This is a follow-up to yesterday's post on Epic Games v. Apple. I have listened for a second time to the appellate hearing with a particular focus on the question of whether Epic's proposal of a single-brand market is dead at this stage due to a "failure of proof." Epic's appellate argument is purely legal; they are not challenging any factual findings under the "clearly erroneous" standard (though they do disagree with some of them).
There are three reasons for which I didn't consider Circuit Judge Milan D. Smith's concern over a potential failure of proof, which he voiced toward the very end of the hearing, necessarily fatal to Epic's case:
Especially during the first 10-15 minutes of the hearing, Circuit Judge Smith appeared inclined to hold that the district court had committed legal error in the market definition context, and that this would require a remand. Epic didn't deny that a new market definition would require a remand. Its counsel even acknowleged it, but argued that it would prevail anyway under the rule of reason. (However, both Epic and the DOJ faced tough questions regarding their position that a final balancing was required.)
I had discussed Epic's and Apple's position on the aftermarket part of the single-brand market test in September, and large parts of the problem with the fourth aftermarket factor are requirements that the district judge thought Epic would have to meet, particularly a policy change and (which would obviously be the case in the event of a policy change) customers' complete unawareness, including their inability to find out even if they wanted to, of the aftermarket restrictions by the time they make their purchasing decision in the foremarket. An incorrect legal standard is different from a failure of proof at the factual level.
The most important question here is whether one considers the district court's finding of Epic not having proved lock-in a legal or factual determination. Apple uses an overbroad definition of what is "factual" and accuses Epic of, conversely, describing actually factual determinations as legal conclusions. So let's look at this part more closely because that's what the appeals court is going to do in the months ahead.
What certainly does help Apple here is that Epic's expert on lock-in, Professor Susan Athey, "got blown up" as Apple's counsel described it: the district judge was "not impressed" by what she presented (as it was not based on consumer surveys or similar data). That is indeed bad for Epic, but not fatal: a party can prevail on a question even if its related expert testimony isn't considered reliable.
According to Apple's counsel, all that Epic has to show in terms of lock-in amounts to only two--and rather old--Apple documents. That is an exaggeration. Let me quote from Epic's opening brief first:
"[The district court] made supporting factual findings, including that there are significant obstacles to switching away from iOS, such as 'time to find and reinstall apps or find substitute apps; to learn a new operating system; and to reconfigure app settings,' [...] and unsurprisingly, 'very low switching rates,' [...] The court nevertheless determined that Epic 'failed to prove lock-in,' primarily because the court incorrectly believed Epic needed to quantify consumers’ switching costs."
"The [district] court's findings and undisputed record evidence prove high switching costs. [...] To switch to Android, consumers must abandon a considerable sunk cost—their smartphone and its apps. [...] But that’s not all; as the district court correctly found, 'it takes time to find and reinstall apps or find substitute apps; to learn a new operating system; and to reconfigure app settings. It is further apparent that one may need to repurchase phone accessories.' [...] The court also found that Apple sought to compete by making its platforms 'stickier,' and its executives touted the difficulties in switching in their internal correspondence. [...] These costs, which result in persistent lock-in to iOS past the lifespan of a single device, far surpass those in Kodak, where a buyer could readily switch to a different brand once its first Kodak photocopier became obsolete."
In its reply brief, Epic then countered Apple's suggestion that it was confusing customer satisfaction (voluntary) with lock-in (an unwanted consequence of a previous decision):
"Second, the [district] court complained that Epic 'ignor[ed] the issue of customer satisfaction,' [...] and noted that 'the features that create lock-in also make Apple’s products more attractive,' [...] Nothing in Kodak or its progeny suggests that if consumers like the features that create lock-in, switching costs should be discounted, and the district court provided no legal support for this position."
"Third, the district court said that it 'is left entirely in the dark about the magnitude of the switching costs and whether they present a meaningful barrier to switching in practice.' [...] That assertion contradicts the court's findings, which identified numerous barriers with which all modern smartphone users are familiar, [...] and confirmed there is no meaningful migration between platforms [...]. Moreover, Kodak does not require a plaintiff to quantify the magnitude of switching costs; in Kodak, it sufficed that plaintiffs 'offered evidence that the heavy initial outlay,' along with support and parts, made switching costs very high for existing customers."
Epic basically argued in its two appellate briefs that the district court found all the ingredients of lock-in, and just made the mistake of not determining that if it looks like a duck, walks like a duck, and quacks like a duck, it probably is a duck. The way Epic actually stated it in its reply brief is this:
"[T]his is a classic single-brand product aftermarket case. Kodak and Newcal are good law, and Epic has proved all facts required for such a determination."
If one agreed with that, it would mean that the district court erred in its application of Kodak and Newcal to the facts it found, unless one considers the finding of a failure of proof of lock-in a dispositive factual determination and ends the analysis there. But even if one just considered it a mixed question of fact and law, the standard for review would be de novo and the appeals court could easily reverse.
Epic also argued that Apple's monopoly power was shown (such as that Apple only reduce its App Store commission under regulatory or litigation pressure).
What's very important to consider--and could have been my fourth bullet point further above--is that Circuit Judge Smith, shortly after raising the question of a potential failure of proof, noted that the district court's ruling was inconsistent about these types of facts and the conclusions drawn from them.
I remain optimistic that there will be a partial reversal and remand. Even if one accepts the district court's factual findings (as Epic elected to do on appeal), the elements are in place for the appeals court to tell the district court to give market definition another try.
It became very clear yesterday that Circuit Judge Smith does not believe that the district court got everything right, and that the critical part on which this case turns is market definition. I know Epic says market definition doesn't matter, but that's where I disagree with them. I don't think the district court's market definition should be affirmed, but if it was (and the risk is obviously greater than zero), it would seem pretty reasonable to me (though not my preferred outcome) to affirm the fate of Epic's federal antitrust claims. Epic's counsel argued that he couldn't think of any market definition under which Apple's conduct should be accepted, and that's an extreme position I respectfully disagree with: in a broad enough market, no one has market power.
The hearing would have gone really bad for Epic if the appeals court had said that the decision looks correct in every outcome-determinative respect. That was not the case. Circuit Judge Smith wanted to explore the practical consequences of a remand centered around market definition. That Epic would be in a stronger position if its lock-in expert had been more persuasive does not mean that it's "game over."
There's also a potential Plan B here: if the appeals court indicated that something was wrong with the district court's market definition, the DOJ--which supports Epic on certain questions without taking a final position on how the case should be decided--would benefit from it if and when bringing its own case against Apple's App Store terms. But we're not there yet. I still consider affirmance of the Epic v. Apple district judgment much less likely than a remand.
