlight] controlled Inter- connect, it has to hire Comcast [Spotlight] as its ad sales rep- resentative.” Id., A835. And a Comcast employee responded to the question “were you also clear … that Comcast Spotlight was interested only in a turnkey deal?” with “Direct relation- ship, full turnkey, yes, we made that clear to [WOW!].” A811. Both Comcast and the partial dissent offer explanations and rationales to try to defang these unusually explicit pieces of evidence. See Post at 136–37. On review of summary judg- ment, of course, Viamedia is entitled to the benefit of reason- able inferences and interpretations in its favor. The opposing 76 No. 18-2852

arguments are suitable for a trial but are not grounds for af- firmance. A jury could easily find that Comcast improperly forced the smaller MVPDs to buy its ad rep services using its monop- oly in the Interconnect services market. The entire purpose of its refusal to deal with Viamedia—conceded repeatedly by Comcast—was to force RCN and WOW! to become full-turn- key clients for ad rep services. A789. Every party involved un- derstood that this was the practical effect of banning from the Interconnects MVPDs that received ad rep services from Vi- amedia. The fact that the arrangements were structured so that ownership of the slot avails passed from the MVPDs to Vi- amedia does not affect this analysis. In applying the antitrust laws, we care more about economic substance than about form. See Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 760 (1984). Smaller MVPDs do not have their own ad rep services divisions. As a practical matter, they cannot self-pro- vide ad rep services and must work with an ad rep to interface with the Interconnects. Given these dynamics, Viamedia of- fered sufficient evidence that Comcast explicitly used its con- trol over the Interconnects to deny access to its competitor MVPDs or their agent to force RCN and WOW! to use Com- cast’s ad rep services for all spot avails, including the two- thirds of spot avails sold outside of the Interconnects, many of which used to be sold locally in competition with Comcast. See Collins Inkjet Corp. v. Eastman Kodak Co., 781 F.3d 264, 272 (6th Cir. 2015) (“When a defendant adopts a policy that makes it unreasonably difficult or costly to buy the tying product (over which the defendant has market power) without buying the tied product from the defendant, it ‘forces’ buyers to buy No. 18-2852 77

the tied product from the defendant and not from competi- tors.”); Microsoft, 253 F.3d at 64 (liability appropriate where monopolist bars rival from “cost-efficient” means of distribu- tion even if some means of distribution remain open).17 Because this is not a typical bundling case, Comcast’s and the partial dissent’s reliance on Aerotec Int’l, Inc. v. Honeywell Int’l, Inc., 836 F.3d 1171 (9th Cir. 2016), is misplaced. In that case, the plaintiff purchased Honeywell’s replacement parts and then “bundle[d] parts and repairs in an effort to woo” the ultimate airline customers, in part reselling defendant’s prod- ucts. Id. at 1176. Moreover, Aerotec presented no evidence that Honeywell “explicitly or implicitly” tied or conditioned the sale of the tying product of parts to the ultimate airline customers’ purchase of the tied product of maintenance. Id. at 1179. The court found compelling that “Honeywell allows

17 The partial dissent insists that that there was no conditioning and that “RCN and WOW! maintained the ability to deal directly with Com- cast and access the Interconnect without any ad representative should they choose not to employ Comcast.” Post at 118–19 This is wrong as a matter of fact. As described above, every party involved—Comcast, RCN, and WOW!—understood that RCN and WOW! would be unable to access the Interconnects unless they hired Comcast to provide ad rep services. Even though its reasoning relies on the possibility that RCN and WOW! could access the Interconnects without Viamedia or Comcast, the partial dissent points to no evidence supporting that possibility. The record con- tains evidence of the opposite: RCN and WOW! needed to employ an ad rep services provider, and once Comcast refused to deal through their cho- sen intermediary, they had no practical choice but to obtain ad rep services from Comcast. As described above, substantial evidence shows that both RCN and WOW! understood themselves to be forced by Comcast into purchasing its ad rep services. A215 n.81. We cannot affirm summary judgment by overlooking that evidence about the realities of the parties’ dealings and the economic realities of the market. 78 No. 18-2852

airlines to purchase parts and services in separate transac- tions from whichever supplier they please.” Id. But Viamedia is not a bundler. It has presented substantial evidence of an explicit tie, and its former customers could not separately ob- tain Interconnect access. There are other fundamental differences between the product offered in Aerotec and the one offered here. Honey- well sold airplane parts and repair services. Comcast, by con- trast, operates a platform that is necessarily cooperative. RCN and WOW! are not just potential customers of Comcast as the airlines were potential customers of Honeywell. They are also Comcast’s rivals in the retail cable market. Because self- providing ad rep services was not a viable option for RCN and WOW!, refusing to deal with their chosen intermediary had the effect of forcing them into much less desirable direct relationships with Comcast, their monopolist-competitor. Vi- amedia has presented evidence that the MVPDs were reluc- tant to be forced into their more powerful rivals’ arms. A886– 87. There was no evidence in Aerotec that the airlines felt sim- ilarly threatened, that providing repairs in-house was eco- nomically infeasible, or that the airlines viewed Honeywell as a necessary intermediary. Comcast’s reliance on It’s My Party, Inc. v. Live Nation, Inc., 811 F.3d 676 (4th Cir. 2016), is even further off-base. That case involved “a world of robust market competition where [the ultimate customers] were free to take a package deal of pro- motion and venues, free to purchase those products sepa- rately, free to turn down both, and where they in fact exer- cised all those options to their advantage.” Id. at 687. That does not bear even a passing resemblance to the markets here. No. 18-2852 79

The district court’s holding that Viamedia failed to offer evidence of forcing was premised on three assumptions: first, that the refusal-to-deal claim was correctly dismissed; second, that Viamedia acted as a reseller of Interconnect services ra- ther than as the MVPD’s agent for Interconnect access; and third, that RCN and WOW! could have purchased Intercon- nect-only access separately. Viewing the record as a whole, as we must, we disagree. We address each assumption in turn. First, the district court acknowledged Viamedia’s evi- dence that Interconnect access was conditioned on the pur- chase of Comcast’s ad rep services. Yet the court discounted this evidence because such a condition could be explained by Comcast’s “legal refusal to deal” rather than “an illegal ty- ing.” 335 F. Supp. 3d at 1061. This analysis fails on its own terms. As described above, Viamedia alleged a prima facie re- fusal-to-deal claim. Such potentially illegal conduct cannot justify Comcast’s related tying of Interconnect services to ad rep services, and more fundamentally, a tying claim does not fail as a matter of law simply because it was implemented by refusing to deal with an intermediary. Second, the district court conflated access to the coopera- tive Interconnects, formerly granted uniformly to competing MVPDs or their ad rep services agents, with the Interconnect services themselves. The district court’s conclusion that the MVPDs wanted a single entity to “make available to them both Interconnect Services and Ad Rep Services” is simply wrong as a matter of fact. Before Comcast excluded Viamedia, RCN and WOW! received Interconnect services from Com- cast and ad rep services separately from Viamedia. The district court misunderstood Viamedia to be a reseller of Interconnect services as part of a bundle that included ad rep services, and 80 No. 18-2852

concluded there was no impermissible tying simply because Viamedia “could not offer that bundle” any longer due to Comcast’s (supposedly) legal refusal to deal with Viamedia. 335 F. Supp. 3d at 1058 n.12. That characterization does not have a factual basis in the record, and it is certainly not be- yond reasonable factual dispute. Even Comcast characterized Viamedia not as a reseller of Interconnect services but as a re- seller of MVPD spot avails to the Interconnect operator. DA638. Comcast’s characterization of ad rep services—in- cluding its own ad rep services provided to customer/com- petitor MVPDs—as reselling spot avails and disconnected from its MVPDs customers’ best interests is troubling in its own right, but we return to that below. Neither the district court’s nor Comcast’s characterization of Viamedia’s role was accurate. As an ad rep services provider, Viamedia acted in the best interests of its MVPD customers and served as their agent or interface with the Comcast-controlled Interconnects for the one-third of MVPD spot avails that they sold cooperatively— not competitively—on a regional, DMA-wide basis. The dis- trict court acknowledged as much elsewhere in its opinion, explaining that MVPDs view “themselves as receiving Inter- connect Services from interconnect operators (like Comcast) even when they have hired an unaffiliated Ad Rep (like Vi- amedia) on a full-turnkey basis.” 335 F. Supp. 3d at 1058 n.10. We agree with the district court’s latter characterization. Third, in considering the summary judgment evidence, the district court drew inferences in favor of Comcast in con- cluding that RCN’s and WOW!’s purchases of Interconnect services and ad rep services from Comcast were not forced. It did not credit RCN’s and WOW!’s reasonable understanding No. 18-2852 81

that Comcast was tying Interconnect services to ad rep ser- vices. The court (1) read Comcast’s statements as more ambig- uous than they actually were; (2) concluded that, because the MVPDs ultimately purchased Comcast’s ad rep services, they must have wanted both services to be provided by Comcast; and (3) pointed to out-of-market evidence to infer that Com- cast would have offered an Interconnect-only deal if only RCN and WOW! had asked for one. This analysis departed from summary judgment standards in several respects. Even if Comcast’s statements had been ambiguous, a plaintiff does not need an express, written declaration of a proposed tying arrangement. A sale on the announced or im- plied condition that the buyer purchase the tied goods from the seller ordinarily satisfies the tying-agreement require- ment. Areeda & Hovenkamp ¶ 1754b–c, at 315–20. Although it is not enough for the services to be merely complementary, a seller is not immunized from a tying claim if there is a fac- tual dispute as to whether the buyer wished to purchase the tied service (here ad rep services) from the defendant with market power in the tying service market (here Interconnect services). The MVPDs’ ultimate decisions, after much finan- cial pain, to sign with Comcast for ad rep services do not dis- prove an illegal tie. (And notably, in Hartford, Frontier has continued to resist signing with Comcast for ad rep services and remains cut off from the Interconnects.) After all, “the great majority of tying and exclusive-dealing provisions that exclude rivals are engaged in by one market-dominating party and one party that is ‘innocent’ in the sense that it can- not profit from monopoly in the market, but is agreeing to the exclusivity only at the behest of the other party.” Areeda & Hovenkamp ¶ 1803a, at 107 n.5. 82 No. 18-2852

Finally, the district court incorrectly inferred from the fact that Comcast offers Interconnect-only access in other local markets that RCN and WOW! could have obtained Intercon- nect-only access if only they had asked. That reasoning denies the non-moving party the benefit of reasonable inferences from the evidence. It also overlooks the evidence showing that in-house provision of ad rep services simply was not a practical option for RCN or WOW! in these markets. More specifically, the evidence from other markets actu- ally supports Viamedia’s case, not Comcast’s defense. It is un- disputed that competition takes place within metropolitan (DMA) markets. If an advertiser wishes to purchase advertis- ing time in the Chicago DMA, buying a spot avail in New Or- leans is not an adequate substitute. Similarly, if Comcast was not permitting RCN an Interconnect-only deal in Chicago, an Interconnect-only deal in Denver was not an adequate substi- tute. And we summarized above at pages 24–35 the evidence indicating that RCN and WOW! did not willingly agree to Comcast’s terms. The Interconnect-only arrangements that Comcast offered in other DMAs, where it did not have (or exercise) so much market power, help Viamedia. Those arrangements show that Interconnect and ad rep services are indeed separate prod- ucts. They need not be sold together. At the same time, Com- cast’s willingness to offer Interconnect-only access in other DMAs may reflect that Comcast does not have an overlapping footprint—i.e., does not compete—in those markets. In con- trast, the evidence could support a finding that in Chicago, Detroit, and Hartford, Comcast tied Interconnect services to ad rep services to exclude its competitor in ad rep services and thereby force its MVPD competitors into its not-so-tender No. 18-2852 83

arms. See Areeda & Hovenkamp ¶ 1744g, at 198–99 (explain- ing that bundling in non-competitive markets does not neces- sarily provide insight into whether a tie is efficient rather than reflective of increased market power exploitation, while un- bundling in competitive markets likely reflects efficiencies). In drawing inferences in favor of the non-moving party, we must also recognize the possibility that Comcast was testing the waters in Chicago, Detroit, and Hartford with an eye to- ward expanding its tying demands to other markets if it is not held accountable.18 Outside of the district court’s inconclusive comparison of other, more competitive markets, there was no basis in the record to support Comcast’s speculation that if RCN and WOW! had just asked once more, Comcast would have aban- doned its strategic plan and agreed to standalone Chicago and Detroit Interconnect access. To make such an inference, we would have to assume that the MVPDs went several years without access to the Interconnects, lost millions of dollars in advertising revenue, complained to Comcast and federal reg- ulators, and then chose the ad rep services provider that they least preferred, all because of a mere misunderstanding.

18 The partial dissent incorrectly looks to arrangements in other geo- graphic markets to draw conclusions about forcing in the relevant mar- kets. Post at 132–33. The ad rep services and Interconnect markets are dis- tinct for each DMA. Viamedia’s tying allegations and evidence focused on Comcast’s conduct in Chicago, Detroit, and Hartford. The fact that Com- cast did not simultaneously monopolize or attempt to monopolize ad rep services in other geographic markets is not a defense to its monopolization of Chicago, Detroit, and Hartford. The dissent’s only support for its novel market-aggregating approach comes from a calculation of profit sacrifice in Novell, 731 F.3d at 1077, which has nothing to do with market definition or tying. See post at 133. 84 No. 18-2852

Comcast is free to offer this misunderstanding theory at trial, but the theory cannot support summary judgment. We re- verse summary judgment for Comcast on Viamedia’s tying claim. C. Section 2 Monopolization: Harms, Efficiencies, & Remedies Undisputed geographic markets, service markets, and market power make this case unusual. In addition, Viamedia has offered sufficient evidence to demonstrate prima facie claims for monopolization of the ad rep service markets in three DMAs through refusal to deal and tying. Comcast’s ac- tions also forced a new, intimate, and unwelcome relationship upon its smaller MVPD competitors. If credited, that evidence will shift to Comcast the burden to prove what would need to be some dazzling procompetitive benefits to justify its con- duct. We set out below considerations for the district court to consider in the rule of reason analysis it will have to conduct on remand. 1. Harm to Competition The potential harm in this case from Comcast’s refusal to deal and tying ripples outward. Prior to Comcast’s conduct, there was competition in three related markets: (1) between Comcast and other MVPDs for subscribers; (2) between Vi- amedia and Comcast in ad rep services; and (3) between Com- cast and other MVPDs for the sale of spot avails in the local DMA market. By forcing out its only competitor in the market for ad rep services and forcing its MVPD competitors to turn over 100% of their spot avails, Comcast eliminated competi- tion in the market for ad rep services and the market for the sale of local spot avails. At the same time, it gained the ability No. 18-2852 85

to impair competition in the market for MVPD subscribers. These harms to competition are prototypical antitrust harms. a. Ad Rep Services The ad rep services market went from two service provid- ers to a single, monopolist provider. With no other ad rep ser- vices providers, the elimination of Viamedia unquestionably harmed competition. The harms that typically flow from a competitive market shifting to total control by a monopolist include potentially higher prices, lower output, and reduced innovation. The market at issue here may not have had time to show ultimate effects from total foreclosure. But Comcast is forthright about the fact that it has refused to allow Viame- dia entry into other DMAs. Appellees’ Br. at 35–36. The DMAs that have never had a competitive ad rep services market may provide useful comparison points to the relevant DMAs here, which are still governed by contracts that were signed while Viamedia was still trying to compete with Comcast. b. The MVPD Market: MVPDs, Advertisers, Cable Subscribers Recall that the MVPDs faced three possible scenarios when Comcast refused to deal with Viamedia and then con- ditioned the MVPDs’ access to the Interconnects on the MVPDs turning over all of their spot avails to Comcast’s ad rep services arm. Above at pages 31–35. Each scenario en- tailed its own potential harm to competition. The MVPDs elected the first scenario, giving in to Comcast and signing up for its ad rep services. To summarize, this ar- rangement has resulted in:  Comcast’s smaller MVPD rivals contributing addi- tional revenue toward their dominant competitor; 86 No. 18-2852

 MVPDs turning over to Comcast the majority of spot avails they had formerly kept out of the Inter- connects, a large portion of which are allocated to local ads (for which the MVPDs formerly competed against each other);  RCN and WOW! being forced to trust that their dominant, incumbent cable rival Comcast will make ad sales decisions in the smaller competitors’ best interests, despite Comcast’s divided loyal- ties—as it is more profitable for Comcast to place ads in ways that are different from how RCN and WOW! might allocate them;  Comcast gaining access to its competitor MVPDs’ competitively sensitive information, including number and location of subscribers, ad sales, pro- motional ad materials to current and potential sub- scribers (including promotions trying to get cus- tomers to switch away from Comcast);  And it is not just the MVPDs that have been harmed. Comcast is now the only seller of spot avails in the local market. As Viamedia’s counsel noted at oral argument, local advertisers used to have several outlets to choose among at various price points when buying spot avails. Comcast is now their only-stop-shop, as well. And recall that this was the MVPDs’ least-bad choice—the one they chose when Comcast denied them their best option. Comcast and the district court hypothesized that Comcast was not actually barring competing MVPDs completely from Interconnect access. The hypothesis was instead that Comcast No. 18-2852 87

would have permitted the MVPDs to have Interconnect-only access without using either Viamedia’s or Comcast’s ad rep services. In this scenario, Comcast would lose ad rep services revenue, but would not lose millions from forgoing Intercon- nect access fees and from its own spot avails being less valua- ble when placed through a degraded Interconnect. But contrary to what the district court seemed to assume, this result still would have been harmful. Forcing RCN and WOW! to forsake the benefits they had gained by outsourcing ad sales to an independent Viamedia would have dramati- cally raised their costs. RCN and WOW! would have needed to hire staff, purchase technology, and pay for services that Viamedia had previously provided at lower cost. Those fixed costs would have been difficult for those MVPDs to afford, as shown by the fact that the option was always available in the- ory, and they never took it, and as shown by RCN’s filing with the FCC and WOW!’s internal emails. A215 n.81. In the other possible scenario, RCN and WOW! could con- tinue working with Viamedia for ad rep services, but Comcast would continue barring them from the cooperative Intercon- nects. If the MVPD competitors had made that choice, they would have remained cut off from a large percentage of their advertising revenue. Any loss of revenue or higher costs from these scenarios is not just a loss for competitors. It leads to a negative feedback loop in the market in which the MVPDs compete for cable subscribers, further harming competition. Higher costs and less advertising revenue lead to fewer promotional offers to subscribers and reduced expansion. See FCC 2007 Report and Order at 8 ¶ 13 (“Revenues from cable services are, in fact, a driver for broadband deployment,” i.e., the build-out of 88 No. 18-2852

additional cable infrastructure). This in turn hampers RCN and WOW! from obtaining new cable subscribers or retaining the subscribers they already have (who may switch to Com- cast). With a weakened RCN and WOW!, Comcast benefits further by not needing to offer the promotions it otherwise would have if it faced a more vibrant RCN or WOW! To the extent that cable subscribers are left with higher priced and lower quality services and competition has been eliminated in the market for the sale of local spot avails, a trier of fact would have to account for that additional anticompet- itive harm. c. Back to the Interconnects Comcast’s conduct also turned a previously procompeti- tive platform into a weapon to decrease competition in related markets. As originally conceived and implemented, the Inter- connects appear to have been comfortably on the “reasona- ble” side of a rule of reason analysis for such cooperative ven- tures among competitors. See generally Broadcast Music, Inc. v. Columbia Broadcasting System, Inc., 441 U.S. 1 (1979). Com- cast has described the Interconnects in these procompetitive terms, as a “collection of two or more cable TV systems that work together to distribute commercials to a wider geo- graphic area than a single system would otherwise reach, giv- ing advertisers the option to reach all cable households within a market with one buy.” First Am. Cplt. ¶ 156. Consistent with that description, Comcast itself has told the FCC that “the revenue share in an interconnect is often the same for all participants, and fairly standardized across interconnects.” Opposition to Petitions to Deny and Response to Comments at 279 n.883, In the Matter of Applications of Comcast Corp., Time Warner Cable Inc., et al. for Consent to Assign or Transfer Control No. 18-2852 89

of Licenses and Authorizations, FCC MB Docket 14-57 (Sept. 23, 2014) (Comcast 2014 FCC Response). Which is as it should be if this cooperative mechanism is procompetitive rather than a weapon to inflict anticompetitive harm. In congressional hearings on its proposed merger with TimeWarner Cable, Comcast highlighted the non-exclusion- ary nature of the Interconnects. A Comcast executive was sent to testify to a congressional committee about the proposed purchase. Much of the concern focused on the potential power of a combined Comcast/TimeWarner Cable to harm other content providers who relied on access to cable compa- nies’ “pipes” into the home (e.g., Netflix, YouTube). One Rep- resentative, however, asked about Comcast’s actions with the Interconnects it controlled. Comcast executive vice president David Cohen was asked to “provide assurances that Comcast will not exclude compet- itors or advertising firms from the advertising interconnects that Comcast operates.” Cohen replied, “We are not in the business of excluding businesses who want to buy advertising from us.” Cohen was again pressed: “So your short answer is that you are not going to exclude competitors or advertising … [f]rom the interconnects.” “Correct.” According to Viame- dia’s evidence, however, Comcast was in the midst of doing just what Cohen was denying.19

19 Comcast abandoned the Time/Warner transaction after investiga- tions by federal enforcement and regulatory agencies. The government’s “significant concerns” about the merger were focused on the likelihood that it “would make Comcast an unavoidable gatekeeper for Internet- based services that rely on a broadband connection to reach consumers.” DOJ Press Release, Comcast Corporation Abandons Proposed Acquisition of Time Warner Cable After Justice Department and Federal Communications 90 No. 18-2852

In this lawsuit, moreover, Comcast now argues that its control over the cooperative Interconnects is the source of its competitive advantage. If that were correct, it would call into question the legality of the Interconnects themselves. Anti- trust law is rightly skeptical of mechanisms that permit com- petitors jointly to set prices and other terms of dealing. Col- laboration between actual or potential competitors “can be rife with opportunities for anticompetitive activity.” American Society of Mech. Engineers, Inc. v. Hydrolevel Corp., 456 U.S. 556, 571 (1982). Based on Comcast’s portrayal and use of the Interconnects in this suit, such skepticism is now warranted. The govern- ment antitrust enforcement agencies provide guidance on competitor collaborations. A number of the factors that show anticompetitive effect appear to be met by the Interconnects as now portrayed by Comcast. The factors include whether the collaboration may:  “[L]imit independent decision making or combine the control of or financial interests in production, key assets, or decisions regarding price, output, or other competitively sensitive variables;”  “[O]therwise reduce the participants’ ability or in- centive to compete independently;”  Potentially “facilitate[] explicit or tacit collusion through facilitating practices such as the exchange

Commission Informed Parties of Concerns (Apr. 24, 2015), at http://www.jus- tice.gov/opa/pr/comcast-corporation-abandons-proposed-acquisition- time-warner-cable-after-justice-department (last visited on Feb. 21, 2020). No. 18-2852 91

or disclosure of competitively sensitive information or through increased market concentration;”  “Successfully eliminate[] procompetitive pre-col- laboration conduct, such as withholding services that were desired by consumers when offered in a competitive market.” Federal Trade Commission and United States Department of Justice, Antitrust Guidelines for Collaborations Among Competi- tors § 2.2, at 6, § 3.31, at 12 (April 2000) (Collaboration Guide- lines). Check, check, check, and check. These red flags were not raised by the truly cooperative original concept of the Interconnects. Having taken control of at least some Interconnects, though, Comcast now has the ability—and now even claims the right—to use the mecha- nism as a source of its competitive advantage over rivals, dis- torting competition in related markets. The use of the Inter- connects to take control of and set prices for competitors’ local ads does not appear related to accomplishing the Intercon- nects’ procompetitive goals. These facts weigh against Com- cast, not for it. 2. Procompetitive Justifications? The potential harms stemming from Comcast’s conduct will not lead to Section 2 liability if Comcast proves that its monopoly in ad rep services “is a consequence of a superior product, business acumen, or historic accident,” or if its con- duct was the result of, or necessary to achieve, much greater procompetitive benefits. United States v. Grinnell Corp., 384 U.S. 563, 571 (1966); see also Areeda & Hovenkamp ¶ 650c, at 94–95. The procompetitive benefits typically recognized in an- titrust law include evidence of “higher output, improved 92 No. 18-2852

product quality, energetic market penetration, successful re- search and development, cost-reducing innovations, and the like.” Areeda & Hovenkamp ¶ 651d, at 119.20 As this issue must also be decided by a trier of fact on remand, we offer some observations. a. The Interconnects We start with Comcast’s monopoly control over the Inter- connects in Chicago, Detroit, and Hartford, which Comcast has identified as the source of its competitive advantage that permitted it, in turn, to gain monopoly control over ad rep services. It appears that the only skill and foresight demon- strated by Comcast in obtaining monopoly control over the Interconnects was its ability to acquire a multitude of other cable MVPD providers without facing a challenge from gov- ernment antitrust enforcers. Comcast’s acquisitions are not in and of themselves evidence of superior skills, services, or ac- cident. What Comcast now identifies as a source of competitive advantage was produced by the kind of mergers that the agency Merger Guidelines describe as anticompetitive. The Guidelines identify as their “unifying theme” the proposition that “mergers should not be permitted to create, enhance, or

20“As a general matter, the evidence supporting a prima facie case need not be as specific as the evidence supporting a procompetitive justi- fication” because “[i]f the defendants have a procompetitive justification, it must have been a motivating factor for the restraint, and the defendants should be able to establish it rather easily.” Herbert J. Hovenkamp, The Rule of Reason, 70 Fla. L. Rev. 81, 107 (2018); see also id. at 110 (“To the extent that the defendants’ expectation of profit came from something other than a restriction of competition, they should have evidence and are in the best position to provide it.”). No. 18-2852 93

entrench market power or facilitate its exercise” by “enhanc- ing [a firm’s] market power.” § 1, at 2. One example of an an- ticompetitive merger is directly on point: Merging Firms A and B operate in a market in which network effects are significant, implying that any firm’s product is significantly more val- uable if it commands a large market share or if it is interconnected with others that in aggregate command such a share. Prior to the merger, they and their rivals voluntarily interconnect with one another. The merger would create an entity with a large enough share that a strategy of ending voluntary interconnection would have a dangerous probability of creating monopoly power in this market. The interests of rivals and consum- ers would be broadly aligned in preventing such a merger. Merger Guidelines § 2, at 6 (emphasis added). Comcast can- not now justify exclusionary conduct by pointing to control over the Interconnects, which was acquired through mergers that themselves may have been anticompetitive precisely be- cause of the risk that they could enable Comcast’s exclusionary con- duct. b. The Ad Rep Services Market Comcast’s new monopoly position in the ad rep services markets in Chicago, Detroit, and Hartford, if we draw reason- able inferences in favor of Viamedia, is a result not of its su- perior services but of its exclusionary conduct. Any claimed benefits from that conduct must be procompetitive and not simply the result of eliminating competition. For example, if 94 No. 18-2852

Comcast has reduced advertising, promotions, or other incen- tives that it previously offered to customers or local retailers when competing with Viamedia, those savings would repre- sent harm to competition. See Covad Communications Co. v. Bell Atlantic Corp., 398 F.3d 666, 674 (D.C. Cir. 2005), citing Bork, Antitrust Paradox at 314. Similarly, any savings gained by for- going investments in research and development, infrastruc- ture, or sales personnel, that otherwise would have been made under competitive conditions are properly categorized as harm to competition, not benefits. Finally, any defense premised upon the proposition that competition itself is inef- ficient, unreasonable, or confusing is not cognizable. See Na- tional Society of Professional of Engineers v. United States, 435 U.S. 679, 696 (1978). Viewing the facts in the light most reasonably favorable to Viamedia, Comcast’s procompetitive justifications seem to fall into this latter category. If Comcast could offer improved efficiencies by offering ad rep services and Interconnect ser- vices together to MVPDs, it was always free to do so. If this were the case, it could have passed on some of those savings to MVPD customers and possibly outcompeted Viamedia. Re- fusing to deal with the MVPD’s representative of choice ap- pears to be an attempt to avoid competition on the merits in the markets for ad rep services. On the procompetitive side of the ledger, evidence of re- duced pricing could offset harm—although current contracts may not yet reflect ultimate post-competition pricing. They were signed when Comcast was still bidding against Viame- dia. And while protection against free-riding is generally rec- ognized as a procompetitive goal, “[w]hen payment is possi- ble, free-riding is not a problem because the ‘ride’ is not free.” No. 18-2852 95

Chicago Prof’l Sports Ltd. Partnership v. NBA, 961 F.2d 667, 675 (7th Cir. 1992); see also Hovenkamp, The Rule of Reason, 70 Fla. L. Rev. at 111, 113 (“Often instances of claimed free riding are really complaints about competition, particularly when there are joint costs,” and “complete market exclusion is a suspi- ciously excessive remedy for claimed free riding, even where a certain amount of free riding actually occurs.”). If Comcast has evidence of truly procompetitive benefits, it should sub- mit that evidence to the trier of fact. But the hypotheses it has offered thus far do not entitle it to summary judgment. 3. Remedies Comcast’s final defense focuses on the challenge of reme- dying its conduct. After all, Comcast points out, “Courts are ill suited ‘to act as central planners, identifying the proper price, quantity, and other terms of dealing.’” Pacific Bell Tele- phone Co. v. Linkline Communications, Inc., 555 U.S. 438, 452 (2009), quoting Verizon Communications Inc. v. Law Offices of Curtis V. Trinko, 540 U.S. 398, 408 (2004). Call this the “So what?” defense. We agree that a court should not “impose a duty to deal that it cannot explain or adequately and reasonably super- vise.” Linkline, 555 U.S. at 452–53, quoting Trinko, 540 U.S. at 415. Yet courts are often called upon to undertake compli- cated, long-term supervision of complex cases and remedies. The judiciary need not and should not adopt a posture of learned helplessness in the face of proven antitrust violations. For example, courts regularly preside over dozens, if not hun- dreds or even thousands, of related cases in multidistrict liti- gation that present complicated questions of liability, not to mention supervising and implementing remedies over years if not decades. See, e.g., MDL 875 In re: Asbestos Products 96 No. 18-2852

Liability Litigation (No. VI) (overseeing thousands of asbestos cases, including class actions, since 1991). Courts oversee the bankruptcy process for companies with complicated corpo- rate structures and far-flung assets, supervising sales of those assets worth hundreds of billions of dollars. See, e.g., In re Leh- man Brothers Holdings, Inc., Case No. 08-13555 (Bankr. S.D.N.Y.). And of course federal courts currently supervise at least 115,000 individuals on supervised release, implementing difficult-to-monitor and intrusive conditions limiting those individuals’ jobs, the family and friends they can see, their drug and alcohol consumption, and locations to which they can travel, among others—all enforced through routine inter- views of family and associates, electronic monitoring, drug tests, and random searches. Administrative Office of the United States Courts, Overview of Probation and Supervised Re- lease Conditions 42–93 (Nov. 2016). In this antitrust case, we are not yet ready to cry “uncle” by affirming dismissal based on the unsubstantiated claim that this case poses “insoluble ad- ministrability problems.” In any event, this defense puts the cart before the horse. The trier of fact must first evaluate the evidence and deter- mine whether Comcast’s procompetitive justifications out- weigh the anticompetitive harms from its conduct. If Comcast is found liable, the district court will then face a decision about appropriate remedies. That will be the time to face the practical problems Comcast hypothesizes. The record thus far offers reasons to think the problems would be manageable. Comcast and Viamedia did business voluntarily, presumably on profitable terms for both. That history may well simplify the problems. As then-Judge Gor- such wrote for the Tenth Circuit in Novell, evidence of that No. 18-2852 97

earlier course of dealing “helps address, at least to some de- gree, administrability concerns—presumably profitable terms already agreed to by the parties may suggest terms a court can use to fashion a remedial order without having to cook them up on its own.” 731 F.3d at 1075. Moreover, Comcast itself told the FCC that this should not be an insoluble problem. It told the FCC that “the revenue share in an interconnect is often the same for all participants, and fairly standardized across interconnects.” Comcast 2014 FCC Response at 279 n.883. Such comparison points (both within a given Interconnect and against other Interconnects in which Comcast sells or buys Interconnect-only access) may be used to establish a remedy that addresses pricing. That being said, “Antitrust courts normally avoid direct price administration, relying on rules and remedies … that are easier to administer.” Linkline, 555 U.S. at 453, quoting Concord v. Boston Edison Co., 915 F.2d 17, 25 (1st Cir. 1990) (Breyer, C.J.). If a pricing remedy proves too complicated, then structural remedies may be preferable. See, e.g., Areeda & Hovenkamp ¶ 600b, at 4 (even “[m]ildly reprehensible be- havior might be enough to challenge a firm whose power is significant” and could justify imposing a more substantial remedy (e.g., “divestiture or dissolution” versus “an injunc- tion”)). Comcast knows that courts are capable of overseeing structural and behavioral remedies (including ones that gov- ern pricing disputes) that ameliorate competitive concerns. It agreed to such an arrangement as a condition for court ap- proval of its challenged 2009 merger with NBC Universal. United States v. Comcast Corp., 808 F. Supp. 2d 145, 147–48 (D.D.C. 2011) (Comcast agreeing to certain remedies to 98 No. 18-2852

prevent anticompetitive conduct post-merger). Such court- imposed and court-supervised remedies can also be imposed without a defendant’s consent after a finding of liability. One obvious possibility would be to prohibit Comcast’s control over Interconnects, which, in light of the evidence of misuse of that power to harm competition, raises serious problems under Section 1 of the Sherman Act.21 D. Antitrust Injury A private civil plaintiff in an antitrust case must also es- tablish “antitrust injury,” which requires proof that its “claimed injuries are of the type the antitrust laws were in- tended to prevent and reflect the anticompetitive effect of ei- ther the violation or of anticompetitive acts made possible by the violation.” Kochert v. Greater Lafayette Health Servs., Inc., 463 F.3d 710, 716 (7th Cir. 2006) (internal quotation marks and citations omitted). Viamedia has offered evidence of antitrust injury. When a monopolist creates a monopoly in the tied market, rivals are often excluded from the market, thereby losing mar- ket share or sales. “In such cases courts ordinarily grant stand- ing to the excluded or impeded rival.” Areeda & Hovenkamp ¶ 1767a, at 449; see also Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 490 n.14 (1977) (“[C]ompetitors may be able to prove injury before they actually are driven from the

21An amicus brief filed in support of Comcast argued that any “relief for arbitrary refusals to deal should be left to the legislature.” Brief for the Chamber of Commerce of the United States of America as Amicus Curiae in Support of Appellees at 25 (internal quotation marks omitted). The Chamber’s argument that courts should not enforce Section 2 in refusal- to-deal claims is a policy position—and not one with which we are free to agree. No. 18-2852 99

market and competition is thereby lessened. Of course, the case for relief will be strongest where competition has been diminished.”). We have also often recognized that competi- tors suffer antitrust injury when they are forced from the mar- ket by exclusionary conduct. See, e.g., Tri-Gen Inc. v. Intʹl Un- ion of Operating Engineers, Local 150, AFL-CIO, 433 F.3d 1024, 1032 (7th Cir. 2006) (“[T]his Court has recognized that com- petitors can bring an antitrust claim when they are excluded from the market and injured by defendants’ actions.”); Serfecz v. Jewel Food Stores, 67 F.3d 591, 597 (7th Cir. 1995) (“When the plaintiff’s injury is linked to the injury inflicted upon the mar- ket, such as when consumers pay higher prices because of a market monopoly or when a competitor is forced out of the market, the compensation of the injured party promotes the desig- nated purpose of the antitrust law—the preservation of com- petition.”) (emphasis added). This rule is integral to an effec- tive antitrust regime because “the foreclosed rival’s injury is entirely independent of the amount or existence of any injury to buyers.” Areeda & Hovenkamp ¶ 1767a, at 449. The general rule is that customers and competitors in the affected market have antitrust standing. See Associated General Contractors, 459 U.S. at 539; McGarry & McGarry, LLC v. Bank- ruptcy Mgmt. Solutions, Inc., 937 F.3d 1056, 1065–66 (7th Cir. 2019), citing In re Aluminum Warehousing Antitrust Litig., 833 F.3d 151, 158 (2d Cir. 2016), citing in turn Serpa Corp. v. McWane, Inc., 199 F.3d 6, 10 (1st Cir. 1999) (“Competitors and consumers in the market where trade is allegedly restrained are presumptively the proper plaintiffs to allege antitrust in- jury.”), and SAS of P.R., Inc., v. P.R. Tel. Co., 48 F.3d 39, 45 (1st Cir. 1995) (“competitors and consumers are favored plaintiffs in antitrust cases”). Viamedia is not seeking relief based on a theory that competition should have been reduced. Cf. 100 No. 18-2852

Brunswick Corp. v. Pueblo Bowl-O-Mat, 429 U.S. at 488 (rejecting claim for loss of income that would have been earned if other competitors had been forced out of the market and competi- tion had thus been reduced). Instead, Viamedia seeks only an opportunity for fair competition in the ad rep services mar- kets, based on the quality and prices of its services. It is an appropriate plaintiff to seek damages based on exclusionary conduct that forced it out of that market. Viamedia claims that Comcast’s exclusionary conduct drove it from the ad rep services markets in Chicago, Detroit, and Hartford, thus reducing competition. Viamedia has pre- sented evidence indicating that if Comcast not tied its sale of Interconnect services to ad rep services, RCN and WOW! likely would have continued to obtain ad rep services from Viamedia. The harm to competition is particularly pro- nounced since Viamedia was Comcast’s only competitor in in the relevant markets. To the extent that Comcast engaged in exclusionary conduct, the evidence indicates that the exclu- sionary conduct caused Viamedia’s injuries by forcing it from the ad rep services markets. The partial dissent expresses great skepticism toward ri- vals’ antitrust suits and argues that Viamedia cannot show an- titrust injury on its tying claim. As a general matter, caution is appropriate. The partial dissent goes astray, however, by con- tending it is “uncommon” for a “single foreclosed rival” to have standing in a tying case. Post at 122–23. The leading ty- ing cases and the leading treatise’s specific, on-point discus- sion of antitrust standing in tying cases teach otherwise. The foundational tying cases of the past 40 years were brought by tied-market rivals. Eastman Kodak, 504 U.S. 451; Jef- ferson Parish, 466 U.S. 2. The partial dissent cites no case law No. 18-2852 101

for its assertion that such cases are “uncommon.” It also does not cite any case in which an excluded rival in the tied market was found to lack antitrust injury. In the single tying case cited in the partial dissent’s discussion of antitrust injury, the plaintiff lacked antitrust injury precisely because she was not participating in the tied market when the tie was allegedly implemented. See Kochert, 463 F.3d at 716. And where Areeda and Hovencamp address antitrust standing in tying cases, they recognize that standing is appro- priate in cases like this. Viamedia was competing with Com- cast in the tied market for ad rep services. It was forced out of that market in Chicago, Detroit, and Hartford. Areeda and Hovencamp explain: Rivals in that market [the tied market] may be “foreclosed” when their entry or expansion is impeded or they lose existing market share or sales. Consumers lose the benefits of any entry, expansion, competition, or innovation that in- dependent rivals might have injected into the tied market. In such cases courts ordinarily grant standing to the excluded or impeded rival. ¶1767a, at 449–50 (emphasis added); accord, Eastman Kodak, 504 U.S. at 479 (reversing summary judgment for defendant: market foreclosure in tied market resulting from illegal tying “is facially anticompetitive and exactly the harm that antitrust laws aim to prevent”). To the extent that the partial dissent suggests that the tying conduct at issue is not illegal or exclu- sionary, that goes to the merits, not to antitrust injury. The partial dissent also suggests that Viamedia’s injuries cannot establish antitrust injury because they are the same as 102 No. 18-2852

those alleged from the refusal to deal. Post at 125. This point also reflects only disagreement on the merits of the tying claim and the confusion that has stemmed from the different treatment of the two claims in the district court. Because the same general course of conduct supports both the refusal-to- deal and tying claims, the two theories necessarily allege sim- ilar injuries and damages. E. Role of Expert Witnesses The final issue is the admissibility of expert testimony. We review the district court’s exclusion for an abuse of discretion. Salgado v. General Motors Corp., 150 F.3d 735, 739 (7th Cir. 1998). If a discretionary ruling is based on an error of law, though, it can often be deemed an abuse of discretion. E.g., Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 402 (1990) (Rule 11 sanctions); Ervin v. OS Restaurant Services, Inc., 632 F.3d 971, 976 (7th Cir. 2011) (class certification). The district court struck Viamedia’s expert testimony largely based on the view that Viamedia’s claims should fail as a matter of law. We disagree with the district court’s view of the law, so we reverse the court’s rulings regarding Viame- dia’s expert witnesses.22 We address separately the district court’s ruling regarding a portion of Viamedia’s expert eco- nomic witness, whom the court perceived as merely offering “lay” testimony, as well as an objection Comcast has raised on

22 Viamedia offered Dr. Lys as an expert on damages issues. The only basis for excluding his opinion was that he assumed that Comcast’s con- duct had violated the antitrust laws and thus caused Viamedia cognizable harm, and the court had concluded that Viamedia could not legally pre- vail on its antitrust claims. As we are reversing the district court’s legal rulings, the court’s exclusion of Dr. Lys’ expert opinion must also be re- versed. No. 18-2852 103

appeal that Viamedia failed to offer a causation expert, which Comcast believes should be fatal. 1. Standard Expert opinion testimony is admissible if “(a) the expert’s scientific, technical, or other specialized knowledge will help the trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts or data; (c) the testimony is the product of reliable principles and methods; and (d) the expert has reliably applied the principles and methods to the facts of the case.” Fed. R. Evid. 702. An expert’s opinion may “overlap[] with the jurors’ own experi- ences” or “cover matters that are within the average juror’s comprehension,” so long as the expert uses some kind of “specialized knowledge” to place the litigated events “into context.” Lawson v. Trowbridge, 153 F.3d 368, 376 (7th Cir. 1998) (citations omitted); see also United States v. Williams, 81 F.3d 1434, 1441 (7th Cir. 1996) (“All you need to be an expert witness is a body of specialized knowledge that can be helpful to the jury.”). 2. Economic Expert Witness Viamedia offered Dr. Furchtgott-Roth as an expert witness on the economic rationales of Comcast’s conduct and the competitive ramifications from such conduct. Setting aside the district court’s ruling regarding Dr. Furchtgott-Roth’s re- port based on the court’s legal holdings, the court also ex- cluded a portion of his testimony as offering only a lay im- pression of the market and Comcast’s conduct. Viamedia argues that Dr. Furchtgott-Roth’s undisputed “specialized knowledge” would be helpful to a jury to place Comcast’s conduct into context, and that there are some 104 No. 18-2852

complex facts in this case, “including the economic incentives faced by a set of interrelated firms in the two-sided market for spot cable advertising,” which are not “obvious to the layper- son.” Dr. Furchtgott-Roth clearly drew conclusions through “ex- pert assessment,” not merely a lay interpretation of the evi- dence. While he did summarize and repeat some relevant facts, he drew significantly on expertise to “add something”— context and supporting information—to the record. Viamedia contends that its exclusion from the Interconnects furthered Comcast’s tying policy. Dr Furchtgott-Roth’s opinion that self-provision is not a viable business option for smaller MVPDs is an expert interpretation of evidence on a highly rel- evant factual point. So too is his opinion that Viamedia cannot make a competitive offer for ad rep services if Comcast con- ditions its competitor MVPDs’ Interconnect access on forgo- ing Viamedia’s services. He drew on his expertise to make these two determinations, both of which required analysis of market conditions. These opinions informed his broader opinion that Comcast’s exclusion of Viamedia from the Inter- connects was integral to its tying conduct. The district court therefore abused its discretion in concluding that Dr. Furcht- gott-Roth’s testimony was not significantly informed by his expertise. It was, and it therefore meets the requirement of Federal Rule of Evidence 702(a). 3. Lack of Expert Witness on Causation Comcast also argues that Viamedia’s case must fail be- cause it has not offered an expert on causation. Hiring another expert on causation is not a legal requirement for successfully bringing an antitrust case. Rather than requiring “§ 2 liability [to] turn on a plaintiff’s ability or inability to reconstruct the No. 18-2852 105

hypothetical marketplace absent a defendant’s anticompeti- tive conduct,” which “would only encourage monopolists to take more and earlier competitive action,” courts have in- ferred causation when a defendant’s conduct “reasonably ap- pear[s] capable of making a significant contribution to … maintaining monopoly power.” United States v. Microsoft Corp., 253 F.3d 34, 78–79 (D.C. Cir. 2001) (en banc) (citation omitted). To the extent there may be an underlying problem of proof, “the defendant is made to suffer the uncertain con- sequences of its own undesirable conduct,” and causation “queries go to questions of remedy, not liability.” Id. at 79–80 (citation omitted). Conclusion Viamedia alleged sufficiently, and at summary judgment offered sufficient evidence, that Comcast violated Section 2 of the Sherman Act. Viewing the allegations and evidence in the light most favorable to Viamedia, Comcast abruptly termi- nated decade-long, profitable agreements and sacrificed short-term profits to obtain and entrench long-term market power, and used its monopoly power in Interconnect services market to force its MVPD competitors into a relationship that makes Comcast a gatekeeper of its competitors’ advertising revenue. This conduct “reveal[s] a distinctly anticompetitive bent.” Trinko, 540 U.S. at 409, discussing Aspen Skiing, 472 U.S. 585. Comcast is free to contest these issues at trial, as well as to try to prove and quantify any procompetitive justifications. The factual disputes in this case are numerous, genuine, and material. The judgment of the district court is REVERSED and the case is REMANDED for further proceedings consistent with this opinion. 106 No. 18-2852

BRENNAN, Circuit Judge, concurring in part and dissenting in part. The majority opinion is synoptic in its coverage, deeply researched, and meticulous in its consideration of the antitrust issues this case presents. It deserves much respect. While I agree Viamedia has plausibly alleged an antitrust vi- olation and is entitled to reversal and remand on its refusal- to-deal claim, I would affirm summary judgment on its tying claim because the undisputed facts do not present evidence of an illegal tie. I also respectfully part company with my col- leagues on some other issues the majority opinion tackles. The last several decades have brought a new regime to an- titrust law in the world of exclusionary conduct. Outdated monopolization doctrines have given way to a sharper and narrower understanding of what constitutes exclusionary be- havior under § 2 of the Sherman Act, 15 U.S.C. § 2. See Novell, Inc. v. Microsoft Corp., 731 F.3d 1064, 1072 (10th Cir. 2013); United States v. Microsoft Corp., 253 F.3d 34, 49 (D.C. Cir. 2001). History teaches this new regime promotes competition and innovation in the marketplace, and it informs the resolution of the claims and defenses before us now. I. Refusal to Deal A 2003 agreement between Viamedia and Comcast granted Comcast the exclusive right to sell on its Chicago, De- troit, and Hartford Interconnects advertising availabilities that Viamedia purchased from multichannel video program- ming distributors (“MVPD”) WOW! and RCN. Viamedia un- derstood that, upon the agreement’s expiration, Comcast had the right to solicit RCN’s and WOW!’s advertising business directly. The agreement between Viamedia and Comcast ex- No. 18-2852 107

pired May 31, 2012. The next day Comcast informed Viame- dia of its intent not to renew their agreement and to seek RCN’s and WOW!’s business directly. Comcast “prefers to deal directly with MVPDs, rather than with intermediaries such as Viamedia, and has found substantial benefits from direct dealings.” Appellee’s Br. 11. When their contracts with Viamedia expired in 2015, RCN and WOW! contracted with Comcast to be their ad repre- sentative in Chicago and Detroit and sell their availabilities on its Interconnects. Up to that point, Viamedia’s agreements with RCN and WOW! prevented them from dealing directly with Comcast. Cut out of the deal, Viamedia sued Comcast for allegedly violating § 2 of the Sherman Act, claiming Comcast’s decision to end Viamedia’s access to the Interconnects lacked a valid business reason under Aspen Skiing Co. v. Aspen Highlands Skiing Corp., 472 U.S. 585 (1985). Specifically, Viamedia al- leged Comcast’s decision caused Comcast to forfeit fees up- front and reduced the economic value of the Interconnects. The allegations do not claim that Comcast’s sacrificed profits later led to monopoly recoupment. The district court dismissed Viamedia’s refusal-to-deal claim. Viamedia, Inc. v. Comcast Corp., 218 F. Supp. 3d 674 (N.D. Ill. 2016). The court found that Viamedia’s own allegations ad- mitted a valid business purpose for Comcast’s refusal: remov- ing an intermediary, Viamedia, to deal directly with MVPD customers.1 Id. at 698. The court further found that Viamedia

1 This court has characterized this common business practice of verti- cal integration or disintermediation as pro-competitive and efficient. See 108 No. 18-2852

failed to allege or explain how Comcast’s refusal to deal with it had “no rational procompetitive purpose.” Id. A. The Refusal-to-Deal Claim Survives a Motion to Dismiss. We review the grant of a Federal Rule of Civil Procedure 12(b)(6) motion to dismiss de novo and ask whether there is “plausibility in the complaint.” Christy Sports, LLC v. Deer Valley Resort Co., 555 F.3d 1188, 1191 (10th Cir. 2009) (citing Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 564 (2007)) (inter- nal citations omitted); see also Deppe v. NCAA, 893 F.3d 498, 500 (7th Cir. 2018). “A claim has facial plausibility when the plaintiff pleads factual content that allows the court to draw the reasonable inference that the defendant is liable for the misconduct alleged.” Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009). Plausibility does not ask whether the allegations are likely true; the court must assume they are. Instead, the inquiry is “whether, if the allegations are true, it is plausible and not merely possible that the plaintiff is entitled to relief.” Christy Sports, 555 F.3d at 1191–92. The plaintiff must plausibly allege more than “wholly conclusory statements” that a defendant violated § 2 of the Sherman Act to advance past the pleadings stage. Id. In considering the alleged violation at the root of Viame- dia’s refusal-to-deal claim, we must be cognizant that “the an- titrust laws rarely impose on firms—even dominant firms—a duty to deal with rivals.” Novell, 731 F.3d at 1066; see also Verizon Commc’ns v. Law Offices of Curtis V. Trinko, 540 U.S. 398,

Jack Walters & Sons Corp. v. Morton Bldg., Inc., 737 F.2d 698, 710 (7th Cir. 1984) (holding “vertical integration usually is procompetitive”). No. 18-2852 109

411 (2004) (holding even a monopolist has no duty to cooper- ate with rivals). “As a general rule purely unilateral conduct does not run afoul of section 2—businesses are free to choose whether or not to do business with others and free to assign what prices they hope to secure for their own products.” 731 F.3d at 1072. In the past, “some courts suggested that a mo- nopolist must lend smaller rivals a helping hand,” but today it is understood that forcing rivals to cooperate “usually leaves consumers paying more for less.” Id.; see also Frank H. Easterbrook, The Chicago School & Exclusionary Conduct, 31 HARV. J.L. & PUB. POL’Y 439, 441–42 (2008) (antitrust themes incentivizing cooperation among firms largely “bit the dust in Verizon v. Trinko” and now “the main goal of antitrust is to compel firms to be rivals”); Olympia Equip. Leasing v. Western Union Telegraph, 797 F.2d 370, 375 (7th Cir. 1986) (“Today it is clear that a firm with lawful monopoly power has no general duty to help its competitors, whether by holding a price um- brella over their heads or by otherwise pulling its competitive punches.”). Despite this hesitancy to condemn refusals to deal, a mo- nopolist’s behavior violates § 2 if it is “irrational but for its anticompetitive effect.” Novell, 731 F.3d at 1075; see also 3 Phillip E. Areeda & Herbert Hovenkamp, ANTITRUST LAW: AN ANALYSIS OF ANTITRUST PRINCIPLES AND THEIR APPLICATION ¶ 651b3, p. 107 (4th ed. 2015) (monopolizing conduct is “irra- tional” if “only explanation that makes it seem profitable is destruction or discipline of rivals”) (hereinafter Areeda & Hovenkamp, ANTITRUST LAW); Trinko, 540 U.S. at 407 (defend- ant must be seeking “an anticompetitive end”); Aspen Skiing, 472 U.S. at 605 (“If a firm has been attempting to exclude rivals on some basis other than efficiency, it is fair to characterize its behavior as predatory.”) (internal citations omitted); Christy 110 No. 18-2852

Sports, 555 F.3d at 1194 (“in rare circumstances a refusal to co- operate with competitors might constitute a § 2 violation”). It is unclear—particularly on a motion to dismiss when a plausibly alleged violation is all that is required—whether conduct “irrational but for its anticompetitive effect” is to be treated the same as conduct with “no rational procompetitive purpose.” See 218 F. Supp. 3d at 698. Although slight, there is a difference: the former provides an antitrust plaintiff the op- portunity to argue that, despite some efficiency justification proffered by an antitrust defendant, the rational or intended goal of the conduct was its anticompetitive impact. The latter, in contrast, requires the antitrust defendant to put forward any evidence of some business reason for its conduct, regard- less of potential anticompetitive effect. The district court ap- plied the latter and found there was a rational procompetitive purpose because Comcast offered evidence of vertical integra- tion and disintermediation—motives that I agree reflect law- ful and procompetitive marketplace conduct.2 But it did so by disregarding the plausibility of Viamedia’s allegations of anticompetitive conduct and weighing the evidence in Com- cast’s favor. This is not the court’s role on a Rule 12(b)(6) mo- tion. Any confusion here may stem from a misunderstanding of how to handle conflicting evidence of conduct in the alle- gations. If only one party advances evidence showing pro- competitive or anticompetitive conduct, the court may find

2 The majority opinion reviews vertical integration and disintermedi- ation cases cited by Comcast and concludes they are inapposite here on the question of procompetitive conduct. See Majority op. at pp. 63–67. While it does not change my agreement with the majority to remand the refusal-to-deal claim, I do not reach this same conclusion. No. 18-2852 111

the lack of any opposing evidence shows either rationality or irrationality for § 2 purposes. See Novell, 731 F.3d at 1076 (on review of judgment as a matter of law, finding no evidence from which a reasonable jury could conclude monopolist’s conduct was irrational); see also Aspen Skiing, 472 U.S. at 610 (on review of summary judgment, finding no evidence of ef- ficiency justification for the refusal to deal). But even these in- stances are post-pleading. What of the parties who, at the pleading stage, proffer allegations of competing economic justifications for behavior? Are we to accept the defendant’s proffered justification as conclusive of procompetitive ration- ality without considering the plaintiff’s allegations of anticompetitive conduct? If so, can an antitrust plaintiff ever advance past the pleadings stage when a defendant asserts a procompetitive justification? The district court effectively held the plaintiff cannot, 218 F. Supp. 3d at 698, but this is up for debate. See 3 Areeda & Hovenkamp, ANTITRUST LAW ¶ 651b3, pp. 106–07 (criticizing approach that relies on facts which “benefit the defendant very slightly while doing considerable harm to the rest of the economy;” “[n]ot all mo- nopolizing conduct that we might wish to condemn is ‘irra- tional’”); Microsoft Corp., 253 F.3d at 59 (defendant bears the burden of presenting a “nonpretextual claim” and proving procompetitive justification on the facts); Illinois ex rel. Burris v. Panhandle Eastern Pipe Line Co., 935 F.2d 1469, 1481–82 (7th Cir. 1991) (finding “the presence of a legitimate business jus- tification reduces the likelihood that the conduct will produce undesirable effects on the competitive process”; “[w]hether valid business reasons motivated a monopolist’s conduct is a question of fact” for a fact-finder); Olympia, 797 F.2d at 378 (reasoning the lack of “a clear business justification” among conflicting evidence “may indicate probable anticompetitive 112 No. 18-2852

effect” but is not conclusive). We need not settle this debate, though, as it does not impact the decision to remand. As the majority opinion recognizes, “the calculation of procompetitive benefits net of anticompetitive harms does not easily lend itself to a pleading standard.” Majority op. at p. 59. Viamedia alleged Comcast’s conduct could achieve “no pro- competitive justifications” because there were no “problems in allowing Viamedia to participate in Interconnects” on be- half of its MVPD customers. Viamedia’s allegations show more than market power; they allege Comcast’s exclusionary conduct was anticompetitive and harmful to the economic purpose of the Interconnects, see Christy Sports, 555 F.3d at 1192 (anticompetitive allegation must appear in the pleadings), specifically by denying Viamedia access to the In- terconnects. Giving Viamedia the benefit of its allegations, its refusal-to-deal claim clears the Rule 12(b)(6) pleading bar be- cause it plausibly alleges it was “excluded from the market and injured by defendant[‘s] actions,” Tri-Gen Inc. v. Int’l Un. of Op. Eng’s Local 150, 433 F.3d 1024, 1032 (7th Cir. 2006)―an alleged injury that “harms both competitors and competi- tion,” Cargill, Inc. v. Monfort of Colo., Inc., 479 U.S. 104, 118 (1986). Down the road, the facts Viamedia has pleaded or other facts discovered may render its refusal-to-deal claim a candidate for summary disposition. See Collins v. Associated Pathologists, Ltd., 844 F.2d 473, 475 (7th Cir. 1988) (considering various antitrust claims: “the very nature of antitrust litiga- tion encourages summary disposition of such cases when per- missible”). But viewing the facts in a light favorable to Viamedia, its refusal-to-deal claim has not reached that point. No. 18-2852 113

On a motion to dismiss, an antitrust plaintiff seeking § 2 damages must point to plausible allegations showing its ri- val’s refusal to deal was irrational but for its anticompetitive effect. Regardless of how this court in the future resolves competing justifications at the dismissal stage, Viamedia has plausibly alleged refusal to deal in violation of § 2 and was entitled to advance that claim beyond the pleadings. B. The Refusal-to-Deal Claim Is Different than the Claim in Aspen Skiing. The majority opinion concludes this case is indistinguish- able from Aspen Skiing, 472 U.S. 585. Majority op. at p. 54. I respectfully do not join the majority’s conclusion that Aspen Skiing “maps onto Comcast’s conduct,” id. at 44, or that on a refusal-to-deal theory this case “appears stronger than Aspen Skiing.” Id. at 52. The familiar holding of Aspen Skiing is that the defendant’s “failure to offer any efficiency justification whatever for its pattern of conduct” resulted in a § 2 violation. 472 U.S. at 608; see also Christy Sports, 555 F.3d at 1197 (finding the “critical fact in Aspen Skiing was that there were no valid business rea- sons for the refusal” to deal). Its holding is specific: the de- fendant could not justify either its insistence on breaking up the joint lift pass or its refusal to sell the other owner its passes for the same value just to keep the pass together. 472 U.S. at 603. There was no efficiency reason offered for the defend- ant’s conduct; the only apparent purpose was to eliminate competition. Id. at 608. The Supreme Court later explained that this conduct constituted a violation because the “unilat- eral termination of a voluntary (and thus presumably profit- able) course of dealing suggested a willingness to forsake short-term profits to achieve an anticompetitive end.” Trinko, 114 No. 18-2852

540 U.S. at 409. The defendant’s “unwillingness to renew the ticket even if compensated at retail price revealed a distinctly anticompetitive bent.” Id; see also Goldwasser v. Ameritech Corp., 222 F.3d 390, 398 (7th Cir. 2000) (explaining the decision “to forgo cash revenues and efficient methods of doing busi- ness for the sole purpose of driving its rival out of the market” is a § 2 violation). The application of Aspen Skiing’s holding has been the sub- ject of substantial debate. Albeit a seminal antitrust opinion, Aspen Skiing is recognized as a factual and legal exception un- der current antitrust law. See Trinko, 540 U.S. at 409 (“Aspen Skiing is at or near the outer boundary of § 2 liability.”). To- day, it fits within the “narrow world of refusal to deal cases.”3 Novell, 731 F.3d at 1079; see Easterbrook, The Chicago School & Exclusionary Conduct, at 441–42 (describing the Court’s hold- ing in Aspen Skiing as “the last gasp of the old school of anti- trust”). Not all refusal-to-deal challenges fall within the “limited exception” of Aspen Skiing. Novell, 731 F.3d at 1074–75. To in- voke that exception, there must be a preexisting, voluntary, and presumably profitable course of dealing between a mo- nopolist and a rival, and the discontinuation of that dealing must reveal irrational willingness to forsake short-term

3 See Gov’t’s Br. in Supp. of Neither Party at 6, 8–11, ECF No. 33 (arguing only limited circumstances have been recognized in which a mo- nopolist violates § 2 by refusing to deal with a rival, and those circum- stances are tightly circumscribed due to the negatives of coerced dealing; recommending this court follow Novell and its test that a refusal to deal does not violate § 2 “unless it would make no economic sense for the de- fendant but for its tendency to eliminate or lessen competition”). No. 18-2852 115

profits to achieve an anti-competitive end. Id. Evidence of for- saking short-term profits may “isolate conduct that has no possible efficiency justification.” Id. at 1077. When consider- ing sacrificed profits, though, courts should not “disaggregate profits from different lines of business” as “[p]arsing profits” would defeat the purpose of “holding firms liable for making moves that enhance their overall efficiency.” Id. (disaggregat- ing profits would make it difficult to assess firm’s goal of “maximizing overall profits” and is inconsistent with the Court’s reasoning in Aspen Skiing and Trinko); see also Christy Sports, 555 F.3d at 1194 (businesses have ability “to recoup [their] investment[s]” in any number of ways). In finding the Aspen Skiing exception applies here, the majority opinion points to three key factors underlying the Supreme Court’s decision in that case: (1) an important change in a pattern of distribution that had persisted for years; (2) conduct in the market with arrangements in comparable markets; and (3) forgoing profitable transactions. Majority op. at pp. 49–51. On these factors, I see Aspen Skiing and this case as different. First, in Aspen Skiing there was more than an “important change” in the distribution pattern, as in that case the joint pass was terminated altogether. 472 U.S. at 603. Here, no termination of the Interconnect occurred; instead, Comcast encouraged additional MVPD participation and sought to se- cure their access to the Interconnects by contracting with them directly. Second, in Aspen Skiing, the monopolist’s con- duct in comparable markets where it lacked dominance in- cluded the use of cooperative tickets in areas that apparently were competitive. Id. at 603–04. But Comcast participates in Interconnects across other DMAs, and, where it is the largest MVPD in a market, it operates the Interconnect there, too. 335 116 No. 18-2852

F. Supp. 3d at 1046. Third, in Aspen Skiing the forgoing of prof- itable transactions was not alone, but “in exchange for per- ceived long-run impact on its smaller rival,” and referenced in Aspen Skiing only in the context of finding no efficiency jus- tification for the refusal to deal. 472 U.S. at 608, 610. That is not the case here, as Comcast was able to proffer an efficiency justification (disintermediation and vertical integration) for its conduct. See Appellee’s Br. 3. Aspen Skiing would be more analogous to this case if a third-party vendor had managed the sales and advertising of the joint pass directly to skiers, and then Aspen Skiing Company (the monopolist) took over that role for vertical in- tegration or other efficiency reasons. Instead, Aspen Skiing Company terminated the joint pass altogether; that would be like Comcast terminating the Interconnects to create a sub-op- timal new platform to sell advertising, which of course did not happen here. Rather than terminate the Interconnects, Comcast encouraged MVPD participation and sought to se- cure MVPDs’ access to the Interconnects by contracting di- rectly with them. The record also contains evidence that Com- cast acted pursuant to a rational business purpose: Comcast claimed it sought vertical integration and disintermediation, and Viamedia admitted such an efficiency justification in its allegations. See id. at 3. Still, the reinstatement of Viamedia’s refusal-to-deal claim does not depend on Aspen Skiing. Because that claim survives the plausibility requirement we apply under Rule 12(b)(6), that portion of the judgment should be reversed and re- manded. No. 18-2852 117

C. Opinion Testimony on the Refusal-to-Deal Claim May Be Allowed. Lastly on the refusal-to-deal claim, the district court’s or- der regarding expert witnesses, which we review for abuse of discretion, is properly vacated. See Salgado ex rel. Salgado v. Gen. Motors Corp., 150 F.3d 735, 739 (7th Cir. 1998). On re- mand, the landscape of this case will have sufficiently changed to allow for this method of proof on the refusal-to- deal claim. This reversal allows for Viamedia to name expert witnesses on this claim, subject to the usual later motion prac- tice to exclude or limit their testimony on this cause of action. II. Tying After their agreements with Viamedia expired, RCN and WOW! sought exclusive, full-turnkey relationships with Comcast. Neither RCN nor WOW! ever sought Interconnect- only services. The record contains no evidence that Comcast has ever declined an MVPD’s request for Interconnect-only services; in fact, 14 percent of Comcast’s agreements with MVPDs across all DMAs since December 2011 were Interconnect-only agreements. Viamedia, Inc. v. Comcast Corp., 335 F. Supp. 3d 1036, 1058 (N.D. Ill. 2018). None of those Interconnect-only agreements prevent an MVPD from hiring another ad representative for local sales or conducting their own ad representation. Under its tying theory, Viamedia claimed Comcast condi- tioned access to the Interconnect (the “tying” product) on the purchase of Comcast’s ad representation services (the “tied” product). But for whom? Viamedia conceded that both RCN and WOW! maintained the ability to deal directly with Com- 118 No. 18-2852

cast and access the Interconnect without any ad representa- tive should they choose not to employ Comcast at the termi- nation of their agreements with Viamedia. The district court granted summary judgment on Viame- dia’s tying claim. 335 F. Supp. 3d at 1074. In so doing the court found there was no evidence that Comcast conditioned access to the Interconnect on the purchase of ad representation ser- vices. Id. at 1058 (finding the fact that RCN and WOW! both requested full-turnkey representation and that 14 percent of Comcast’s agreements with MVPDs are Interconnect-only “belies any inference that Comcast tied its services.”). The court further found there was no triable issue as to antitrust injury or damages, both necessary elements of an actionable claim. Notably, Viamedia admitted its injuries were “fully attributable to Comcast’s decision to deny Viamedia Intercon- nect access,” id. at 1070, thereby failing to establish a cogniza- ble antitrust injury separate from the refusal to deal. Viamedia likewise failed to separately prove damages caused by the al- leged tying conduct, again collapsing that showing into the refusal-to-deal claim. Id. at 1072–73. To the district court, the crux of Viamedia’s tying claim was that Comcast withheld the alleged tying product from its rival, Viamedia, not from its customers, WOW! and RCN. As the court found, Viamedia aimed to sell MVPDs a bundle of Comcast’s Interconnect services with Viamedia’s ad represen- tation services, but “Viamedia has no antitrust right to force Comcast to help it sell such a bundle to their mutual custom- ers.” Id. at 1064. Finding the undisputed evidence did not show tying conduct, the district court granted summary judg- ment. Id. at 1074. No. 18-2852 119

A. Antitrust Injury and Standing are Lacking on the Tying Claim. As a threshold matter, a plaintiff must have antitrust standing to bring an antitrust claim. See McGarry & McGarry v. Bankr. Man. Solut., 937 F.3d 1056, 1063 (7th Cir. 2019) (anti- trust standing required to identify “which plaintiffs may bring the cause of action”); see also Novell, 731 F.3d at 1080 (“[A] private party must establish some link between the de- fendant’s alleged anticompetitive conduct, on the one hand, and its injuries and the consumer’s, on the other.”). A show- ing of antitrust standing requires more than the standing in- quiry under Article III. See McGarry, 937 F.3d at 1063 (“[T]he Sherman Act has additional rules for determining whether the plaintiff is the proper party to bring a private antitrust ac- tion.”) (citing Associated Gen. Contractors of Cal., Inc. v. Cal. State Council of Carpenters, 459 U.S. 519, 535 n.31 (1983)); see also Kochert v. Greater Lafayette Health Servs., Inc., 463 F.3d 710, 716 (7th Cir. 2006) (“Antitrust standing requires more than the ‘injury in fact’ and the ‘case or controversy’ required by Arti- cle III of the Constitution.”). To establish antitrust standing, a plaintiff must first show it was injured by anticompetitive conduct. See McGarry, 937 F.3d at 1063–64. Antitrust standing is limited to “(1) those who have suffered the type of injury that the antitrust laws were intended to prevent and (2) those whose injuries are a result of the defendant’s unlawful conduct.” Id. (quoting Serfecz v. Jewel Food Stores, 67 F.3d 591, 595 (7th Cir. 1995)); see also Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 120 No. 18-2852

489 (1977) (antitrust plaintiff must be able to show that its in- jury is “of the type the antitrust laws were intended to pre- vent”). Even if a plaintiff can show its injuries are the result of un- lawful anticompetitive conduct, it must be able to “efficiently vindicate the purposes of the antitrust laws” to gain antitrust standing. Kochert, 463 F.3d at 716; see also Associated Gen. Contractors, 459 U.S. at 537–44 (plaintiff’s ability to efficiently vindicate the law’s purpose confers antitrust standing if shown antitrust injury); In re Industrial Gas Antitrust Litig., 681 F.2d 514, 526 (7th Cir. 1982) (“not all persons who have suf- fered an injury flowing from [an] antitrust violation have standing to sue”). Simply, the successful antitrust plaintiff must prove both antitrust injury and antitrust standing. See Kochert, 463 F.3d at 716. “We usually presume that competitors and consumers in the relevant market are the only parties who suffer antitrust injuries and are in a position to efficiently vindicate the anti- trust laws.” McGarry, 937 F.3d at 1065 (citing Associated Gen. Contractors, 459 U.S. at 538; In re Aluminum Warehousing Anti- trust Litig., 833 F.3d 151, 158 (2d Cir. 2016)). “But ‘presump- tively’ does not mean always,” SAS of Puerto Rico v. Puerto Rico Telephone Co., 48 F.3d 39, 45 (1st Cir. 1995) (finding third-party suppliers are not “automatically improper antitrust plaintiffs”), and often consumers or competitors are denied antitrust standing. The majority opinion proffers a bright-line rule “that cus- tomers and competitors in the affected market have antitrust standing,” Majority op. at p. 99. But in each case cited in the majority opinion the court denied antitrust standing, includ- ing to competitors. See Kochert, 463 F.3d at 718; Tri-Gen Inc. v. No. 18-2852 121

Int’l Union of Operating Engineers, Local 150, AFL-CIO, 433 F.3d 1024, 1031–32 (7th Cir. 2006); Brunswick, 429 U.S. at 477–78, 484, 488 (collectively, competitors); see also McGarry, 937 F.3d at 1063–66; In re Aluminum, 833 F.3d at 158; Serfecz, 67 F.3d at 597–98; Serpa Corp. v. McWane, Inc., 199 F.3d 6, 10 (1st Cir. 1999); SAS of P.R., Inc., 48 F.3d at 44; Associated Gen. Contrac- tors, 459 U.S. at 540 (collectively, neither consumers nor com- petitors). Instead, on this prerequisite the relevant case law priori- tizes the type, directness, and cause of an antitrust injury, ra- ther than applying a bright-line rule for antitrust standing based on the plaintiff’s status. See McGarry, 937 F.3d at 1064– 65 (examining the “type of injury” alleged, the “remote[ness]” or “direct link” between the alleged antitrust violation and the claimed antitrust injury, and the “causal connection”) (citing Associated Gen. Contractors, 459 U.S. at 537–40 (explaining these factors)); see, e.g., Blue Shield of Va. v. McCready, 457 U.S. 465, 476–77 (1982) (plaintiff’s status was neither consumer nor competitor, but granted antitrust standing based on type, di- rectness, and cause of antitrust injury suffered). Even with the presumption favoring consumers and com- petitors, standing is granted to rivals only when it “serves an- titrust policy.” 2A Areeda & Hovenkamp, ANTITRUST LAW ¶ 348a, p. 232 (4th ed. 2014) (examples include when a rival is in a position to detect a violation earlier than consumers, or the rival’s injury is large while consumers’ injuries are small or their suits less likely). “[T]he elimination of a single com- petitor, standing alone, does not prove anticompetitive ef- fect.” Austin v. McNamara, 979 F.2d 728, 739 (9th Cir. 1992) (quoting Kaplan v. Buroughs Corp., 611 F.2d 286, 291 (9th Cir. 1979)). Thus, it is uncommon that a suit by a single foreclosed 122 No. 18-2852

rival disputing a tying arrangement will have antitrust stand- ing.4 See 2A Areeda & Hovenkamp, ANTITRUST LAW ¶ 348a, p. 232 (“courts are properly skeptical of many rivals’ suits, par- ticularly when the practices are not obviously ‘exclusion- ary’”). First, the rival must show it suffered the type of injury the antitrust laws intend to prevent. This is consistent with antitrust jurisprudence that looks to market injury, like raised prices and decreased output or quality, to determine a com- petitor’s antitrust injury. See Tri-Gen, 433 F.3d at 1031 (“To have standing as a competitor, [plaintiff] needed to show that ‘its loss comes from acts that reduce output or raise prices to consumers.’”) (quoting Stamatakis Industries, Inc. v. King, 965 F.2d 469, 471 (7th Cir. 1992)); see also Serfecz, 67 F.3d at 597 (plaintiff’s injury was “linked to the injury inflicted upon the market,” including consumers paying higher prices); Associ- ated Gen. Contractors, 459 U.S. at 539 n.40 (finding no antitrust injury from alleged predatory behavior because competitor failed to show “that output has been curtailed or prices en- hanced throughout an entire competitive market”). Next, the rival must show it is the party “who can most efficiently vindicate the purposes of the antitrust laws”

4 The leading treatise acknowledges the possibility that a rival could have standing if an illegal tying arrangement creates, enlarges, or perpet- uates a monopoly for a tied product, and that rival may be foreclosed if it loses existing market share or sales. But “[i]njury to the foreclosed rival occurs, of course, only because a tie has forced buyers to purchase the de- fendant’s tied product rather than the rival’s.” 10 Areeda & Hovenkamp, ANTITRUST LAW ¶ 1767a, p. 449. This is a high bar: the rival must show (a) a monopoly in the tied-market, (b) the loss of its existing market share or sales (not limited to certain customers), and (c) an injury based only on buyers being forced to purchase defendant’s product over their own. No. 18-2852 123

against unlawful ties. Kochert, 463 F.3d at 718 (quoting Serfecz, 67 F.3d at 598). But, having been forced to purchase a product it did not want, the tied consumer is almost always in the su- perior position to sue the violator. McGarry, 937 F.3d at 1066 (concluding “[t]here is, after all, a more appropriate person to pursue [a] claim” when that person’s “self-interest would normally motivate them to vindicate the public interest in an- titrust enforcement”); see also Kochert, 463 F.3d at 718; 2A Areeda & Hovenkamp, ANTITRUST LAW ¶ 348a, p. 232 (recog- nizing that “consumers almost always have the correct incen- tives for suit, [but] rivals do not”). The presence of a more appropriate person to bring a claim “diminishes the justifica- tion for allowing a more remote party” to step in. McGarry, 937 F.3d at 1066 (quoting Associated Gen. Contractors, 459 U.S. at 542). Whether or not there is a more appropriate plaintiff, a rival may not pursue an antitrust injury that is “entirely de- rivative” of other injuries. Id. And even when standing is rec- ognized, a foreclosed rival cannot oppose efficient, legitimate, or even aggressive lawful competition by its rivals. 2A Areeda & Hovenkamp, ANTITRUST LAW ¶ 348a, p. at 231 (“[A] rival may allege an antitrust violation by its rivals not to protect competition but to protect itself from competition. … Such losses are not antitrust injury, so the rival is [] denied stand- ing.”). Here, Viamedia claims it is a foreclosed rival harmed by Comcast’s alleged tying conduct. First, Viamedia must show it suffered an injury the antitrust laws intend to prevent that is directly linked to or caused by Comcast’s alleged tying con- duct. But Viamedia relies on the injuries and damages it claims from Comcast’s refusal to deal rather than any distinct tying injury. See 335 F. Supp. 3d at 1069–73 (Viamedia’s expert testified that foreclosure from the market and resulting injury 124 No. 18-2852

or damages “flows directly from Viamedia’s inability to ac- cess” the Interconnects, not tying); see also Novell, 731 F.3d at 1080 (declining to recognize separate, cognizable antitrust in- jury for other exclusionary conduct that relied on alleged re- fusal-to-deal injury). Viamedia has conceded that any injury it suffered is derivative of Comcast’s refusal to deal, not the alleged tie. So Viamedia lacks an independent basis for ad- vancing a tying claim, and recognizing it as a foreclosed rival with standing to sue under the narrow exception described above would not “serve[] antitrust policy.” 2A Areeda & Hovenkamp, ANTITRUST LAW ¶ 348a, p. 232. While Comcast offered its customers, RCN and WOW!, access to the Intercon- nect and exclusive ad representation services, it was not re- quired to make such an offer to its rival, Viamedia. This re- veals the fatal flaw in Viamedia’s tying theory: even if Com- cast acted precisely as Viamedia claims by tying Interconnect access and ad representation services, Viamedia would suffer no separate, cognizable antitrust injury. 335 F. Supp. 3d at 1071; see Novell, 731 F.3d at 1080 (“Even if Microsoft had be- haved just as Novell says it should have, it would have helped Novell not at all.”). Next, Viamedia must show it is the party “who can most efficiently vindicate the antitrust laws” regarding tying. Kochert, 463 F.3d at 718 (quoting Serfecz, 67 F.3d at 598). While Viamedia enjoys the general presumption of antitrust stand- ing as a competitor, this presumption is limited by the court’s determination of whether an antitrust injury occurred. See 2A Areeda & Hovenkamp, ANTITRUST LAW ¶ 348a, p. 232. Here, RCN and WOW! would have been directly impacted by Com- cast’s alleged tying conduct and would be in the superior po- sition to pursue a claim against Comcast. Viamedia could not participate directly in this alleged tying arrangement, and it No. 18-2852 125

now seems motivated to protect itself from competition rather than enforce the antitrust laws. See Majority op. at p. 66 (“Vi- amedia simply wants to ensure that MVPDs can freely choose Viamedia as their supplier of ad rep services if that is their preferred choice.”).5 Because Viamedia has not shown a cognizable injury that the antitrust laws intend to prevent or that was directly linked to an antitrust violation caused by Comcast’s conduct, it has no antitrust injury. Without an antitrust injury, and not able to efficiently enforce the law against illegal tying, Viamedia lacks antitrust standing to bring this claim. B. Summary Judgment Was Properly Granted on the Tying Claim. The district court’s grant of summary judgment on Viame- dia’s tying claim, which we review de novo, was proper. We construe all facts and reasonable inferences in favor of the nonmoving party, and we refrain from weighing any evi- dence. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255 (1986); see also In re High Fructose Corn Syrup Antitrust Litigation, 295 F.3d 651, 655 (7th Cir. 2002) (describing the

5 Even if somehow Viamedia could be considered a foreclosed rival with antitrust injury and standing, the leading treatise notes the possibil- ity of injunctive relief—which the district court correctly concluded was misplaced on these facts, 335 F. Supp. 3d at 1074—and that treatise ob- serves how difficult it is in such a circumstance to prove damages without speculation. 10 Areeda & Hovenkamp, ANTITRUST LAW ¶ 1767a, p. 450 (“Even for those established in the market, estimating the number of sales lost as a result of the tying arrangement is elusive at best.”). Viamedia’s status as a foreclosed rival, though, would depend on the existence of a tie. I conclude below the undisputed facts show no evidence of tying conduct because there was no conditioned sale of services. 126 No. 18-2852

weighing of evidence as a “trap” to avoid). But the nonmov- ing party must “go beyond the pleadings” at summary judg- ment. Celotex Corp. v. Catrett, 477 U.S. 317, 324 (1986). It must affirmatively demonstrate through evidence that there is a genuine issue for trial. Anderson, 477 U.S. at 249. Liability under § 2 requires anticompetitive conduct. Mercatus Grp., 641 F.3d at 854; Endsley v. City of Chicago, 230 F.3d 276, 282 (7th Cir. 2000). On summary judgment, the § 2 plaintiff must present evidence tending to exclude the possibility that the monopolist’s conduct is as likely to be pro- competitive as anticompetitive. Matsushita Elec. Indus. Co. v. Zenith Radio, 475 U.S. 574, 588 (1986) (“conduct as consistent with permissible competition as with illegal conspiracy does not, standing alone, support an inference of antitrust conspir- acy”); Monsanto Co. v. Spray-Rite Service Corp., 465 U.S. 752, 768 (1984) (“[T]here must be evidence that tends to exclude the possibility of independent [pro-competitive] action by the manufacturer and distributor.”). As noted previously, firms—even monopolies—generally have the right to decide with whom they will do business. In this vein, “antitrust law does not require monopolists to co- operate with rivals by selling them products that would help the rivals to compete.” Schor v. Abbott Labs., 457 F.3d 608, 610 (7th Cir. 2006); see Pac. Bell Tel. Co. v. Linkline Commc’ns, 555 U.S. 438, 450 (2009) (monopolist may wield “upstream” power “to prevent rival firms from competing effectively” in a downstream market); see also Trinko, 540 U.S. at 409–10 (a monopolist has no duty to deal with a rival, let alone a duty to deal on favorable terms). In presenting its tying claim, though, Viamedia argued that Comcast’s conduct constituted more than a “mere” refusal to deal. 335 F. Supp. 3d at 1057. No. 18-2852 127

Viamedia claimed that when Comcast excluded Viamedia from the Interconnects, Comcast engaged in the “distinct” practice of tying, pressing forward with this claim as an “al- ternative theor[y]” of relief. Id. After discovery on Viamedia’s tying theory, the district court ruled that “the record leaves no genuine issue of material fact.” Id. The district court correctly evaluated Viamedia’s tying claim and rightly concluded that Viamedia’s proffered evidence did not tend to exclude the possibility that Comcast’s alleged tying conduct was as likely procompetitive as anticompetitive. Id. at 1055–64. Under the new antitrust regime, “[o]utright condemna- tion of product tying has been reversed.” Hon. Richard D. Cudahy & Alan Devlin, Anticompetitive Effect, 95 MINN. L. REV. 59, 76 (2010); see Microsoft Corp., 253 F.3d at 49 (“[N]ot all ties are bad.”). The “essential characteristic of an invalid tying ar- rangement lies in the seller’s exploitation of its control over the tying product to force the buyer into the purchase of a tied product that the buyer either did not want at all, or might have preferred to purchase elsewhere on different terms.” Jefferson Par. Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 12 (1984). An illegal tie, whether express or as applied, exists only when “the defendant improperly imposes conditions that explicitly or practically require buyers to take the second product if they want the first one.” 10 Areeda & Hovenkamp, ANTITRUST LAW ¶ 1752b, p. 291 (4th ed. 2018). The fundamental feature of a tying claim is the conditioned sale, including by force. See Sheridan v. Marathon Petrol. Co. LLC, 530 F.3d 590, 592 (7th Cir. 2008). No conditioning occurs if a buyer wants to purchase a bundle of the tied and tying products from the same seller. See Will v. Comprehensive Acct. Corp., 776 F.2d 665, 669 (7th Cir. 1985) (“A tie within the meaning of antitrust depends on showing that the buyer did not want to take both products 128 No. 18-2852

from the same vendor.”). Without the actual, conditioned sale of the tied product, there is no tie. See It’s My Party, Inc. v. Live Nation, Inc., 811 F.3d 676, 684 (4th Cir. 2016) (“If ... the buyer is free to decline the tied product ..., then by definition there is no unlawful tying.”). Other circuits have acknowledged the essential and indis- pensable role conditioning plays in a tying claim. In Aerotec Int’l, Inc. v. Honeywell Int’l, Inc., 836 F.3d 1171 (9th Cir. 2016),6

6 The majority opinion relies on certain factual assumptions to attempt

to distinguish Aerotec. First, the majority assumes “this is not a typical bundling case,” Ma- jority op. at 77, so Aerotec should not control. But Aerotec provides an in- structive framework for determining whether a tie has occurred and illus- trates the important role that conditioning plays in an unlawful tying ar- rangement, see 836 F.3d at 1179. Second, the majority assumes “self-providing ad rep services was not a viable option for RCN and WOW!,” Majority op. at 78, and so they “needed to employ [a third-party] ad rep services provider.” Id. at 77 n.17. As discussed elsewhere in this opinion, infra at 28–29, this assumption relies on the majority opinion’s conclusion that it must not have been eco- nomically feasible for RCN and WOW! to internalize ad representation, either regionally or locally. This is not supported by the record, which con- tains undisputed evidence of various economically feasible options MVPDs choose in structuring their advertising sales. See infra at 25–26. Regardless, this assumption does not prevent Aerotec’s application here. Third, the majority assumes RCN and WOW! had “no choice but to obtain ad rep services from Comcast,” Majority op. at 34 n.8, and were “forced by Comcast” to do so. Id. at 77 n.17. This in turn assumes that RCN and WOW! did not want ad representation services from Comcast. But the evidence shows the opposite: neither RCN nor WOW! ever requested In- terconnect-only services from Comcast, and both sought full-turnkey re- lationships with Comcast to receive ad representation and Interconnect services together. Ultimately, Comcast’s and Viamedia’s offers to solicit RCN’s and WOW!’s business were “nowhere near equal,” and it was “not a very difficult decision” for the MVPDs to make. 335 F. Supp. 3d at 1048. No. 18-2852 129

the Ninth Circuit concluded there was no evidence that Hon- eywell, a monopolist manufacturer of replacement airplane parts, “explicitly or implicitly tie[d] or condition[ed] the sale” of replacement parts on a requirement that its customers use its in-house repair services. Id. at 1179. When Honeywell ceased supplying parts to Aerotec—a third-party repair ser- vice provider—for repairs, Aerotec sued Honeywell for al- leged unlawful tying. Id. at 1177. Honeywell explained it had long preferred the benefits it achieved from internalizing its repair services and working with its affiliated servicers, rather than with independent ones like Aerotec. Id. at 1176–77. The court found it did not matter whether Honeywell had refused to deal with Aerotec, even if that made it more difficult for Aerotec to compete. Id. at 1179–80. The Ninth Circuit “de- cline[d] to stretch the tying construct to accommodate the claim that … conduct toward third party servicers … acts as an effective, or ‘de facto,’” tying condition. Id. at 1178. In Serv. & Training, Inc. v. Data Gen. Corp., 963 F.2d 680 (4th Cir. 1992), the Fourth Circuit reached the same conclusion, finding there was no conditioned sale of a licensed product when customers demanded access to it, even though this hin- dered the ability of third-party servicers to compete with the company. Id. at 687–88. The same is true with the D.C. Circuit, see Microsoft Corp., 253 F.3d at 85 (tying requires consumer to have “no choice but to purchase the tied product”), and this court, see Reifert v. South Cent. Wisconsin MLS Corp., 450 F.3d

Comcast was able to offer superior terms and better prices with full Inter- connect access and ad representation services. The assumption that any MVPD was forced or threatened to purchase ad representation services to gain access to the Interconnects does not follow from the evidence and is not a distinguishing factor from Aerotec. 130 No. 18-2852

312, 318 (7th Cir. 2006) (“conditioning access” to a service based on the forced purchase of a membership was essential element of tying). Like these other third-party rivals, Viamedia has offered no evidence of conditioning. Such evidence is simply absent from the undisputed facts on which Viamedia attempted to build its tying claim:  An ad representative is responsible for managing and selling an MVPD’s avails to advertisers and can represent their MVPD customers: (1) locally, selling only a part of an MVPD’s avails in a DMA to local advertisers; (2) regionally, selling all of the MVPD’s avails in a DMA; or (3) nationally. Id. at 1044.  The industry standard relationship between an ad representative and an MVPD is exclusive, region- wide, full-turnkey representation. Id. at 1045.  MVPDs in a full-turnkey relationship with an ad representative have an exclusive agreement with that ad representative, which makes Interconnect services available to them without requiring a di- rect relationship with the Interconnect operator. Id. at 1063–64.  Interconnect operators may pursue a direct rela- tionship with MVPDs to sell a portion of their avails regionally without a third-party ad representative. Id. at 1046. This is called an “Interconnect-only” agreement. Id.  Some MVPDs retain a portion of their avails for lo- cal advertising and hire an ad representative to rep- resent them locally. Id. at 1045. No. 18-2852 131

 Some MVPDs choose to conduct their own ad rep- resentation—regionally, locally, or both. Id.  RCN and WOW! are MVPD customers of Comcast and Viamedia, which compete for MVPDs’ busi- ness on the relevant Interconnects. Id. at 1046.  Comcast refused to deal with Viamedia by disal- lowing it access to the relevant Interconnects. Id. at 1057.7  No evidence shows that Comcast told MVPDs across all DMAs, expressly or impliedly, that they could only access the Interconnects on the condi- tion that they also purchase ad representative ser- vices. Id. at 1058.  Fourteen percent of Comcast’s agreements with MVPDs across all DMAs are Interconnect-only. Id.  RCN and WOW! never requested Interconnect- only services from Comcast, and both pursued full- turnkey relationships with Comcast. Id. at 1059. These facts do not show illegal tying conduct. Both parties stipulate that Comcast never denied an MVPD’s request to ac- cess the Interconnect on a standalone basis. In fact, 14 percent

7 The majority opinion casts Comcast’s refusal to deal as a negative fact, noting that Comcast “conceded repeatedly” that it had done so. Ma- jority op. at p. 76. It is undisputed that Comcast refused to deal with Viamedia; the question on remand is whether that refusal was anticom- petitive, which Viamedia will have to prove. As discussed, refusals to deal are now not disfavored, and in fact the opposite—cooperation among ri- vals—is a red flag under antitrust law. See Easterbrook, The Chicago School & Exclusionary Conduct, at 442 (“cooperation is to be feared rather than welcomed”). 132 No. 18-2852

of Comcast’s agreements with MVPDs across all DMAs are Interconnect-only. Id. at 1058; see, e.g., It’s My Party, 811 F.3d at 685 (14 percent of standalone, non-tied sales “exceed [] suf- ficiently” the minimum threshold required “to cast doubt on any allegation of tying”); see also 10 Areeda & Hovenkamp, ANTITRUST LAW ¶ 1756b2, p. 334 (“10 percent unbundle[ed]” sales rebuts any “established or presumed inference of a tying condition”); see also Novell, 731 F.3d at 1077 (aggregating sales and profits across entire market to assess firm’s “overall effi- ciency” and not singling out smaller market or product line). And there is no evidence Comcast withheld Interconnect ac- cess from MVPDs unless they also purchased ad representa- tive services from Comcast. Instead, the evidence shows that RCN and WOW! sought a full-turnkey relationship with Comcast to receive access to the Interconnects and ad repre- sentation services as a bundle. They were neither forced to purchase ad representation services from Comcast nor denied access to the Interconnects unless they purchased ad repre- sentation from Comcast. Further, if they wished, they were free to contract with a third-party ad representative for local sales. Even while under exclusive contract with Viamedia, RCN and WOW! were not forced to purchase ad representa- tion services from Comcast. Comcast never poached RCN or WOW! during that time to capture their business in the ad market, and RCN and WOW! could have chosen to forgo ad representation altogether at the conclusion of those exclusive contracts. That Comcast did not affirmatively offer RCN or WOW! Interconnect-only access does not alter these facts. Courts need not assume antitrust laws require a business to offer its customer a less profitable or less efficient option than the one the customer seeks. Here, it is undisputed that RCN and No. 18-2852 133

WOW! sought Interconnect access and ad representation ser- vices as a bundle in a full-turnkey relationship with Comcast. See 9 Areeda & Hovenkamp, ANTITRUST LAW ¶ 1700i, p. 12 (4th ed. 2018) (“[F]inding two products does not mean that they are tied together. The franchisee may have preferred a ‘turnkey’ franchise and never asked for the” tying product “separately”); see also Will, 776 F.2d at 670 (the “voluntary pur- chase of two products together” is “not a tie at all”). The dis- trict court recognized, and correctly rejected, how broad a view of tying Viamedia attempts to advance. 335 F. Supp. 3d at 1059 (“the constraining of consumer choice is of course a feature of a tying arrangement … but there must still be an actual ‘tie’ of products or services”).8 Viewing the record in the light most favorable to Viamedia, the evidence reflects that WOW! and RCN wanted full-turnkey representation, and they were prepared to hire the company with the ability to deliver both Interconnect access and ad representation ser- vices. Here, that company was Comcast. My colleagues in the majority conclude, at various points, that it was not economically feasible for RCN or WOW! to conduct their own ad representation. See Majority op. at p. 34 n.8 (finding RCN and WOW! “had always chosen to buy [ad representation] services from outside companies, suggesting that in-house was not an economically viable option”); id. at 77 n.17 (concluding sufficient evidence existed of a “forced” sale because “RCN and WOW! needed to employ [a third- party] ad rep services provider”); id. at 78 (deciding “self- providing ad rep services was not a viable option for RCN

8 Further,if tying occurred, one would expect to see higher prices or lower output. See Washington Legal Foundation’s Br. in Supp. of Def. at 12, ECF No. 45. There is no such evidence here. 134 No. 18-2852

and WOW!”); id. at 82 (finding evidence “showing that in- house provision of ad rep services simply was not a practical option for RCN or WOW! in these markets”). This conclusion assumes that if RCN and WOW! chose to outsource ad repre- sentation to a third-party, either locally or regionally, then it must have been economically feasible to do so. It also assumes the inverse: that internalizing these services must be econom- ically infeasible. Such assumptions go beyond the court’s role and presume underlying facts about the parties’ business practices and strategies that we simply do not know and can- not accurately predict. These assumptions are not supported by the record, which instead contains evidence of various eco- nomically feasible options MVPDs choose in structuring their advertising sales, including internalizing part of or all ad rep- resentation services. 335 F. Supp. 3d at 1044. I respectfully do not share this economic feasibility conclusion and rely only on the evidence presented. The majority opinion also references “[a]mple evidence” the jury could have relied upon to “easily find that Comcast improperly forced the smaller MVPDs to buy its ad rep ser- vices” through a tying arrangement. Majority op. at pp. 75– 76. In particular, the majority points to internal Comcast and WOW! documents, deposition responses, and testimonies be- fore the Federal Communications Commission and the Department of Justice. But the relevant evidence shows Com- cast’s desire to solicit RCN’s and WOW!’s business directly, not the forced purchase of a service that neither RCN nor WOW! wanted. For example, the majority references an email from a WOW! employee to his colleagues explaining his un- derstanding that “[WOW!] can be in the [Interconnect] but only if [Comcast] rep[s] us directly.” See Majority op. at p. 75. No. 18-2852 135

The majority opinion reads this as revealing Comcast’s “de- mands” and “threats” with which RCN and WOW! did not “willingly” comply. Id. But in context, the email reflects only ongoing business negotiations between WOW! and Comcast, in which Comcast expressed its desire to no longer accommo- date third-party ad representatives (its competitors) on the Interconnects it operates. This is consistent with Comcast’s position all along—that it wishes to increase efficiency by in- ternalizing services—and does not constitute tying as de- fined. Some other examples: in a different email, a Comcast em- ployee expressed to a WOW! employee Comcast’s “desire to have a direct relationship” with MVPDs on the Interconnects and noted that “Comcast would be thrilled to do business di- rectly with WOW,” reiterating the “tremendous value and benefits [Comcast] can deliver for WOW” in a direct relation- ship. Another Comcast employee noted in a deposition that working “through a middleman” like Viamedia “really brought no value to the table” and the decision not to renew its contract with Viamedia was primarily to “have a direct re- lationship with WOW and RCN.” Explaining RCN’s decision to contract directly with Comcast, an RCN executive testified to the superior terms Comcast could provide, noting that Comcast’s and Viamedia’s offers were “nowhere near equal” and that it was “not a very difficult decision” for RCN to make. 335 F. Supp. 3d at 1048. WOW!, too, selected Comcast’s direct representation based on “better financial terms.” Id. These statements are consistent with Comcast’s position to directly solicit the business of MVPDs. When questioned by the Department of Justice, a Comcast executive acknowl- edged it was Comcast’s “business practice” to inform MVPDs 136 No. 18-2852

that want “to get access to a Comcast controlled Interconnect, it has to hire Comcast as its ad sales representative” instead of a third-party rival. Id. at 1061. This does not reveal a forced purchase that neither RCN nor WOW! wanted. If anything, such evidence depicts “hard-nosed” business practices like those the Tenth Circuit noted but did not find anticompetitive in Novell, 731 F.3d at 1078 (finding email evidence may sug- gest “an uncharitable intent toward rivals” or even an “intent to undo a competitor,” but did not show that conduct was “irrational but for its exclusionary tendencies”). None of this evidence “tends to exclude the possibility” that Comcast’s conduct “was as consistent” with lawful conduct as with ille- gal tying conduct. Mercatus Grp., 641 F.3d at 856. Importantly, the internal documents the majority opinion references are dated 2014, when WOW! and RCN were still under exclusive contract with Viamedia but soliciting bids from other representatives to sell advertising avails on the In- terconnects. While soliciting bids, RCN and WOW! engaged in back-and-forth negotiations with Comcast and Viamedia, leading ultimately to new contracts with Comcast. The timing of these discussions, as preserved in the record, is key. When reviewing the evidence here, this court is limited by the prospective nature of these then-ongoing negotiations. Our perspective is necessarily predictive not retrospective. These internal documents do not contain evidence of the economic impact after the negotiated deals were made, but only before when offers were being made, accepted, and rejected. They do not describe the economic results of these then-prospective re- lationships. Because of their prospective nature, they are not a reliable metric for understanding what has transpired since, but merely what the parties had hoped to achieve. Under our No. 18-2852 137

antitrust regime, “[w]hat distinguishes exclusion from effi- ciency is what happens in the future,” Easterbrook, The Chi- cago School & Exclusionary Conduct, at 443, not what the parties want to happen. None of this evidence measures whether Comcast’s con- duct has resulted in a market injury, such as raised prices or decreased output. It does not show that RCN or WOW! understand their contracts with Comcast to be coercive or conditioned, and there is no evidence indicating RCN and WOW! are unhappy with their current relationship with Comcast. While viewing this evidence in Viamedia’s favor, it is crucial to acknowledge its limitations in this case. As the district court noted, the real “gravamen of Viame- dia’s tying claim [is] that Comcast’s refusal to provide [Viamedia] Interconnect access prevents [Viamedia] from selling the kind of full-turnkey Ad Rep Services that WOW! and RCN desire.” Id. at 1063 (internal quotations omitted). That differs from the paradigmatic tying claim, which re- quires conditioning and a forced sale to a customer rather than a withheld advantage from a rival. See 9 Areeda & Hovenkamp, ANTITRUST LAW ¶ 1700a, p. 4. Antitrust law does not require Comcast to help Viamedia sell the same bundle it offers to their mutual customers. See 10 Areeda & Hovenkamp, ANTITRUST LAW ¶ 1748b, pp. 251–53 (no tied product “when the plaintiff’s theory of injury is not that cus- tomers of the defendant’s [tied] bundle would buy the items unbundled if they could, but rather that a rival could sell the same bundle if only the defendant would sell it a particular in- put”). 138 No. 18-2852

Finally, even if Viamedia’s refusal-to-deal claim had sur- vived a motion to dismiss, summary judgment was appropri- ate on the tying claim because the undisputed facts do not show Comcast conditioned access to the Interconnects on the purchase of ad representation services. Although acknowl- edging that “[t]hese related claims are both based on the same course of conduct, resulted in the same anticompetitive harms, and would be subject to the same procompetitive jus- tifications or defenses,” Majority op. at p. 38, the majority opinion does not explain how Viamedia has established a dis- tinct tying claim that should have advanced beyond summary judgment on its own merits. In framing its tying analysis, the majority opinion focuses on whether, when viewing Com- cast’s conduct as a whole, “[Comcast] has unreasonably main- tained or enhanced its monopoly position.” Id. at 73. While this framework would not necessarily reveal anticompetitive conduct in violation of § 2, I hesitate to apply it here for the simpler reason that our antitrust jurisprudence does not in- struct us to do so. Instead, looking to the undisputed facts, Viamedia has failed to prove the elements of a tying claim. To the extent the majority opinion also concludes that the existence of a viable refusal-to-deal claim saves Viamedia’s ty- ing claim, I am not convinced of that connection. In compari- son to refusal to deal, tying is a form of exclusionary conduct with more specific requirements, which the district court noted. 218 F. Supp. 3d at 698 (a refusal-to-deal claim is differ- ent than a tying claim); accord Novell, 731 F.3d at 1072 (distin- guishing unilateral refusals to deal from other recognized forms of anticompetitive conduct like tying). Viamedia, seek- ing an “escape route” and “trying to recast” Comcast’s refusal to deal as unlawful tying conduct, see Novell, 731 F.3 at 1078, has treated these two claims interchangeably. But parties’ No. 18-2852 139

claims stand and fall on their own merits, so the district court was correct to consider them separately. Conditioning, an essential element of tying, requires the forced sale of a product the buyer did not want. Based on the record before us, to which we are limited, there was no con- ditioning. Because the undisputed facts show no evidence of tying conduct separate from Viamedia’s refusal-to-deal claim, I would affirm that portion of the district court’s judgment. III. Section 2 Monopolization Analysis The majority opinion acknowledges Viamedia raised only two claims on appeal: refusal-to-deal and tying. Majority op. at p. 42 n.11. It also suggests an alternative means of § 2 re- covery should Viamedia fail on either of its alleged claims. Per the majority opinion, Viamedia has sufficiently pleaded and presented evidence for the court to find an alternative § 2 mo- nopolization claim. Such an open-ended approach may place courts in the role of the decision-maker on dense and complex economic issues better left to the free market. See Easterbrook, The Chicago School & Exclusionary Conduct, at 442 (“Markets are much better than judges at sifting efficient from anticom- petitive practices.”). We are no more skilled at predicting mar- ket shifts than anyone else, and courts are very rarely the best forum for discerning between exclusionary and efficient con- duct. See id. at 442–45. The district court hewed to the particular types of exclu- sionary conduct Viamedia alleged (tying and refusal to deal; its exclusive dealing claim was abandoned on appeal). The court found Viamedia had “affirmatively disavowed” any “free-standing” § 2 monopolization claim. 335 F. Supp. 3d at 1068. Having failed to plead or preserve an alternative claim 140 No. 18-2852

before the district court, Viamedia should not be entitled to pursue a vague § 2 monopolization theory on appeal. King v. Kramer, 763 F.3d 635, 641 (7th Cir. 2014) (this court not in a position to advance claims that party abandons or fails to pre- serve below); Geva v. Leo Burnett Co., 931 F.2d 1220, 1225 (7th Cir. 1991) (an issue not “properly preserved below” in the district court is generally waived). Nor should the court be required to consider the restraint on trade or the impact on monopoly powers generally that may arise from Viamedia’s allegations. This rule is particularly relevant in antitrust law, which is susceptible to high rates of “false positives” that occur when the court confuses real competition with exclusion. See Easter- brook, The Chicago School & Exclusionary Conduct, at 445; see also Microsoft Corp., 253 F.3d at 87 (discussing false positives). The new antitrust regime identifies problems like false posi- tives, unpredictability, and past court confusion, and it seeks to protect consumers and promote competition in a techno- logically advancing marketplace. Many cases and authorities point in this direction. An alternative § 2 monopolization theory is not consistent with how courts now consider poten- tially exclusionary conduct. Accordingly, I am unable to join this portion of the majority opinion. IV. Conclusion The majority opinion dives deep into this case’s compli- cated facts and thoroughly covers swaths of antitrust law in an insightful manner. As Viamedia has plausibly alleged an anticompetitive refusal to deal, I join the majority in reversing and remanding on that claim. But the undisputed facts do not show an illegal tie, so summary judgment was proper on that No. 18-2852 141

allegation. For the reasons above, I respectfully concur in part and dissent in part.