Epic v Apple Day 10 - Daily Pool Report
|
Dave, Paresh (Reuters)
Hello, Paresh from Reuters here with Michael Acton from MLEX. I’m at 415-565-1302 or paresh.dave@.... We’re on the judge’s far right in rows 2 and 3 of the gallery.
A reminder that exhibits can be found here: http://tinyurl.com/epicvapple
Judge Yvonne Gonzalez Rogers (YGR) is on the bench in a bright pink top, with a stone paneled wall behind her, the US flag behind her right shoulder and the US seal high on the wall behind and above her.
The court reporter and clerk are in front/below the judge and the attorney asking the questions is on the other side of a shield as the court reporter.
Epic’s attorneys are seated in rose brown leather seats around a rectangular table that’s in line with the court reporter, attorney podium and judge. Apple’s rectangular table with the same seats are on the judge’s left, beside the Epic table. The jury box is on the judge’s right with a couple of people unknown to me sitting in it.
Apple’s folks have Crystal Geyser normal-sized water bottles on their table, Epic has mostly the tiny Nestle water bottles and couple normal-sized Arrowhead bottles.
There’s two computer monitors on each table with the slides, one big screen facing gallery, one big screen facing judge, one monitor in the witness box and two monitors on either side of attorney podium.
Today’s teams include:
Epic: Katherine Forrest Gary Bornstein Yonatan Even Monica Coscia Tim Sweeney
Apple: Richard Doren Heather Grenier Jen Brown Dan Swanson Phil Schiller
*** KF asks for time to be set aside for closing argument. YGR skeptical.
YGR: So nothing has really changed, I heard your opening, read your briefing, the hundreds of pages. So perhaps for the public you want to do this?I just don’t know how it’s going to be any different. What would be most helpful, and what I don't think you would be interested in doing Ms. Forrest, is what I do in my patent cases. For every term that the lawyers are disputing, both lawyers are up here and we go back and forth. It’s more like a debate. I take each issue and each side gets to go back and forth and respond to my questions and each other. Your [arguments] talk past each other very frequently...Until I have a chance to go back and really assemble and re-assemble and look at all this evidence, another single presentation doesn’t help me. If the two of you are willing to stand up there and debate these elements, I would find that helpful.
KF: We would be prepared to do that.
RD: Certainly would be fine with us.
YGR: It doesn’t seem to me that anyone really disputes the legal approach i took with the preliminary injunction order. If each of you have a shared framework, we could go back through that legal framework and work through it that way. ...If you cannot agree that is what I would use myself. These are important cases. And it’s tse factual elements that are going to drive the decision. If you all want to agree on a one page outline of the various elements, then we are all on the same page. … We would do this (back-and-forth debate) after close of evidence.
KF: We can meet and confer over the weekend and present court with an agenda Monday, so court can weigh in. And be prepared to do this next week.
RD: (asks to present on Monday the 24th)
YGR: (Agrees, says she expected case to hang over and had already reserved courtroom for 24th)
First up David S Evans, chairman at Global Economics GrouP takes the stand for Epic with a face shield, examined by Gary Bornstein
DE says he listened to Schmalensee, Hitt testimonies.
GB asks about Hitt’s testimony about Apple’s success.
DE: There has been tremendous growth in the smartphone business over the last 10-12 years and there’s been incredible growth for Apple.
GB: Does that cause you to rethink your views?
DE: Not at all. It’s useful background. But in terms of being relevant to assess conduct here, it’s not relevant. It doesn’t permit antitrust economist to make causal inference about whether the conduct has interfered with competition.
GB: Do you recall Hitt’s criticism of your testimony? And that one of the criticisms leveled was there was not an obvious change in trajectory of Apple’s growth. What’s your reaction?
DE: That’s not a very sensible way of thinking about market power in this context. It’s not like a switch goes off. The other point very important for this kind of business is in a rapidly growing industry it’s not possible to look at time-series of growth and refer to the growth to [assess] changes in market power.
Brings up DX-4767, showing 115 million total users and 35.9% being multi-platform users. From Hitt’s testimony.
GB: Can you tell us what a population of Fortnite users this slide covers?
DE: The title says iOS Fortnite user accounts. It’s anyone who has touched iOS. Someone who logged in for five minutes to someone who uses it all the time. It’s the full range of interactions. The second dimension is time period. This slide covers march 2018 to July 2020, it’s about a 29-month period.
GB: What does 35.9% mean?
DE: I would say 64.1% were single-homing on the iOS Fortnite app. Over that 29-month period, 64.1 were only using IoS. What it says is there is a large group of individuals who are only using their iPhones to play Fortnite. That tells me first of all there is a large group of individuals who just like using their smartphone to play. I can’t say this with 100% certainty but they probably don’t have a game console.
Brings up “switching after hotfix” slide in DX-4824 from Hitt’s testimony
GB: 81-88% of Fortnite spending by iOS users was retained. What does that mean?
DE: The calculation is the amount of spending that Epic had with the Fortnite app for this group of users after this app was removed by the App Store divided by how much spending would have occurred as an estimate had it not been removed.
GB: Is it a good measure of substitution?
DE: It’s not a measure of substitution. Let’s take an individual spending $100 a month. They spend $80 using PS4 and $20 using their iPhone to play Fortnite. Before removal. In this extreme example, let’s imagine this person continues to use their game console but without having access to Fortnite app, they drop from $20 to $0. He would calculate in this example a retention rate of 80%.
YGR: If someone has access to both devices and they have V-bucks and they know they can spend it they don’t spend it just because they don’t have the phone?
(Judge appears frustrated, swivels chair other way, that she doesn’t really get an answer after GB tries to ask her question a different way.)
DE: The better measure of substitution is how much of the $20 was spent on game console. That’s the number we want to focus on. Out of the $20, what does that user end up switching over.
YGR: What if they spend $20 on other games on iOS?
That would be a different form of substitution. That is a form of substitution that could very well happen in the marketplace. But in that case, they are not substituting to game consoles. They are substituting within the iOS ecosystem. ….That behavior is consistent with a market definition that does include PCs and consoles.
GB asking several basic questions about Schmalensee testimony criticizing Evans’.
GB: (You were in US v Microsoft. What was market definition accepted by court in Microsoft case?)
DE: OS for computers. It included Windows. It included Linux, which was and is a free operating system.
GB: (Do you agree with Schmalensee saying you’re also condemning the business models of gaming consoles?)
DE: The analysis I’ve done in this case has been very specific to the facts and circumstances in this case and the analysis of the operating system. In the case of game consoles, that’s a very different business. In the foremarket, there’s competition to sell game consoles. In game consoles, the main thing in the aftermarket consumers are buying is games…. It is very hard for me to imagine it would have the anticompetitive concerns we have in this matter. … Game consoles as we talked about is an industry. When we talk about iOS or the iPhone, we’re talking about large platforms that are foundational to the digital economy. They are foundational platforms that support the app economy, which is a big portion of the digital economy, which is sucking up more and more of the physical economy.
YGR: Are they utilities? Given you keep calling them foundational. Are smartphones utilities?
DE: I would not use the word utility as an economist.
YGR: In the antitrust context? It sounds like you are saying these are so fundamental they are utilities.
DE: It’s a matter of gradation. The electric company that I depend upon to have electricity to function. Economists would attach the word utility to that because of the history of utility regulation in this country. I’m not prepared to say not being able to access a smartphone at this time rises to the same level as not being able to access electricity or water.
Oftentimes when economists talk about utilities we are talking about one company and it’s an essential service.
YGR: Let’s say they are two companies and that’s why despite the fact apple and android have created all this technology, they give it to developers. I’m having a hard time understanding the gradation given how essential you say it is.(she says while moving her left hand up and down in the air while chair swiveled to the left to face Evans, who is in the witness box on her right. ) Do you don’t think the government should break them up to add competition like they did with steel?
DE: I am a moderate on that topic. There is significant debate going on concerning what do with the technology companies including Google and Apple. I would not advocate a reg solution or breaking them up. I would advocate that the antitrust laws could deal with these companies. I would argue it would be a mistake to go down the road of extreme solutions. Antitrust law is a flexible method that can be useful to reining in competition in the digital economy. I would have colleagues that would take a more extreme view.
Okay, that’s all from me.
GB: (Asks about Schmalensee testimony - brings up slide PDX0043-008) Asks about “second lane” in slide. Physical in-app purchases.
DE: The starting point for this exhibit is with respect to the install base of iPhone users, these are the only roads into town. These are the only ways for developers to get consumers with an app. There is also the notion of having a toll on the path.
Brings up WRD-005...paragraph 4 of DE’s testimony--- Apple’s experts do not accept some of these principles…
GB: Do you agree with Lafontaine?
DE: I do not. The starting point for the market definition is the supplier. (Then) what substitute suppliers are out there to which consumers could turn. ….It always comes back are there enough customers that have good enough substitution options that they would move and discipline the hypothetical firm with market power.
DE takes a moment to sip water from a plastic bottle (he takes another one amid the questioning by Swanson) and Swanson/Apple takes over questioning with GB planning to ask more questions in closed session.
DS: (View Apple as providing an essential facility?)
DE: The way I am using gatekeeper and foundational platform, I am not connecting to the legal term of essential facility.
DS: Aren’t you saying Apple has to offer its IP for free?
DE: I am not saying that.
DS: If Apple were required to offer the typical deal to avoid anticompetitive conduct, wouldn’t that be the antithesis of innovation? Asks several times, Evans doesn’t understand each time and DS moves on.
DS: Who are the successful mobile OS providers in the U.S. who have offered access to developers for distribution for completely free?
DE: You’re combining two things together. Apple requires certain developers to pay for app distribution, and Google does as well. Google Play and the App Store are the two app stores globally outside of China and in the United States.
DS: (What about Windows Mobile?)
DE: Windows Mobile I believe had the same model as Apple. They provided access to the platform for a nominal fee and they had an App Store and charged a commission.
(Several basic questions)
DS: Does a firm have right to decide who to deal with?
DE: As a general matter
DS: Restricting duty to deal to extreme cases?
DE: I do [agree]
DS: Do you agree if a firm went to monopolist and said for the sake of competition give me a loan so I can compete with that that would not be anticompetitive conduct?
DE: I agree.
DS: Asks about paragraph 49 in rebuttal written testimony about average dollar commission paid by developers has increased dramatically. You agree each developer selects the price of IAP?
DE: Yes, subject to Apple’s guidelines.
DS: You understand from Hitt that average IAP transaction price has been rising?
DE: Yes
DS: You understand developers have raised prices 400-500%?
DE: I believe there is an exhibit from Professor Hitt that says that.
DS: You didn’t study why there has been that increase?
DE: That’s true.
DS: You don’t have a firm opinion on why developers have raised prices?
DE: If that’s what I said in the deposition, I don’t disagree.
DS: Anti-steering in AmEx. Do you understand they prevent merchants from expressing preference for non-AmEx cards?
DE: Yes in the sense that if I an AmEx cardholder go to a merchant, the American Express anti-steering rules prevent the merchant from doing things to dissuade the customer from using that card, or otherwise interfering with the consumer’s preference to use the American Express card. It prevents the merchant from discouraging… I do [view that as anticompetitive]. It’s targeted to a situation where a consumer with multiple cards has a desire to American Express card and a merchant has put logo on door but turns around and says to consumer - I'm saying this imprecisely - I would really prefer you not use that card.
DS: The purpose of the antisteering provision is to prevent merchants from steering business to American Express’ consumers? Yes?
DE: The rules discourage the merchant from preventing the consumer from doing what they would otherwise do.
DS: But they don’t prevent from steering to debit cards, checks or cash?
DE: With respect to cash, you’re right. With respect to debit cards, I would have to get a refresher course. ...My recollection is they prevented merchants from steering customers to other credit cards, not to cash for sure. The part I didn't recall is what the rules had been for debit cards.
YGR: Is Apple saying it is a competitor to developers?
DS: It is a competitor to platforms developers would steer business to.
DS: Aren’t you assuming those other platforms and websites are competitors?
DE: The two situations you talked about Mr. Swanson are absolutely positively not analogous.
DS: I asked you a question about whether they are competitors.
DE: For conducting that transaction, given the prices, those are alternative avenues and therefore substitutes for doing the transaction.
DS: There must be independent competition for an anti-steering provision to have a rationale
DE: The rationale is it prevents developers from undoing the tie, which is different from American Express.
DS: If you're right that Apple has no competitors, then why is Apple concerned about steering?
DE: Because with a 30% commission, developers have an incentive to direct consumers where they have an ability to operate on a different platform in the absence of anti-steering rules to have consumers not use IAP but make that purchase somewhere else. To prevent the developer from informing the consumer that there is another alternative available to them.
DS: Apple’s rules prevent developers from providing any information about purchases on other platforms?
DE: My understanding of the app store’s antisteering restrictions is they prevent developers from providing any information, any call to action that would provide that app user with information that they could or should go off the app store and somewhere else for that transaction.
DS: You testified Epic is not allowed to message users that they could go to the web and get v-bucks more cheaply?
DE: I don’t recall that statement without qualifications.
DS: Does epic sell v-bucks more cheaply on the web than they do on consoles?
DE: The price on consoles and the price on the web is the same obviously they don't sell them on the ios app.
DS: And that’s obviously a choice they made.
DS: Do they sell V-Bucks more cheaply on the web than any other platform?
DE: Not to my understanding.
DS: (Did you familiarize yourself with Epic’s communications to users on different options?
DE: (No didn’t collect data on that)
DS: Would you agree competitors shouldn’t get a free ride?
DE: Yes
DS: Are you aware Spotify has anit-circumvention rules for artists?
DE: I don’t know one way or the other.
DS: Has Spotify ever been a client?
DE: I am not free to disclose that.
(Same questions around Samsung Galaxy app store, Airbnb, eBay having anti-circumvention rules. DE says Airbnb and eBay do have rules. Not sure on others. DS moves on to asking various questions about App Store Review Guidelines)
DS: Who would decide if Apple’s criteria were objective enough?
DE: If we’re talking about this case, it would be ground rules typically set with antitrust remedies concerning preventing the party from engaging in anticompetitive behavior. Typically when remedies are established that’s something thought about.
DS: You’re saying the court would decide?
DE: Or no one would decide because the ground rules are clear. You can’t engage in antocompetitive behavior.
YGR: Would a rogue app store be one that offers pornography?
DE: So I don’t know if a good answer for you.
YGR: That’s not stark enough for you?
(DE starts speaking, judge speaks over)
YGR: I’m trying to understand what you classify as a rogue app store
DE: The only reason I am hesitating is..in the case of an app store that had pornography i would like a world in which Apple is able to prevent that. I do think that is something Apple could do. The one - I don’t want to make it a qualification- but it’s not like the iPhone is designed now in a way that iPhones can’t access pornography. Apple has a matter of business policy could decide it doesn’t want pornographic app stores to use iOS.
DS: In the but-for world, Apple isn’t allowed to decide who competes with it with an App Store but it could revoke a license it is compelled to give if an app store has gone rogue?
DE: I don’t have a view on what scrutiny it could give before and after.
(DS asks several questions about privacy, meanwhile he’s getting little notes from Apple’s table handed off to him.)
DE: The whole issue here - we’ve talked about privacy - we’re basically to Apple being able to dictate what other App Stores can compete. ...It’s possible Apple could make decisions with respect to privacy that are for pro-competition reasons or ...it could be for advantaging its advertising-supported business model.
DS: Do you view competition on Android in China as the road you would like to see Apple travel?
DE: The road I would like to see Apple travel is not replicating China. I would like to see it having competition. In my but-for world, Apple still has the App Store, is able to bundle it with the iPhone and there is the possibility of competition.
DS: Do you believe Apple’s App Store is a monopolist in China?
DE: (Yes)
(YGR asks for ballpark estimates of iPhone market share in China.)
DE: I think iPhone in China is above 20% of total smartphone revenue.
(DS asks about malware, IP theft on Android Chinese app stores. DE says he’s not aware of these issues.)
DS: Isn’t it true there’s 50% commission on game apps in those stores?
DE: (Yes that’s headline rate, but there’s negotiations)
DS: (Asking about how DE’s analysis applies to console industry. Do you realize Sony was sued since this trial began for similar antitrust issues?)
DE: (Not aware)
GB takes over questioning.
GB: Do you know what Supreme Court said about whether AmEx had market power?
DE: The government had not proven AmEx had market power.
GB: (Anything in anitsteering of AmEx that prevents prominently displaying saying accepts Visa, etc?)
DE: (There is not, whereas Apple essentially prohibits that in its similar rules)
GB: Does market power mean a firm has absolutely no competition?
DE: No. When a firm raises its price and has substantial market power, consumers at the margin think about substitutes. To put it plainly, boy this price is so high what else could I turn to to defeat it?
GB: What is purpose of antisteering rules then?
DE: It is to prevent customers from seeking those other alternatives.
Breaking at 10:15, returning at 10:35 in sealed session.
***
GB resumes questioning DE at 10:58 with a couple of basics.
DS resumes recross at 11:00. Phil Schiller and Apple’s two internal attorneys, both carrying iPhones, return at 11:00 in the courtroom.
I would note that the two attorneys’ computers easily visible to me at the Apple desk are not Apple laptops.
DE excused at 11:02.
Epic calls Ned S Barnes, managing director at Berkeley Research Group (BRG) and a CPA and forensic account. He is wearing a face shield. Being questioned by Brent Byars.
NB: I was asked to investigate, analyze and report on operating profits for the Apple App Store and certain online marketplaces, based on parameters and information provided to me by Dr. Evans. BB: What is operating profit?
NB: Operating profits is basically the income that is generated by a business or business unit from its core business activities. It’s revenue less the expenses required to generate that revenue. There are two expenses principally excluded, interest expense and income taxes.
BB: Did you see Schmalensee saying operating margin is not a measure of profitability? (And his example)
NB: Yes. // I disagree with that. () Depreciation expense is how businesses and accountants account for the cost of acquiring capital assets.
BB: What were your estimates for Apple?
NB: 79% in 2018 and 2019.
BB: What steps to get those estimates?
NB: (looked at documents and depositions and then specifically P/L statements for the Apple App Store and Mark Rollins’ deposition about certain expenses missing. I tried to estimate what those operating expenses would be and sued those to derive my estimates). After my report was submitted, additional documents were made available produced by corporate finance department at Apple that came directly from Mr. Cook’s files and were circulated with the CFO. They calculate operating margin for App Store and other business units.
Veronica S Moyé objects, saying sealed information. BB says it comes from unsealed testimony. So they look up paragraph 2 of his written testimony. Sentence of operating margin percentages there.
NB: These documents show Apple calculated its own operating margin at 77.8% for FY2019 and 74.9% for FY2018
BB: Were you able to determine that Apple had allocated costs to its App Store?
NB: Yes
BB: Could you compare to public financial reports? Confirmed that they match?
NB: Yes, and I was able to total them up and agree to what Apple reported publicly for its companywide revenue and expenses.
BB: What is a joint or shared cost?
NB: Generally an expense that has benefits to more than one business unit, product line or service line in an organization.
NB: (Reports I reviewed clearly showed Apple was making these allocations respectively and prospectively to calculate margins for different units.)
BB: (How did you ensure your comparisons were substantially similar between App Store and other marketplaces?)
NB: When a company resells goods, including digital goods like apps, there are two ways revenue can be recognized under GAAP. If you think about App Store for example, when a hypothetical app is sold for $10 and Apple keeps $3 commission, Apple recognizes just $3 as revenue. That’s revenue recognition on net basis. If Apple were to recognize revenue of $10 and report cost of sale of $7, that’s revenue on gross basis. (Apple looks at net basis). I wanted to limit my investigation to marketplaces that recognized revenue on net basis (as a result). I wanted to identify firms primarily operating online marketplaces or they had separated information on their marketplace activities.
VM, one of the few people of color in the courtroom, takes over questioning for Apple. Starts by asking about Sweeney’s testimony on how Epic accounts
VM: (Isn’t any attempt to spread out shared costs artificial / arbitrary?)
NB: I have not investigated how Epic internally accounts for its operation.
(...Several similar questions pulling from Sweeney’s testimony… Schiller and RD exchange whispers and laughs at one point during this time.
After VM expresses frustration about NB’s answers going beyond her direct question, VM admonishes him and then asks again,)
VM: Do you have an opinion as to whether it’s a good way to run a business to focus on innovation rather than business unit P&L?
NB: I don’t have an opinion on that.
(NB says he made some adjustments after Rollins’ testimony that lowered the op margin estimates)
(KF leaves the courtroom, holding her phone, as VM keeps asking about NB’s calculation methods)
VM: (Did Rollins talk about other categories, for which you didn’t make adjustments?)
NB: I don’t recall that.
(VM asks to look at Rollins deposition, 216, line 9)
(KF back in the courtroom)
VM: (Does that call you to recall that Rollins may have been thinking there were other categories?)
NB: I think he provided a pretty comprehensive list.
NB: (There’s actually four adjustments. One category is three)
VM: If the level of expenses attributed to the App Store were higher than you estimated, wouldn’t the margin decrease?
NB: Yes it would go down. That is math.
VM: You don’t know how the documents were used, How often they were prepared. Are both of those statements true?
NB: I don’t agree with the statement.
VM: I am talking to you about the three documents that confirm your conclusions. Are you with me? Have you seen testimony about those three documents?
NB: No. I have seen no testimony.
VM: And you don’t know if they were prepared on fully burdened basis? And if an Apple witness were to testify it was not calculated on a fully burned basis?
NB: No that would be incorrect. I am not assessing the credibility of witnesses. But that statement would be incorrect.
VM: Do you know the level of cost for R&D?
NB: Generally yes.
VM: Do you feel like you know that costs have been properly attributed to the app store in these new documents?
(NB starts talking)
VM: Answer my question, sir
NB: The answer is I don’t know.
VM: Now let’s talk about your online marketplace work. (VM’s ring and gold watch on your left hand and wrist sparkle as she raises her arm while asking couple of questions about those calculations)
VM: You didn't search for online marketplaces comparable to the App Store?
NB: That is correct.
VM: Did you try to calculate op margin for Google Play Store?
NB: They did not have sufficiently separable information. No I did not.
(Goes on same way)
Sony?
No
Microsoft?
No
Samsung? Nintendo?
I did not
VM: Those companies do not report separate financial reports for their app stores, just like Apple. Yes?
NB: I’m struggling with the last part of your question.
VM: Does eBay operate a hardware business?
NB: I don’t believe so.
VM: Does Etsy? Alibaba? Mercadolibre? Rakuten?
NB: I don’t think so.
Epic’s BB back on redirect
BB: Why no testimony about the later documents?
NB: The documents were produced on the last day of discovery in this case.
PX2391, which is under seal, being discussed. PX2385, also under seal and also being discussed.
BB: It depicts the way in which Apple assigns costs, whether they be direct, shared or allocated costs?
VM objects for foundation
BB: Did you use this page for your analysis?
NB: Yes. It was indicative that Apple was using a rigorous method on how to allocate costs.
NB: Right hand side of document has figures of costs that go in different product buckets. I was able to reconcile that with other pages of this presentation, where the App Store’s cost allocations were flowing.
BB: What other pages?
NB: dot-11 and dot-12. / (For fiscal year 20, a forecast period, shows breakdown of revenue, gross margin, operating expenses and operating margin across different Apple businesses). I was able to tie this to dot-24.
BB: How do you tie something out?
NB: It’s just simply an accountant’s term for taking two sets of data, adding up the numbers and getting them to agree.
(Schiller rocking back and forth in his chair. Currently wearing glasses, but sometimes takes them off throughout the morning)
YGR: Dot-24, looking at iPhone number?
NB: No, your honor, I was looking at last item on right-hand corner.
NB excused at 12:14
***
Peter Eric Rossi, professor of marketing, economics and statistics at UCLA, next up for Epic. Expert in survey methodologies and statistics. Lauren Moskowitz asking the questions.
LM: What was your assignment?
(Exhibit PDX-0091.4 on screen)
PR: (Undertaking reactions of consumers to a 5% increase in cost of certain purchases on App Store.) (Slide say in-app purchases and subscriptions)
LM: You asked for actual purchase data from respondents. For how long?
PR: 30 day period. (The quality of the data would be compromised if went too long. And people also think about purchases on monthly basis)
YGR: When was this?
PR: January of this year. I believe survey was administered late January.
YGR: After the holidays?
PR: Yes
From the slide:
Hypothetical Scenario
“Imagine that, starting 30 days ago, the Apple App Store increased the prices of all IAPs/subscriptions by 5%.
You told us that your spending on IAPs/Subscriptions during the past 30 days was $4.04. The higher prices mean that the same purchases would have instead cost you $4.24.
Nothing else about the apps or IAPs/subscriptions has changed. Prices at other app stores (e.g. the Google Play store) or websites remain the same.”
Another slide shows this survey question: Thinking about the same 30-day period, would you have made the same purchases of IAPs/subscriptions from the Apple App Store with the higher prices?
Yes, I would have made the same purchases and spent $4.24
No, I would have changed my purchases and spent less than $4.24
Not sure what I’d have done
YGR: (What about the fact things are on discount during holidays and consumers believe that prices rise in January?)
PR: That I don’t think applies to this class of goods. And purchases happening in post-holiday period.
(Next slide shows Q17 about where costs would have shifted PDX-0091.6. Other questions about how much spending would change)
LM: (What were results?)
PR: The total sample size was about 2,600 (2,595 to be exact). 81% were stickers, 19% were decrementers. 1.3% were switchers. -11% overall spending reduction. Elasticity -2.19 (+- .66)
Break at 12:36 until 1:15
****
(Parties agree withhold the Rossi exhibits from the Box until the two sides can agree on what can be publicly released)
Back in session with Rossi on the stand
(PDX-0091.10 on the screen about the reliability of survey results to show survey population representative of target population, had small margin of error and robustness testing.
LM asks some questions to walk through those^)
VM begins cross of Rossi
(Asking about draft of survey)
VM: In the initial draft you were asking survey respondents what they would do about hypothetical price increase in the future?
PR: Yes
VM: You had never done a structured pre-test ever before?
PR: No I had not.
(VM also calls out Rossi through question for using the word “permanent” in terms of price increase his written testimony but not having word “permanent” in the question that was actually used, which he noted on direct was a key word purposely omitted because it was too ambiguous)
VM: Weissinger (Matthew, the Epic VP) said in May 10 testimony that primary demographic are boys/young men 13-25.
PR: I haven’t seen any research on their demographics.
VM: If it’s true that Epic’s core demographic is 13-25, your survey only targeted people 17 and older?
PR: That’s true.
VM: Does that mean your survey results represent people 17 and older?
PR: Yes but may represent more.
(VM through questions points out survey offered starting Jan. 30, so 30 days prior included holiday period purchases)
(Summary statistics on spending- figure 9 on the screen from Rossi report)
VM: 25% of respondents responding spending $5 or less.
(Slide shows 50% spent $12.99 or less 75% spent $29.97 or less
Apple’s transaction data from Jan. 2019 on 25%, 50% and 7% is redacted on slide)
(Rossi exhibit, page 36, question 17 Asks What would have done to spend less than $4.24 in the Apple App Store? In considering your reposes, keep in mind potential costs and time required to shift purchases to other devices. See more detail.)
(The see more detail -- For example, the cost of the new device and accessories, installing and/or repurchasing apps from a different app store, compatibility with other devices and learning how to use new features of the device and the apps.)
(Options for the question 17- Would have kept my IAP/subscription spending in the Apple App Store but spent less than $4.24 Would have shifted some or all of my IAP/subscription spending to my other existing devices Would have shifted...to a new device None of the above)
LM on redirect of Rossi
LM: Why did you do structured pre-tests here?
PR: One unique feature of this survey is we ask people to enter their actual purchase history and what I learned from process of pre-test is IT’S hard to monitor conscientiousness of entering it. We’re better able to see the reliability and diligence of respondents. Made substantive changes based on that, like introducing reliability test.
LM: Does Fortnite demo have anything to do with survey?
PR: (No looking at all in-app purchases. Our data would reflect joint purchases made my older household members. Also know parental consent makes it difficult to get reliable data from under-17 population.
PR: Including paid apps would not affect my conclusions. (Reviewed highly confidential Apple docs to verify this)
YGR: Where can I find your data that shows me your thinking process for taking out words and getting to this formation?
PR: It’s described in detail in the appendices.
LM: It was not attached but can add it.
VM on recross asks couple of small questions.
YGR: You didn’t do any comparisons of spending habits of downloads vs IAPs/subscriptions?
PR: I just reviewed documents that looked at magnitudes of that spending. PR excused at 1:58pm.
***
James Mickens, professor of computer science at Harvard University and former researcher at Microsoft, now on for Epic. His first time being an expert witness, he says. Justin Clarke asking the questions.
JC: What was your assignment?
JM: I was asked to review security of iPhones.
JC: What were your conclusions?
JM: I found if you look at security properties iPhone provides to its user, most are enforced by the operating system, which in this case is iOS. When we look at the benefit the app review provides, it provides minimal additional security benefits. In lay person terms, the safety experience on the iPhone is mostly guaranteed by the OS. If third party app stores were allowed, that would not lead to a meaningful less secure experience.
(Showing PDX-0081.1 and .2)
JM: Describes middle layer as most important (slides show that as -- digital signature validation, sandboxing, address space layout randomization, execute never, memory isolation, kernel integrity protection, page protection layer)
YGR: Why does Android have more security issues than the iPhone?
JM: If you look at various commodity OS, they are in the same equivalence class when it comes to susceptibility.
YGR: So you don’t think there is a difference?
JM: I don’t think there is a meaningful difference. Why am I qualifying it, if we look at the big picture, they are roughly equivalent.
JC: (Walk us through how OS works)
JM: At lowest level is kernel, similar to chef in restaurant. Then middleware, like wait staff. When application wants to do something it sends request to middleware, which sends it to kernel, which then decides whether to fulfill and how. (Compares customers to apps and hardware to kitchen)
JM: It is the most important security layer because when we think what makes an application malicious, what makes a program malware, that’s a behavioral definition. It’s bad because it tries to do bad things. Those attempts to do bad things are happening on the client side. When we think about protecting users, we think about how to protest users on that device. That’s why the OS is the most important layer of security.
JM: (Explaining memory protection for the kernel. Compares to person, bookshelf and desk. The bookshelf is a storage device full of things apps might need. But when apps want to work on a specific set of things, it needs to bring info from bookshelf to desk. These protection mechanisms prevent stealing information.)
(YGR looking at JM throughout his testimony, moreso it seems than with other witnesses as he explains all this tech and he looks in her direction)
YGR: Isn’t there risk of someone inadvertently having that on their device and then the whole system is compromised? That’s why you don't open emails where they come from?
JM: Sandboxing helps address some of those concerns. It’s one of the critical mechanisms to limit the scope of damage if somehow a malicious app ends up on the device.
YGR: (So even if an app ends up device through some means outside of Apple’s control, this will offer protection?)
JM: If you look at how sandboxing works, all of these security mechanisms are agnostic to the method by which the app gets onto the phone.
JC: Are of the layers you just described dependent on the means by which the app arrives on the device?
JM: No. It’s all about what does that application try to do once it’s there.
YGR: Have you ever created an app?
JM: Yes but none publicly released
YGR: Have you yourself gone through the process with Apple?
JM: No
JC: What security properties does review process screen for?
JM: (Per public facing guidelines it purports to screen for sandbox compliance, exploit resistance, malware exclusion, user consent for private data and legal compliance)
JC: (How does it compare to OS review?)
JM: (Walks through PDX0081.11 showing sandbox compliance, exploit resistance, malware exclusion can all be enforced by the OS and that user consent app review is weak at best and legal compliance is difficult by both OS and App Review.
YGR: (Are you saying there is no value in human review?)
JM: What I am suggesting is the benefit of human review is marginal at best.
(Apple objects to these questions since JM didn’t address human review in his report.
JC moves on to asking about Apple Developer Enterprise Program and Internal Apple Testing Distribution)
(Now showing PDX-0081.13 App Store, ADEP and IATD -- all protected by on-device security. But no protection from off-device security (app review, developer identification and code signing).
ADEP has no App Review, partial developer ID and code signing is not by Apple.
IATD has no App Review, does have developer ID, but no code signing.)
JM: Apple already has other distribution mechanisms.
JC: (What about macOS?)
JM: Apple looks at the app, checks for malware and if answer is no, Apple generates a signed message saying we didn’t find malware in this app and then the third party developer is free to distribute the app by any method of its choosing and it is signed by developer, not Apple.
YGR: In your industry, is iOS viewed as more secure?
JM: Google people wouldn’t agree. In my expert opinion, there is rough equivalence between Android and iOS.
YGR: I said macOS. Is iOS more secure than macOS?
JM: I wouldn’t say it’s meaningfully more secure.
JC: Are iOS and macOS different?
JM: (Not totally). They share plumbing.
YGR: Did you review proprietary data to draw these conclusions?
JM: No
YGR: If Apple’s experts come in and say these systems are different, you wouldn’t have a basis to disagree?
JM: (Says he can prove core components exists by looking at public materials, books written by Apple people, reverse engineering done by others. Speaks of the shared “Darwin” OS kernel).
JM: That shared kernel is responsible for memory protection, the infrastructure/the plumbing that makes sandboxing possible.
JC: How does security get enforced on macOS?
JM: There’s a lot of overlap (with iOS).
JC: Does macOS do anything differently to enforce security?
(Notarization, Gatekeeper and Malware Scanners now highlighted on slide PDX-0081.18)
(JM talks about how they could be brought to iOS)
YGR: But they would have to create it new?
JM: The code that would have to be written would be quite minimal.
(JM explains Gatekeeper tries to figure where the app came from, distribution wise and based on that may prompt user, this app hasn’t been signed by anyone or isn’t notarized)
JM: It shows empirically Apple devices already support multiple distribution models. Second, in those distribution models, the one-device security measures are supported.
JC: Would opening up app distribution mean Apple’s App Review Process (would have to go away)?
JM: If iOS were opened up, the App Store wouldn’t have been changed at all.
JC: Why focus on macOS vs Android?
JM: MacOS and iOS - these are two operating systems which Apple itself has written and advertised as being extremely secure. But macOS which has been touted as very secure allows distribution channels in which apps have not been reviewed at all. That’s why the comparison is so important. YGR: Do you know how many people fall into this third-party developer category for macOS?
JM: I don’t know
YGR: Do you know how many apps?
JM: I don’t know
YGR: Would it make a difference if the third-party developer number were so small it could be managed vs hundreds of thousands of apps that go through this process?
JM: Well, I think that it’s a subtle and complex topic. Fundamentally what determines whether an application is malicious. It’s behavior. Let’s say the third-party universe was so small, the manager has a relationship with that developer and say tsk tsk don’t do that or punish that developer in some way. You still need insurance (Client-side protection). Plan for worst.
YGR: Could it be though that those third-party developers are so small (in number) it’s an acceptable risk? Versus dealing with hundreds of thousands if not millions of inputs and that risk is not necessarily a risk worth taking?
JM: Your question is touching on issue of scalability. If you do have a very large app universe, it can be difficult to find enough human reviewers with enough time.
YGR: Your analysis doesn’t deal with that scalability issue?
JM: Those on-device client-side mechanisms can be implemented regardless of the (quantity of apps).The human-driven, that aspect is harder to scale. I don’t think that is sufficient to open up a world of abject and total security horrors. We want to get to these quantitative numbers, but those numbers are oftentimes difficult to provide.
YGR: (MIT Technology Review article from a couple of months ago about Apple security informing my questions)
YGR tells JM he’s not getting back to Massachusetts this weekend but hopefully he can get to Wine Country in between cross-examine prep.
He says he will tell them “the judge sent me.”
Paresh Dave Covering Alphabet for Reuters m +1-415-565-1302 (on text, Signal, Telegram and WhatsApp) @peard33 | My articles | LinkedIn *Paresh is indeed the first name
|