Epic v. Google: everything we’re learning live in Fortnite court

We’re looking at a 2018 blog post from Epic that gives Google credit for contributing on-site engineering resources to make the game work — a blog post that Epic seems to have removed from its website, but you can access the text here.

Grant says he doesn’t know if partner support was “crucial,” as the blog post claims, “but it was very helpful,” he says.

In an old deposition, Grant was asked if Epic compensated Google for any of its services. He had a very diplomatic way of saying “not directly,” saying Fortnite fixes would lead to changes in Unreal Engine that would provide better performance for all mobile games that use the Unreal Engine.

He concluded: “I think it was a win-win as to compensation.”

(Big picture: Google is arguing it helped launch the game, only to watch Fortnite skip its store, twice, to avoid paying Google its cut of the proceeds.)

We’ve now got Epic Games senior director of engineering Andrew Grant on the stand, and Google attorney Kyle Mach asked straight-up: “You know that Project Liberty violated Epic’s contract with Google?”

With no hesitation whatsoever, Grant replied: “Yes.”

Of course, Google has him dead to rights in internal communications. In 2020, Grant wrote: “The needle we are trying to thread is to not fall foul of any agreements we have with Google, other than as needed for Project Liberty.”

(Google has counterclaims against Epic for breaching its contract.)

We’ve known for years that Epic’s Project Liberty was a hush-hush plan to draw Apple and Google into a legal antitrust battle — and Epic just admitted it built a “coalition” lobbying group to help achieve its aim.

But the company’s Project Liberty never targeted Samsung, Weissinger just admitted on the stand — after also admitting that, yes, Samsung does things that the Coalition for App Fairness rejects.

“You do not believe that Samsung abides by the principles for app fairness that are on your coalition’s website?” asked Google’s attorney.

“I agree with your statement,” he replied.

“Epic has a special deal with Samsung Galaxy, correct?” Correct.

“Epic chose not to focus Project Liberty on Samsung, correct?”

“Yes, that’s correct,” said Weissinger.

Epic’s attorney got a chance to redirect, and Weissinger started off weak, suggesting that Samsung are “exceptional promotional partners that really go above and beyond.”

Epic attorney Moskowitz tried again. “Does Samsung have the same reach and control and power as Google?”

“No it does not,” said Weissinger.

I personally broke this story in 2021, but Epic would not confirm or deny the details to me back then.

Weissinger just did that today on the stand. “Yes, it was our idea,” he said, confirming that Epic paid a consultant $100,000 to start the lobbying group, and that there were no other founding members.

In 2020, the “Coalition” told press that Epic Games, Spotify, Basecamp, Match Group, Tile, Blix, and Deezer were all “founding members” of the organization, but Google just asked if any of them were indeed founding members.

“If you’re referring to the incorporation, no they were not,” said Weissinger.

Here’s how my request for comment went in 2021, in case you’re curious:

But Epic would not confirm or deny whether it provided the original funding, hired a consultant to form the group, or bound that consultant to report directly to Epic, after repeated requests, and no CAF member we contacted would comment on that.

We’ve heard a lot about the friction Google erected to keep “unknown” sources from installing apps on Android phones and how Epic believes it’s detrimental, but Google just showed that Epic knew Fortnite’s high minimum requirements to play on Android phones and its huge download size were also barriers.

Epic estimated that 37 percent of users gave up between “Clicked Download on Web” and opening the Epic Games app — but 40 percent dropped off after opening the Epic Games app and before passing a minimum spec check.

Weissinger says 40 percent is not bigger than 37 percent “in terms of absolute players,” suggesting more players were dissuaded by the former, but had to admit that 40 is bigger than 37 “as a number.”

Update: Weissinger has now explained that the 40 percent was of the remaining players after 37 percent of those attempting to download had already been dissuaded.

Only 260M phones met Fortnite’s minimum system requirements, compared to billions for competitors, the document showed.

“Download and Patching Process remains a challenge on mobile due to Fortnite Download Size relative to competitors,” reads another line from Epic’s document, showing that Fortnite’s first download took over 14 minutes, compared to under a minute for PUBG and Call of Duty: Mobile.

That’s Epic’s VP of marketing today.

In a February 2020 internal document, Epic wrote that “mobile drives new accounts,” with 38 percent of daily new accounts in 2019 coming from a mobile platform.

But it also wrote that “mobile retention is low compared to other platforms,” with iOS retaining just 6 percent of players and Android retaining just 4 percent after the 30-day mark, compared to 16 percent on PS4, 11 percent on Xbox One, and 21 percent on Nintendo Switch.

And as we’ve already covered, players weren’t buying their V-Bucks on Android but, rather, on console. Just 0.7 percent of purchases were on Android as of 2018. Players who were introduced to Fortnite on PS4 bought 94.2 percent of their V-Bucks on PS4, but only 21.6 percent of those who started with Android bought their V-Bucks there. (34.8 percent of them bought them on PS4. Over 5 percent bought them on iOS.)

These all predate Fortnite’s launch on the Play Store — they’d have been from the sideloading days.

Google’s Paul Gennai was dismissed just before lunch, and we’re now hearing from Epic’s marketing chief — but he’s Google’s witness, which means Google is asking the questions, which means we’re not starting with pleasantries.

Google has just asked about Fortnite’s low retention of players on mobile — a point I myself brought up the other day.

In a March 26th, 2019, email from Google’s Paul Gennai to Atul Kumar (bolding mine):

In terms of the Samsung proposal, a few thoughts:

I realize that Jim noted a $50M a year number, but it’d be helpful to put that into perspective. What would that be in terms rev-share number, recognizing we might prefer to hide it in some way?

Note that I do think that a rev-share would be better at disincentivizing other app stores from being preloaded (given it’d defray revenues), which is why I liked the idea of offering a rev-share across Search and Play, but not breaking out the figures in our accounting to them.

I’m confused on the homescreen exclusivity part. If they’re distributing via Play, is the fight to have them remove Galaxy Store from the homescreen really worth fighting? I can see them digging in here and I’m not sure it’s worth us being tied to it.

Google did wind up doing the revenue sharing arrangements with RSA 3.0. Wouldn’t companies have to factor in this lost revenue share before preloading another store? asked Epic. “Yes, another app store would have to compete on a similar basis,” said Gennai.

We’re going to lunch now.

Epic just showed us an internal Google proposal document from May 23rd, 2016, where it was figuring out its “next steps with Samsung.”

Within that document, there’s a list of what “Google Gets” and what “Samsung Gets” for each of the proposed items.

One of those proposed items was “Samsung Zone in Play replaces Galaxy Store.”

The “Google Gets” field reads: “No Galaxy Store.”

The “Samsung Gets” field reads: “Stop wasting resources on Galaxy Store.”

Epic’s attorney summed up: so, there was an offer for Google to pay Samsung $50 million a year to close the Galaxy Store?

Gennai said he didn’t recall it like that. “I recall it being a separate storefront that would use Google Play infrastructure.”

That’s Google’s Paul Gennai, who pushed back reasonably well against Epic’s assertion that these were so secret as to be illicit. “I don’t think Google publishes most commercial relationships with companies,” he said.

When he pointed out that most companies in Hug deals “continue to pay the published rate,” though, Epic leapt at that.

“Yes, exactly, they just got millions and millions of dollars on the side.”

Gennai protested that those companies work with Google.

“To the tune of millions of dollars in cash and prizes,” asserted Moskowitz.

“I don’t know the numbers,” replied Gennai.

We’re looking at Gennai’s big document again, the one Judge Donato said would magically change, and I’m pleased to say I have now copied down a huge chunk in full for your reading pleasure.

Epic’s lawyer is asking a variety of questions about various passages, but the most compelling ones were these:

Was “contagion” the driver behind Project Hug?

“Yes, we did not want the loss of titles from Google Play to start a chain reaction of more and more titles on Google Play leaving.”

Was Google worried about losing them to other competing app distribution mechanisms on Android?

“They could just leave Android entirely and keep focusing on iOS.”

Would a developer like Activision Blizzard King just up and abandon Android’s billions of users?

“They can use it as a threat. We didn’t see it,” says Gennai.

January 11th, 2019, in an email from Gennai to a “Sri” who’s being tasked with a big project:

The project that Mike will talk to you about is our biggest Play priority for Q1, but one that we don’t have the resources to staff (our Play strategy team had a lot of departures into the business).

Essentially, Jamie and Sameer are asking us to consider what Play should do in the face of increasing app store competition on Android, both from OEMs (who are looking to increase services revenue / differentiate devices with content) and other large platforms (like Epic). In that context, they’re willing to consider all options, including changes to our business model, considering paying our OEMs and carrier partners, and more.

We see it as two phases: setting the scene for why it’s important (some scenarios for impact on Play, Android and Google), then onto solution design.

It’s a huge project and one where we’d definitely appreciate the help. Jamie and Sameer have already confirmed they can allocate an hour every other week to keep track of progress and make decisions.

Two impressive questions from Epic attorney Moskowitz:

“During that time the App Store was charging this lower fee and Google wasn’t, you’re not aware of a single developer that pulled out of the Play Store because of this differential pricing?”

And, about the “loophole” where some developers were able to pay 0 percent because they hadn’t adopted Google Play Billing: “Are you aware of any developer that abandoned iOS and came over to Android exclusively?”

He wasn’t aware of either. Epic’s implication is that Google didn’t truly compete with Apple on app store pricing. (Google, mind you, often argues it competes for consumer purchases of phones instead.)

Why did Google seemingly sit on its hands for well over a year after Apple dropped its App Store fee for subscriptions, before eventually matching that rate? That’s what Epic is pushing Gennai to explain on the stand, and the answers aren’t clear.

Google’s primary answer so far is that these things take time, but we’re seeing internal Google executive summaries that at least suggest Google was taking that time intentionally.

I haven’t seen any proof Google was being greedy, but it did seem worried about how developers would react: one executive summary of a meeting between top Play Store execs shows they were so concerned with the “Spotify plan” that changing the Play Store policy seemingly hinged on Spotify.

“Make go-no go pending Spotify status” was listed as a recommendation right under “Change rev share to 30(3mo/15)” and “Change policy for all (with risk mitigation measures and exceptions).” I wonder what those exceptions were.

“We can’t count it accurately,” says Gennai.

“In the order of magnitude, it’s reliable, but I can’t attest to it being correct.”

Even so, at least one Google internal document suggested that switching wasn’t common. “While the number of OS switchers is small, those that switched to iOS tend to be younger and live in urban areas,” reads one line that Epic highlighted for the jury.

“There is more switching behavior between Android OEMs than churn to iOS,” reads another.

In March 2021, a Google document shows 22 percent of Apple iOS users were listed as considering another OS, but only 9 percent were listed as intending to switch, with actual switchers labeled TBD.

For Android, it was 17 percent considering another OS, just 7 percent intending to switch, and again, TBD on actual switching. Google has already argued that small percents of billions add up to big numbers.

“If I love Gmail, and I do, and if I love Google Maps — and I really do — I can buy an Android or an iPhone.”

Gennai says that’s correct.

“You understand that Google makes money off those apps on iPhone?”

“In some instances,” he says.

(I missed what she was driving at this time, I’m afraid, just thought I’d share an interesting moment in the room.)

Epic attorney Lauren Moskowitz is asking a rapid-fire series of questions about the things Gennai does not know — because apparently, his approach to an old deposition was to say he didn’t know things over and over again, and she’s taking full advantage.

So he’s seemingly being forced to admit he doesn’t know various facts and figures and processes now — like how often Google Play is preinstalled on phones or the quantity of apps on Play versus the Amazon Appstore.

She also tried to suggest he ignored the 3.4-star rating for the Move to iOS app he demoed, choosing to focus on its 100 million downloads instead of its 180,000 reviews, and points out he skipped steps in his demo.

“You skipped messages for example?”

“I did, yes,” he admits.

Judge Donato gets another laugh (he says it’s just how the rules of evidence work in this particular instance).

Here’s what’s at the top of that document, written by Gennai:

For a variety of reasons, the long-standing growth and (potentially) fundamental business model of Google Play is under pressure in 2019. This is being driven by a confluence of factors, including:

The size and margins of the market are making it attractive for new entrants, either drafting on existing distribution (OEMs) or competing on rev-share rates (Epic)

A meme is emerging related to the “app store tax”, unfairly comparing the app store revenue share to basic payment processing and putting pressure on Play’s 30%. Larger developers who have more standalone capabilities (and other differentiated content) are leveraging this discontent to attempt to negotiate differentiated rates or go-it-alone.

These problems are very real and playing out in the market in 2019. Epic/Fortnite, Galaxy Store, conversations with large developers at GDC, etc...

Combined, these two risks could have major effects on Google Play’s business in the near term...

We propose to solve these challenges through a combination of tactics that are required in conjunction:

Firstly, competing on price (rev-share) is prone to be a race to the bottom and is not necessary across the board (only to the most strategic partners where loss of the titles would disproportionately affect the business). We are interested in solving for this in a way that accomplishes two goals simultaneously: (1) lowers the effective rev-share of Play while maintaining the prevailing rate, and (2) returning this value in a way that accrues value back to Google. This is Hug.

Secondly, the defense of Google Play as the preeminent Android app distribution store can be focused around major points of distribution, particularly in aligning Android’s largest OEMs around Play as an app distribution source. There is also a user privacy / security benefit of doing this. This is Samsung (and maybe Huawei).

Finally, the prevailing 30% rev-share rate may not be sustainable over time, and it’s in our interests to innovate in that regard before the decision is taken away from us. There are options that are likely Googley, good for the ecosystem, and allow us to continue to derive the right level of value from our investment. This is the business model project.

Why did he write the “competing on price (rev-share) is prone to be a race to the bottom” bit, asks Google’s lawyer.

“If that that were to become known, other developers would come to us to do the same, and... negotiate it down, and that would play out badly for us,” said Gennai.

Correction: I originally missed a bunch of text in the middle of the document; it has now been added.

Gennai on the “Move to iOS” app — we just watched a taped demonstration of data transfer from an Android phone to an iPhone that took less than three minutes.

Google’s point seems to be that switching between Android and iOS is not that difficult, and I largely agree, but not because of what was just shown. This demo doesn’t show lots of other switching costs beyond initial setup (including carrier-gated things like making visual voicemail work properly), and I would argue it’s harder to go from iPhone to Android than the other way around.

“Developers were working with us in partnership and they were offered a 15 percent revenue share as part of that,” says Google’s Paul Gennai, referring to Google’s Living Room Accelerator Program (LRAP) and parallel ADAP program for audio apps.

What about subscription devs who were paying 0 percent back in the day?

There was a component of our billing policy at the time that was likely confusing and could be misinterpreted; there were developers selling goods and digital subscriptions at the time that were not using Google Play Billing.

“We knew we were going to have to come around and correct that problem,” he adds.

This does not explain Spotify’s current 0 percent rate, or its special offer to Netflix, or other deals we’ve seen in court, though.

But it is a defense for another point Epic has made about how Google was not enforcing Play Store Billing evenly.

Google’s Paul Gennai seems to be here to demonstrate that Google isn’t just pocketing its Play Store service fee — it’s investing in the Google Play store because it knows it needs to compete with Apple. It’s important when consumers “choose which phone they’re going to buy,” he says.

“Why did Google choose to invest in the Play Developer Console the way you described?” asks Google’s attorney Jonathan Kravis.

“Our business model was aligned with the business model of developers, when they succeed, we succeed,” was the core of his answer.

Now, Google’s showing that Google didn’t just copy Apple App Store features — Google introduced promo codes and pre-registration for apps first, he claims, and quickly followed Apple with apps for kids and a subscription game library. The reason: “We wanted to stay competitive to Apple’s App Store.”

I’ve been wondering for many days — what will the members of the jury actually be allowed to decide? They’re not going to define the relevant product market all on their own, right?

I can now give you a slightly better idea — last night, Epic and Google submitted dueling proposed jury verdict forms (pdf). As you’ll see right away, Epic is trying to avoid defining the market, while Google is asking jurors if Epic proved that an “Android app distribution market” and an “Android in-app billing services market” exist to begin with.

It can’t be both ways: only one final judge-edited verdict form will reach the jury.

We’ve nearly spent a fortnight in Fortnite court.

Here’s how we ended the day: Judge Donato told Epic that he doesn’t believe any of Google’s conduct reaches the standard of being “per se anticompetitive.”

“I’m perfectly fine with rule of reason but I’ve having trouble saying they’re so egregious on their face,” he said.

Epic lead attorney Bornstein says he’s only seeking per se for the Project Hug agreements with Riot, Activision, and Supercell, and the judge says he’s willing to hear more but isn’t inclined to take this away from the jury. He thinks we’ll be deciding everything under the rule of reason standard instead, partly because of a Circuit Court decision (presumably referring to where Epic v. Apple went on appeal.)

Judge Donato also wanted to know where Epic thought we’d land on the relevant product market, and Bornstein said he’s handling it differently than in Apple because of Google’s very different structure of business relationships. The judge wound up saying, “I don’t think aftermarket concepts are in play here.”

I am not a lawyer or even a legal reporter, so I’m not going to translate any of this for you on the spot. It’s not worth getting it wrong.