Epic v Apple Day 9 - Daily Pool Report


Leah Nylen
 

 

Day 9: I’m Elizabeth Lopatto of The Verge, today’s pool reporter, but you can call me Liz. Seated next to me is Dorothy Atkins of Law 360. You can DM me (I’m @mslopatto) or email me: liz@...

Exhibits can be found here: http://tinyurl.com/epicvapple

Judge Yvonne Gonzalez Rogers (YGR) is on the bench wearing a black mask.

Today’s teams include: 

Epic:
Yonatan Evans 
Tim Sweeney 
Katherine Forrest
Alex Carvajal
Jin Niu

Apple: 
Cynthia Richman
Richard Doren
Heather Grenier 
Karen Dunn
Harry Phillips
Heather Grenier
Phil Schiller

Lorin Hitt, professor of operations, information, and decisions at the University of Pennsylvania Wharton returned to the stand for Apple, with Apple’s Cynthia Richman performing the direct examination.

YGR begins by discussing whether it is appropriate or not for a witness to speak to counsel on the break. Trial judges do have discretion, she tells us. Katherine Forrest for Epic thinks the conferring during cross is prejudicial. She asks for a standing order against it. Apple does not object. YGR approves.

Hitt was discussing analysis of competitive effects yesterday. We see a slide from yesterday title “Output Growth."

Discussion of commission rates. There’s a slide, DX-4083, that shows the average commission rate for in-game purchases. DX-2806 is a graph showing Average in-game purchase price and average commission. Hitt says that developers can charge higher prices on IAP, and because Apple’s commission is steady, Apple is earning more in commissions. Epic is one of the developers that’s raised its price over time, Hitt says. (Hitt is talking very quickly; YGR has asked him to slow down but unfortunately it didn’t take.) Looking at DX-4757, Hitt says that Apple’s prices have been constant and developers are making greater use of IAP and charging higher prices. “It does suggest that game developers and consumers are getting more value for these transactions.”

Generally, for all apps, the commission rates are steady or declining. The decline is due to changes in Apple policies, Hitt says. There are details of small-developer commission programs — small developers pay less to Apple. Hitt says Epic could have taken advantage of subscriptions, where Apple has lowered subscription commissions for long-term subscribers. (Change happened in 2016.) 

//average commission users pay Apple is 27.7% b/c it ignores free apps in the App Store. When you factor in free apps, the avg commissions rate on all transactions is 4.7% and 8.1% on game transactions, he says//


Richman: Did Apple raise its prices after it allegedly gained a monopoly (a date Epic’s witnesses have put around 2010)?

Hitt: No. “Apple’s commissions have been steadily declining throughout.”

YGR gets him to confirm this applies to both initial and in-app purchases.

Says David Evans, who you may remember from Day 7, has done very little analysis of commission rate. Evans omitted free transactions from his commission — which means that the results are different. Excluding those “makes the commission rate seem higher than they are.” DX-5469 shows a bar graph, with Evans’ “incorrect” calculation, compared to Hitt’s “correct” calculations. Fun labeling! The difference is due entirely to omitting free transactions, Hitt testifies.

There is some discussion of how to directly measure “quality” — anticompetitive measures make quality decline, Hitt says. One argument against a quality decline is that devs make more money, Hitt says: “They’ve been able to raise their prices because consumers perceive the value in what they’re offering."

Now innovation. There’s nothing to suggest Apple has decreased innovation, Hitt says. “That has improved the kinds of services devs can acces and the kinds of games that are available.” This is bc of technical improvements to the iPhone.

We now discuss game streaming services. If you’d like to play some high-end games, one way to do so is to stream it through a web browser. He lists Stadia, X Cloud, Luna, and Geforce Now. Weird — didn’t Microsoft and Nvidia testify Apple rejected their apps?

Onto user lock-in. Previous testimony suggested iPhone users were locked in, but Hitt doesn’t agree. “For one thing, consumers actually do switch.” As many as 26 percent switch platforms when it’s time to replace a phone. “People do indeed switch platforms when they can.” The 26 percent number comes from an industry study. A sealed document has a different document, but Hitt won’t say what’s in it because it’s sealed. 

High rates of customer retention aren’t evidence of high switching costs, Hitt testifies. “I stick with the product I have because I like the product I have.” He says he’s not sure this was considered in other analyses. Drivers of loyalty were put into switching costs, he says, and he doesn’t think that’s the right way to think about it.

Developers have a role in facilitating switching, Hitt testifies. Switching apps, locating apps or moving data from platforms may inhibit people from switching, but developers control those costs. In nearly all cases, the major apps are available on both platforms, Hitt says. Devs can make content portable, letting you transfer from one service to another, or portability of subscription content. “Developers do have a lot of control about how difficult that switching process is.”

Generally, re-purchasing apps isn’t a significant cost, but in the games market, the vast majority of games are free to download, Hitt says. 

Even if it’s free to download, unless there’s cross-platform or cross-wallet play in games, then if you haven't used all the money in your wallet on an app in an iOS device then you'd better use it before you switch to an Android device.” Is that right, YGR asks? Yes, Hitt says, but that’s a developer’s call. (You may remember earlier that some consoles do control this. Apple doesn’t, but Sony does, for instance.)

The relevant market is digital game transactions and 62 percent of all revenue in the App Store is game transactions, Hitt says. He looked at other choices for game transactions, per slide DX-4758. Says he identified 14 different platforms, including Google Play, Steam, Nintendo, and xBox.

Developers have choices for game transactions, Hitt says. They consider customers on platforms, the tools the platforms provide, services and technical support, capabilities of the platform, and cost — how to develop and transact. Those are the factors devs consider, he testifies.

“It’d be hard to play Pokemon Go with a console,” YGR observes.

One example is Epic deciding not to go through Google Play, initially. They later decided to offer their games on Google Play, Hitt says. Apple doesn’t provide restrictions on where you develop your games — Apple does not require exclusivity on iOS games, Hitt says.

Onto consumers. There are many platforms, mutli-homing for the same game (that is, to use multiple platforms), and the ability of consumers to switch. There are natural experiments, Hitt says, that provide insight onto how consumers move among platforms. 

Consumers switching between games matters, Hitt says. That makes these platforms potentially substitutable even if consumers aren't playing exactly the same game. That might be same IP but different games, or something else. 

Hitt analyzed multihoming for Fortnite. “That’s not something you normally are able to see.” Slide DX-4767: 35.9 percent of iOS users are multiplatform users. (There were 115 million iOS users.) On PlayStation 4, 104 million users, 32 percent on another platform. On other platforms, there’s a range between 20 and 50 percent for platform switching.

Another slide that Hitt seems particularly proud of DX-4769. 75.9 percent of consumers buy nothing. 15.8 percent only buy outside iOS. 5.6 percent purchase on iOS, and 2.8 percent buy on iOS and elsewhere. The data are restricted in time period, but it’s all transactions — no sampling.

Hours played and revenue, March 2018-July 2020, DX-5339. On iOS, users played 2,752 hours, or 10 percent of all time users played Fortnite. (90 percent of the time was on other platforms.) iOS was $745 million, and 13.2 percent. “The vast majority of spending and play is on other platforms.” iOS is slightly better at getting revenue per time played, Hitt says. 

Now we are discussing consumer substitution. Hitt analyzed switching after the hotfix, which may give an indication of whether consumers can move across platforms when given a reason to do so. (Other situations include introduction of Nintendo Switch, console and PC gaming companion app analysis, Spotify subscriptions, and Netflix subscriptions.)

We are now arguing about analysis methods; Evans used a sample. 

Post hotfix switching: 22 percent to 51 percent of the expected spending relative to what it would have been was retained for all iOS users. 22 percent to 38 percent for iOS only users. 81 to 88 percent of Fortnite spending was retained, per DX-4824. “Not all the iOS spending moved over, but as we saw a few slides back, a lot of the spending was on other platforms,” Hitt says. Some iOS spending moved to other platforms. “Consumers are willing to and able to move across platforms when they have reasons to do so." Says Evan’s data are different because of sampling, analyses. He uses a different methodology, which Hitt has “issues” with.

The Nintendo Switch. Fortnite launched there June 2018. Slide DX-4823. Customers who used both iOS and Switch spent less on iOS after the Switch introduction relative to users who were on iOS but not Switch. Spending varies from month to month, and is driven by new “seasons.”

Analysis of Console/PC Gamers substitute. Those who downloaded console or PC gaming apps had spending growth of 19 percent on iOS; those who didn’t download those apps had spending growth of 24 percent. When you have an additional option and spending on iOS slows, that indicates consumers are substituting, Hitt says.

Final two natural experiments: Spotify and Netflix. The two decided not to offer new subscriptions through the iOS app, so you can see whether that affected revenue. No change in revenue associated with this, Hitt says.

Hitt says that there are not high barriers to substitution.

Now friction. Other app developers do indicate they can participate on multiple transaction platforms, Hitt says. “Consumers can buy V-bucks through a web browser.”  Qualitative analysis is appropriate for this, YGR asks. Yes, Hitt says.

One of the things that indicate frictions are low is that consumers already do this, Hitt says. Web browsers are widely available, and you can transact there, as well as on consoles. “Maybe we’ve just got habituated to things are easy to do.” But if the friction is leaving an app and spending a minute or two on a web browser — that’s more convenient than leaving one convenience store to go to another. “Very small, compared to the ones you see in the real world.”

How can developers ease these frictions? DX-4777 The frictions are under developer control, Hitt says. Two possible things are “single sign on” eg can you sign on to the same account, and is the content portable. In all apps, the top 25 percent by revenue: 96 percent have single sign on, and 92 percent have partial or full  portability. Top 25 by downloads: 80 percent have single sign on, and 84 percent have portability. For game apps; the Top 25 by revenue: 92 percent have single sign on, 84 percent offer partial or full portability; in the top 25 percent of downloads, 32 percent have single sign on, and 32 percent have full or parietal portability.

We now discuss Tinder, which Applle’s counsel jokes that no one in the courtroom has ever used. (For the record: I have used it.)  It has single sign-on, you can use it in an iOS browser. These frictions are very low, Hitt says. 

“Whenever I talk about Tinder in my class, every one of my students has access to Tinder.”

CROSS.
Yonatan Even for Epic. 

Even asks about switching between Android and iOS. 78 million new phones every year, compared to 300 million phones in the US.  7 or 8 million switch out of about 300 million phones, Even says. That’s something like 3 or 4 percent of the installed base, something like that, Hitt says.

Streaming services: “You agree those streaming services are not able to be native apps on iOS.” If Netflix weren’t native, that would hurt Netflix, most likely, Hitt agrees. “It would harm innovation or competition, correct?” Hitt says he doesn’t know. “I haven’t done the analysis. I would think it would be harder to access Netflix.”

In that slide, you argue something wonderful happened to Epic in March 2018, Evens says. Fortnite came out on iOS, hooray! Essentially, Epic was able to sell at $5 before the introduction and $10 after that, doubling Epic’s IAP. Apple’s commission similarly went up from $1.50 to $3. Apple made more money on each transaction, the same 100 percent increase. But the transaction services weren’t different, Evens says. 

Evens: In March 2018, Apple did not introduce any new innovation into the sphere of app store transactions that completely changed the value it delivered, correct?

Hitt: Correct.

In the context of a patent license, they run out, right? Hitt agrees. But Apple’s position is that it’s entitled to 30 percent forever, Evens says. “There’s not a temporal element in that sense… it’s not a comparison I’d draw.”

We are looking at purchasing behaviors from iOS Fortnite users again. I will spare you the re-discussion of the numbers, which are above. “Do you recall that in your report, you found that 15.8 group who purchased only on Sony and other devices accounted for 65.4 of the spend by people who ever accessed Fortnite on iOS,” Evens asks. This seems to confuse Hitt and YGR. “Of that 15.8 percent, 65 percent did what?” YGR asks, apparently also confused. We go through a great deal of confusion, which I will not transcribe in full. There is some shuffling of binders among Epic’s lawyers. 

We are looking at exhibit 13 from Hitt’s rebuttal report. If Apple, for instance, raised the price on distribution of Fortnite on iOS, that would not affect the transactions that never touch iOS, right? If consumers behave as they have in the past, yes, Hitt says. 

E: If you look at what happened after August 13, you included those transactions in the 85 number

Hitt says his number does include that group. That means there's a guaranteed retention rate of 65 percent, right, says Evens. Yes, says Hitt.

More binder confusion as we look for Hitt’s deposition. Evens finds what he’s looking for. He now reads from the deposition. 

E: IF you look at the 23.1 percent that Epic obtained from people only ever purchasing on iOS and the 11.5 percent that sometimes purchase on iOS on others, Epic only retained about 50 percent, correct?

H: Correct.

This is Hitt’s first case on market definition, and he’s never published books or papers about it either, Evens establishes. Nothing on antitrust, either. Nothing about the antitrust implications of two-sided markets, for that matter, we discover.

We are now discussing the small business program, which lowered the rate on developers to 15 percent if they met certain standards. Before that, did any other platform lower its commission? Hitt doesn’t recall timing but thinks Valve lowered its commissions for very large developers, from 30 to 25 to 20. Microsoft may have done something as well.

In late 2020 or early 2021, there’s no decline in demand for ioS app transactions, right? Hitt didn’t examine this specifically but there’s nothing to suggest it. Apple said it was lowering commissions to support innovation, Evens establishes. 

If Apple lowered its price to support innovation, that means the price wasn’t set by supply and demand forces?

Yes, I would say that. It’s a response to enhance the overall size of the ecosystem.

If Apple lowered its price to support innovation, do you believe its current price is [unintelligble] 

I didn’t investigate that, and have no opinion, Hitt says.

Do you see any data that suggests Apple couldn’t increase its price tomorrow? Hitt hasn’t seen data on feasibility or impact on that. It’s not beyond the pale? I don’t like to speculate, Hitt says. People can raise prices from time to time? Given market circumstances, yes. As far as I can tell, Evens seems to be suggesting that Apple isn’t constrained by market forces — its pricing decisions are arbitrary.

Now back to Jobs’ statement about not making money on the app store. Hitt says he hasn’t modeled dev’s revenue if the app store did not make money. Evens also establishes that Hitt hasn’t done any modeling to determine whether the improvements in iPhone affect this. Hitt says that modeling isn’t appropriate.

Cellular bandwidth and speeds have grown exponentially since 2008, Evens says. Hitt agrees. Today we can stream apps on cellular, right? Yes. “We couldn’t do that in 2008?” I think that’s fair. “Today we can have video chats ove Zoom and Facetime?” Yes. Hitt agrees things have gotten dramatically better. Streaming games are now possible over cellular. All this growth has nothing to do with Apple’s restrictions on the App Store, right? Yes. Has Hitt controlled for the immense growth in cellular capabilities? Hitt says he doesn’t think that’s appropriate to the innovations Apple has contributed.

Hitt relied on other experts for Apple in forming his opinions, he says. He says he did not perform a hypothetical monopolies test. He didn’t offer an opinion about a hypothetical monopolies test is required.

We recess at 10:14 for 20 minutes.

We resume at 10:34AM. 

Lorin Hitt, professor of operations, information, and decisions at the University of Pennsylvania Wharton returned to the stand for Apple, with Epic’s Yonatan Even on cross.

YGR wants to know if he did any clustering in testing the market definition he rejected. There were some analyses I did to suggest the underlying economic considerations are different for games. Did he do entertainment, or IAP-driven econ apps, as opposed to others where everything is free, YGR asks. “Games in particular are unique, or at least very different, because they rely on IAP of individual items.”

Spotify brought the same complaints and allegations, but that would require a different analysis, Hitt says. Dev tools would require different analyses. Epic publishes House Party, a social app not within the games market. No analyses on EGS on iOS, either. 

E: When an EGS user obtains an app that’s not a game, is that a digital game transaction? 

H: It’s a digital transaction but not a game transaction, so that would be no.

E: You have not offered any kind of definition of what a game is? No. Do you assume Roblox is a game? 

H: “I don’t assume, I relied on classifications that were provided by developers.” 

E: Have you consulted with anyone at Apple about whether they consider Roblox a game? 

H: No.

Even is trying to establish whether a developer’s opinion of friction will matter separately from Hitt’s definition will influence decisions. 

Was there an analysis of how many users a developer would lose if they forced users to go outside the app to a browser for purchases, Even wants to know. He brings out Hitt’s deposition, where Hitt admits he did not perform this analysis. A game developer who wants to maximize their purchases are highly incentivized for IAP, but in HItt’s depo, he didn’t identify a single developer that was actively trying or had tried to encourage users to transact outside the app. Hitt is arguing about the definition of “trying,” and Even says, “they are not trying in their heart.” He didn’t study if there were discounts or advertising moving purchases to the web from IAP, Hitt admits.

Now onto substitution analyses. Hitt’s Switch analysis focused on half a million iOS users, so all he could show at most was 0.3 or 0.4 percent of Fortnite iOS users moved some of their play to the Switch.

About the commission rate: the argument is whether it’s appropriate to include free transactions in the analysis. Earlier this morning, on direct, Hitt pooh-poohed excluding the free transactions. Even is doing his best to make including the free apps look ridiculous. He does basic math, showing that an average of a $10 commission and a $0 commission is $5.

Real Rick Moranis vibe coming off Hitt during this cross, if I am honest.

We’ve moved onto data showing availability. Frankly, I’m not totally sure what this is meant to prove! The chart shows that 16 percent of the top 25 gaming apps by revenue and downloads aren’t for game consoles — and about 88 percent are available on PC. 100 percent and 96 percent are available on Android, though!

There’s some very confusing stuff about a spreadsheet that’s nearly impossible to make out from the gallery. I believe Even is trying to undermine Hitt’s analysis by showing that some games he’d said were available on PC, such as BitLife, are not, in fact, available there. World Story, another game listed as available on PC in the document, does not say it is available on PC on its website. When we go to the Microsoft store, we discover a “Word Story” exists, but it’s not from the same developer. Even: “Sir, this is not the same developer and not the same game, is it? It’s what is called a ‘fake game.’” It’s also listed on Lagged, but is “based on the popular mobile game” and by another developer. (For what it’s worth, this collection is a Word Story but not a game.)

Next, Helix Game, developed by Voodoo. No mention of PC on the dev website for Helix game. On Voodoo’s site, Crowd City is also listed as being available only on the App Store and Paper.io 2 is available only on the App Store and Google Play. Happy Glass, another game from the spreadsheet developed by Lion Studios, is listed by the developer as available on the App Store and Google play as is Mr. Bullet. Both are listed as things available on the PC, but the app developer page suggests otherwise. I wonder how many more examples of this we are going to get, because it is tiresome to go through this for every single one of these apps.

Even asks if Hitt double-checked the data to make sure it was the same developer across stores. Hitt says yes, and that there are some missing links. Even asks Hitt if he can tell the court under oath that a certain app is from the same developer. This bit gets really contentious. The point of the testimony is to show that devs multi-home to PC and mobile, Even says. “That’s what you said in your written testimony. I can read it to you.”

“Can you give the court the assurance that your 75 percent is correct?” 

“Can you give the court sworn testimony that you have seen with your own eyes that this game is also available on PC?” Not from the same developer, no. Hitt can’t guarantee all the data shows the same developer is behind the game on the different platforms.

We are now at Figure 8 from Hitt’s testimony, about monetization and portability of FY2019 game apps in the App Store. I am dreading the same exercise here. We go back to the terrible spreadsheet. Crowd City is highlighted. Hitt says he’s identified 8 games that let you buy things on the web and use them in iOS. Even says there are 10 yeses for 8 games. 

Had Hitt personally examined the possibility of going to iOS Safari, buying something, and going back to the app? No, personally he did not. His research team did. 

We are now going to go through the “frictionless” process described in direct testimony. We will use Candy Crush as our example. We scroll to the developer, King. The website says we can install, we press “install” and the link sends us to the iOS store. Among our options for playing is king.com, the only possible way to play on the web. Says it’s available on desktop — and only available on desktop. The Facebook option, too, is only for desktop.

“That’s part of the frictionless process you have envisioned?” Even asks, somewhat sarcastically.

Let’s try a different game, Clash Royale, developed by Supercell. Supercell says in its FAQ that it does not store any credit card information — payment processes are though Apple’s App Store or Google play. “And yet you believe that your team managed to go into a website and buy legitimate Clash Royale money and go back to the app? That’s your testimony?"

The typical user of Clash Royale doesn’t have a research team, Even points out.

Big Fish Casino is another app in question. Even has questions about the other five as well, and hopes this will not take so long for each one.

Based on what you have seen, do you have any reason to disagree with our conclusion this cannot be done. 

Onto developers. My team has found this can be done for three apps — one is Fortnite, one is PubG and one is Roblox. Does Hitt have any basis to dispute this? Hitt says he trusts his team.

REDIRECT. Apple’s Cynthia Richman is back up to bat.

Nothing Hitt said changed Even’s analysis, because emulators exist. Sometimes developers license games, too, he points out. Hitt says that Even’s demonstration isn’t a realistic example of the switching costs — it’s not broadly applicable. “Certainly we also have examples where it’s very, very easy, such as the purchase of V-bucks.” Hitt did purchase V-bucks though a mobile browser and found you could do it.

“It looked pretty difficult given the examples you provided,” YGR says. What’s the explanation for why we couldn’t do these things during the cross-examination?

Well, there were other links in there, and it showed games that were very similar, Hitt says.

It’s still true that most of us don’t have a research team. I sure wish I did.
Hitt is now explaining why the improvements in engineering  (iPhone and otherwise) that lead to better performance isn’t relevant to his analysis of how Apple monetizes its ecosystem. The continued innovation is what makes it possible for better games and so on to be delivered. 

Why didn’t Hitt conduct his monopoly test? One, because other methods were used. Second, in order to do those calculations in two-sided markets, you need parameters and information that aren't available. It’s not clear you can considered two-sided market effects, either, Hitt says. 

Onto the Switch. He says the Switch adoption reflects a large population in a natural experiment. He wants the population that’s “most informative” in terms of switching to the Switch.

RECROSS, EVEN AGAIN.

This is a two-sided market and this is a point about the developers side — so if someone is playing Fortnite on iOS and then goes out to play Fortneat on PC that doesn’t help. So comparable games from a different developer — that’s not something that moves the ball for someone who’s pay 30 percent? They need their own PC store, Even says. 

Re: emulators: Does Microsoft make an Android emulator? Hitt doesn’t know Apple? Dunno. Citrix? Dunno. Adobe? Dunno. Does Google? Dunno Do you know who does make an emulator? Hitt knows they exist, doesn’t know the developers.

“I’m just asking because I still remember Apple’s concern about safety and security.” Downloading emulators on the open web would not help that safety and security, Even suggests.

Now onto Switch analysis. While a lot of people have iPhones not as many have the Switch.

We are now looking at the slide that says the App Store is growing 6x faster over teh same period. It’s a much more digital world. Has Hitt done a similar comparison to the sale of phones, microprocessor tech growth, and other improvements? “This is an evolving industry.”

REDIRECT. RICHMAN AGAIN.

In many cases, the games are available, Hitt says. If a developer licenses a second developer, the first developer benefits from substitution because the game goes on multiple platforms, Hitt says. “That’s a perfectly reasonable strategy.”

Back to the digital game transactions slide. Total growth rate over a 9 year period is 448% — it’s 1/6th of the detail games. It says that there’s nothing to suggest that Apple has engaged in anticompetiive conduct, Hitt says.

Hitt is excused. Counsel from each side is invited to get up and grab their binders.

At 12:10, Michael Cragg, principal and chairman at The Brattle Group, takes the stand for Epic. Yonatan Even will conduct the direct examination.

After Cragg is sworn in, it’s binder time — several lawyers approach the bench and witness stand with binders of Cragg’s written testimony. We go through his resume, which does not include that Gawker and Ozzy Osbourne were among his clients (fun fact!). Cragg was given a broad assignment to review the Apple expert reports. He is, thankfully, quite loud.

His opinions re Apple experts: (1) Market definition: Apple experts’ market definition is both too broad and too narrow at the same time and (2) Apple has significant market power.

Apple experts should have performed a test, the hypothetical monopolist test. They start with a market definition that considers digital game transactions on iOS — but is there a larger market where the hypothetical monopolist comes into play? The starting point is to ask the question if a monopolist were to monopolize the transaction market on games on iOS, could they profitably raise price?

What would that test show us? Were that hypothetical monopolist to raise their distribution price, naturally, the publishers, namely game developers, would seek an alternative store. 

Availability of game apps on each platform: Cragg identified the most popular games that are published, and each of the platforms they’re on. The portion of the table expanded here is specifically for mobile games. Fortnite is highlighted, in a “unique position with respect to the other mobile games” in that it’s available for cross platform play. 

For the static platforms, ie the ones you don’t move around to use, they aren’t generally publishing games on mobile devices. Fortnite is, but it’s an anomaly. The Switch assumption is that it would be on a screen on the household. Switch is not analyzed in the mobile context. 

What does this tell you about the availability and substitutability of these platforms? For mobile games, we see that all games are published on both Android and iOS, because that’s the way a developer accesses those different populations. The developer has to go through these two channels. For the static games, the nature of those games is quite a bit different than on a mobile device — they’re more complex, take advantage of the bigger screen, have different ways of interacting — and as a result, those games provide a different experience, which is differentiated from teh mobile devices. They’re not substitutable, Cragg says. Also, the budget for a console game is “very much like producing a movie,” and it can be upwards of a couple hundred million to produce a game. Mobile games, by contrast, cost hundreds of thousands of dollars or less. The games are simpler, the experience has to be limited because of the size of the screen. The use of a mobile device is also when you’re not looking to be absorbed — it’s a fleeting experience.

Cragg has segmented Fornite players by platform. iOS players played an average of 47 minutes per week — whereas those on consoles are playing several times longer, six or seven hours of play on average. The platforms from a consumer perspective aren’t substitutes. Console players also spend more. The chart also suggests that multi-platform users weren’t playing the same way as those who were on a single device — the multiplatform users can go weeks without it.

We recess for lunch at 12:35.

1:15PM: Yonatan Even resumes direct examination of Michael Cragg, principal and chairman at The Brattle Group


We are looking now at a slide entitled “Within any given month, most Fortnite users are single-platform users.” Cragg says he examined that within each month, how many platforms users see. There’s an eightfold or more difference between the single platform users and others, Cragg says. “The hypothesis of the Apple experts is that multi platform play is a way of creating a disciplining force for the Apple App store, and from a practical perspective, that’s not happening in this marketplace,” Cragg says.

Where do people transact? Does this tell us anything? Cragg says the vast majority of people sign in on a single device each month — so that says the majority are not transacting across platforms. It’s going to be on the platform we’re on

YGR: Doesn’t that also suggest that a Forntite user from a consumer’s perspective, the consumer would use the same platform for some other game if they wanted to?

Cragg: yes, depending on whether they own more than one platform that would be true.

PDX-01118: “Fortnite users’ total play time increases when adding a second platform.” It’s a chart with plus signs

“So for mobile devices, for every single platform type that was introduced as a new place of play, we see an increase in total play,” Cragg says. And if someone was playing on something else and added a mobile device, the total play time increased.

In home console, there’s no increase in total play when another one is introduced. 

Cragg decided to test a substitution effect. “This is a particularly important case to look at, because if there were any games where this held true it would have been Fortnite.” If it’s not true here, which is the best case study, it’s unlikely to hold true for other games where the same type of play opportunity doesn’t exist.

When adding a mobile platform, Fornite users’ total play time increases on the original platform. “There’s a feedback where you’ve enhanced the experience on the original device.” You don't substitute away from the original device, you just play on more of them, Cragg says.

“I think the most important piece is that other platforms don’t act as a substitute for iOS on Fortnite,” Cragg says.

Now we are going to go after the Apple experts. Craggs’ results don’t show the substitutes Apple says should be there. Analyses focused on the growth rates — as we saw earlier — lead to different results, depending on the starting point. It’s possible to draw a false inference with growth rates, Craggs says. (We’re looking at the iOS transaction data from Hitt, Cragg clarifies after YGR is confused.) Here’s his example: imagine the year-over-year change were a dollar. You say there’s no difference across two populations. But the group that didn’t download started at a dollar and the other started at $10, so teh console players were already sending more. So in that case, there’d be a 100 percent increase versus a 10 percent increase. 

Ah yes, binder time. It’s time to talk growth rates. The group that downloaded the app, he used that as a decline in their consumption. The actual dollar numbers show that the group that did download the app bought more than the group that didn’t. He points to the numbers we can’t see as the cause of an incorrect inference by Hitt.

We are now discussing the Switch substitution analysis from this morning. This is Fortnite data, not Apple data. Hitt says people switched to Switch — Cragg calls it the “manipulated data.” We’ve already seen the regression results, which Cragg believes are the correct results. “This is really a fairly rudimentary review of the numbers,” Cragg says. We are now going to have a fight about statistical methods, which as a science journalist I must say feels comforting and familiar. 

Oh now we see two charts. One is Hitt’s analysis. The other is Cragg’s re-analysis with a different statistical method. It pretty much reverses Hitt’s analysis, showing the people who accessed Fortnite on both platforms spent more. There are also fun red circles around the March 2018 data, which show that the starting point is different, “a substantial difference.” Cragg’s analysis is done using Hitt’s data — just different methodology. “A substantial difference in spending between the two groups, and the Switch group spent even more.”

Another slide: it shows that after the Switch was introduced, iOS users increased total and iOS spending. There’s no fall in iOS spending, nor is there evidence that such spending went to the Switch.

We are now turning to another slide. It is titled, punchily, “Apple’s opening slide presented Professor HItt’s misleading commission rate and only for a single time period.” Hitt averaged together app downloads and IAP transactions with those where there’s no commission rate. Blending all of that together is misleading, Cragg suggests.

Cragg’s data, which we cannot see, suggests that Apple’s cut increased over time. “They’re providing the same service but getting more for it.” There are no zero transaction sin this graph.

In Fortnite, you can have purchases if your wallet is full and Apple gets nothing. How did Cragg account for those, YGR asks. “That circumstance is not reflected in this chart.’ Yes, Cragg says. If you have V-bucks you bought somewhere else, you would not be doing a transaction through Apple in that case.

Slide: “Apple’s market power is indisputable.” First, you see increasing prices from Apple, as well as high, sustained margins, Cragg says. There’s an absence of competitive pressure from other app stores because of the lack of substitution of transactions across platforms. “All of that together is, to me, strong evidence that there’s monopoly power,” Cragg says.

CROSS. Daniel Swanson for Apple.

Ah, we are now going to roast Cragg for not holding an academic position or getting tenure.  Cragg is qualified as an expert for antitrust in an administrative court, in a matter that involved opioids, he thinks. (He doesn’t remember) It wasn’t a Sherman Act violation, Cragg says. It seems like Swanson is trying to rip apart Cragg’s antitrust credentials. 

Cragg didn’t ask anyone to remove the resume item in question. YGR suggests he do so.

We are now discussing a Coca-Cola company case, where his opinions were rejected per Swanson. Cragg characterizes this differently, says only one area.. The court had problems with his economic opinions — ”wholly unreliable” per Swanson. US tax court found his methodology led to “absurd” results. Cragg recalls this, says it was about a specific calculation.

App Store is a two-sided platform, Cragg says. “There are transactions of the American Express sort happening in the App Store. In iOS, that’s clearly not the environment,” Cragg says. It’s hard to make a sweeping statement about platforms, Cragg says. The central defining question is substitutability, Cragg agrees, but they don’t have to be perfect substitutes. Cragg says you don’t always need to look at both sides of the market to see about substitutability. 

The network effects are the subject of some dispute. I’m having some difficulty following this portion of this testimony, but it sounds like Cragg is saying that at a certain point, network effects were no longer relevant to pricing. At the margin, this indirect network effect is no longer important, Cragg says. US Airways prevailed at trial on this theory — which was then reversed by the Second Circuit. 

From the perspective of the Apple App Store, the same rules apply to all transactions, Cragg says. It’s not correct, is it, Swanson asks. There are a couple of negotiated outcomes — Apple has negotiated in some instances and in others, they’ve chosen to apply rules in ways that are inconsistent, Cragg says.

We are onto the Reader Rule, which I hope I have spelled correctly. “That particular rule allows apps that fall under it to have content used within an iOS app even though it was purchased elsewhere,” Swanson says. That rule doesn’t apply to all apps, right, Swanson says. We are now onto the mulitplatform rule — there are some rules that are restrictions on how you can steer people, others that allow using currency in cross play, and others that relate to what you can advertise about pricing. The multiplatform rule allows developers not subject to the reader rule to have a virtual currency that is paid for on another platform but used in the iOS app right, Swanson says. We are splitting hairs about whether or not these rules apply to all apps; Swanson seems to think they don’t apply to all apps, Cragg appears to think that they apply to all apps but some apps don’t qualify.

Swanson: Isn’t it true that there’s no consensus among economists about how to apply this in two-sided markets?

Cragg: That’s right.

The hypothetical monopolist test you describe is a thought experiment, right, Swanson says. Apparently Cragg described it that way in his deposition. We are now looking at Cragg’s binder. We are circling around whether you can separate games from the overall app market. Swanson reads Cragg’s testimony aloud. Cragg starts bickering — is there one or two sellers of game transaction in his hypothetical monopolist test, Swanson wants to know. They talk over each other. “Is this a hypothetical duopolist test?” Swanson asks. Can you differentiate a game transaction from other transactions is sort of the central question here.

YGR: “I have no idea what you mean by that.”

Cragg. In the real world, some subsidies allow for the purchase of gas at a lower price than teh retail price for us. That gives incentive for that person to resell to us at a higher price. The way that is constrained is by coloring the gas. You and I can’t buy red gas.

I wish I could tell you specifically how this applies.

Cragg is being asked about a chart I can’t see (presumably this one: https://twitter.com/leah_nylen/status/1392954903715319808). Cragg says these are not submarkets — he says he’s trying to identify play styles. Core, midcore, and casual games are on every platform. “Casual games, such as puzzle games, are the most popular on iOS and mobile devices.” There is an ongoing theme of Swanson and Cragg miscommunicating; Cragg has to occasionally backtrack and correct himself. He also apparently can’t read the font on the exhibit he’s been given.

We move onto the slides from the direct analysis. The consumer is identified as having a second device when the consumer starts playing on that device; we don’t know for how long the user has owned the device before playing on it. “The next best thing is to look at how play changed,” Cragg says. “You made an assumption, right?” Swanson says. “I didn’t make the assumption, I understand the point you’re making that we can’t observe that specific event. To test the assumption of what's this correlated with the purchase of the platform, there’s not a way to test it.” Swanson points out you don’t know why the user went to a second device.

There’s no payment necessarily associated with play. In Cragg’s rebuttal report, he’d analyzed consumer transactions as well as play time — that was the original report. But it was corrected. Seems like Swanson is trying to make Cragg seem unreliable.

When two products are complements, when the price of the first product goes up, the demand for the second product goes down. “If the price of hot dogs goes up, the demand for hot dog buns goes down, right?” Swanson says. Cragg says he’d have to do the analysis but he understand the point being made. Swanson asks if mobile device transactions are complements for those on PCs. Cragg says he doesn’t remember. Okay, so but hypothetically if they were complements, and the price on one platform goes up, then the consumption on both platforms would go down, Cragg says. 

When Fortnite releases a new season, it attracts old players back and drives a surge in revenue, Swanson says. It’s not a perfect test but looking at expenditure more directly, it’s closer to the definition of what economic substitute and complement are, Cragg says. The month before Fortnite was released on the Switch was May — so looking at that line and looking at September, it’s almost the same bar as in May. It includes Switch revenue. So it’s a smidge below the bar in may, that implies that revenues on all other platforms declined, right, Swanson says. Cragg agrees. The bar chart implies that Switch has taken share from other devices. Swanson wants Cragg to say this is substitution — but Cragg says you need to do a regression analysis. 

We go through a number of platforms and whether they regard their own online stores as competitors to the App Store (Nintendo, Sony, Google). Cragg says he doesn’t know how they view each other. We then establish that Cragg has an ongoing relationship with Microsoft. He says his confidentiality agreement wouldn’t allow him to answer with specificity what he’s analyzing. The binders come out for the deposition. Deposition says he’s looking at entire portfolio, including xBox. His work includes the competitive properties of Microsoft’s work. He wouldn’t answer questions about pricing, competitive advantages and so on in the deposition.  Cragg doesn’t know whether the xBox is viewed as a competitor to the app store by Microsoft, he says.

Epic Game Store charges a 12 percent commission and EGS’ commission is below its current average cost, Cragg says. He also says Apple’s prices have been increasing. Cragg heard the oral testimony for Evans, but didn’t read his written testimony carefully. Evans spoke to a slide about reasons why Apple had monopoly power — Cragg doesn’t remember the specific slide. The commission rates have been stable, but Apple’s average commission per download in dollar terms has increased, Cragg says in his deposition. Developers set the price for IAP — with some limitations, like that the price has to end in 99. Cragg says he heard much of Hitt’s testimony but missed the redirect. Developers on average increased the amount of purchases by 400 to 500 percent in that testimony. “That increase in prices by developers is why you say Apple’s dollar value on IAP has increased dramatically,” Swanson says.

“Isn’t it true that the average price of a game downloaded on EGS is about $25.” Cragg hasn’t looked at the price of app downloads in the store. But if that’s the average, then Epic gets an average of $5 on the sale. The average paid on iOS game download that Apple gets is 16 cents.

REDIRECT: Yonatan Even

If the price of Fortnite on iOS goes up, it would lead to a reduction in iOS purchases, Cragg says. So for the people who play on PC whose best friends play on iOS, they’ll play less too because that’s how complementarity works.

We return to the resume. Cragg says he assured YGR he’d make sure what was on the website was accurate.

With respect to the other issue, we can move into a sealed session for the next 15 minutes, YGR says.  We go into closed session at 3:05 and I am very politely told to get out.

 

Elizabeth Lopatto

Deputy editor, The Verge

510-612-2139

@mslopatto