More from Epic v. Google: everything we learned in Fortnite court
I don’t really buy that personally, but I did like how Bernheim admitted that switching has gotten easier in some ways. Too many witnesses feel like they’ve been stuck in the past on how Android works.
Now, we’ve moved onto smartphone pricing: he’s suggesting that Apple and Android have targeted different markets with their phones — and that they’ve continued to diverge, with the price of the average Android phone generally getting lower while the average iPhone price gets higher.
If you’re like me, you’ve been wondering for weeks — how can Epic possibly justify that Google isn’t primarily competing with the iPhone?
Bernheim says market definition is about whether good substitutes exist, ones that can keep a monopolist from charging whatever it wants for a product. Picture a Samsung television: “You could have bought an LG television, you could have bought a Panasonic, those would be good substitutes.”
Are movie theaters a good substitute, though, if your goal is to watch movies? “We have to ask ourselves, is that a close enough substitute that manufacturers are providing enough discipline for the product manufacturers? And in that example it’s a little far fetched.”
He claims that neither consumers nor developers have good substitutes for Google Play when it comes to getting apps. App distribution on iPhones, non-smartphones, or web apps don’t compete, he says.
Why might Samsung have decided to back away from seriously competing with an app store even if it didn’t get paid off?
“The economic principles that speak to this issue concern what happens when you have corporations that do many things and have may points of contact with each other — they have an interdependent relationship,” says Bernheim, pointing out how Samsung has manufactured the Pixel’s processor among other things.
“What we know is in those types of settings, companies tend to mutually forebear to some degree... even if they don’t agree not to compete, they find it in their best interests,” he says.
I expect Google will combat that during cross-examination.
Directing our attention to a slide titled “Third-Party App Stores Do Not Constrain Google’s Monopoly Power,” Bernheim shows that every competing app store together adds up to just 5 percent of “entitlements,” i.e., Android app downloads. It’s not clear what the source of the data is here.
Then, he shows us another slide titled “Samsung Galaxy Store Only Has Five of the Top 20 Apps Worldwide (2020),” which is exactly what it sounds like.
The missing top 20 apps, as of 2020: WhatsApp, Zoom, Facebook, Snapchat, Messenger, Facebook Lite, Telegram, Garena Free Fire, Subway Surfers, Among Us, SnackVideo, Hunter Assassin, Ludo King, Netflix, and My Talking Tom Friends.
(Samsung did have TikTok, Instagram, Likee, Gardenscapes, and Shareit, according to the slide.)
Epic’s expert economist (and Stanford professor) Doug Bernheim, explaining why smartphone manufacturers who’ve signed with Google are disincentivized to compete by preloading other app stores — even if they weren’t contractually obligated not to.
He says it gets worse: “If Google ends up reducing its commission rate in response to your competitive entry, you end up reducing your commission on everything else you sell.”
The “20 cents” part isn’t theoretical. Bernheim says he found Xiaomi, Oppo, Vivo, and OnePlus each get 20 percent from Google. HMD (Nokia) and TCL signed 10 percent revshare deals.
We previously saw Google had planned to offer 16 percent to key Chinese OEMs, but I guess it grew to 20 percent.
In August 2022 — according to another slide from Epic’s economist — over 80 percent of Lenovo, over 60 percent of Vivo and Oppo, and roughly 60 percent of Xiaomi phones were activated on Google’s RSA 3.0 premier tier. That’s the one that contractually keeps smartphone makers from choosing to preinstall apps that compete with ones made by Google.
Google has said fewer than 25 percent of phones go that way, and that OnePlus strategically decided not to. While those things might be true, it’s interesting to see how many new phones from the big Chinese manufacturers are coming up Google.
(Realme and OnePlus are shown as having roughly 40 percent of new phones activated on the premier tier; of 830 unnamed OEMs, only around 10 percent of new phones are activated on RSA 3.0. The chart excluded Samsung entirely.)
We’re headed to lunch: see you back here in about 30.
As of July 2021,
Google Play, 100 percent (2.88 billion phones)
Galaxy Store, 39.5 percent
Xiaomi, 10.6 percent
Huawei App Gallery, 6.3 percent (Huawei hit US government blocks)
Vivo App Store, 4.9 percent
Oppo App Market, 2.1 percent
LG SmartWorld, 1.8 percent (LG quit phones in 2021)
ONE Store, 1.6 percent
SHAREit, 0.6 percent
Epic Launcher, 0.6 percent
Amazon Appstore, 0.1 percent
APKPure, 0.002 percent
F-Droid, 0.002 percent
Tencent, 0.0005 percent (0.01 million phones)
APKMirror, 0 percent
Where by “Google Android,” he seems to mean Android with pre-loaded Google apps.
“What is correct is when Google enters these licenses, it doesn’t receive a cash payment from the OEMs... the OEMs are receiving something, and they’re doing something for Google,” says Bernheim.
“They agree to put the suite of GMS apps on their smartphones, and that generates billions of dollars of revenue for Google.”
By positioning the whole web of relationships as a simple matter of buyers (Android app users, Android developers, OEMs) and sellers (Google), Bernheim says Apple isn’t a factor, because Apple doesn’t directly participate in those particular markets — and would still be positioned to take advantage of those buyers even if other markets exist.
“Is there anything wrong with saying... that’s one big market that competes with Apple?” asks Epic’s lead attorney.
“There would be quite a bit wrong with that,” says Bernheim.
“We’re looking to see whether there are pockets of transactions where sellers can take advantage of buyers,” he says, saying it’s part of a framework antitrust economists use to do competition analysis — and there’s a general consensus you do it that way.
I expect Google’s economist will say there’s no such consensus.
We’ve touched on market definition before, and it’s the key to each side’s case.
“I’ve concluded that Google has monopoly power in the market of app distribution on Android smartphones,” says Epic’s economics expert Doug Bernheim, adding that Google “engages in anticompetitive conduct” to maintain its monopoly there.
He says he was assigned to:
1) Determine whether Google’s challenged conduct obstructs competition
2) Identify the relevant antitrust markets for purposes of evaluating Google’s market power
3) Determine whether Google possesses monopoly power
4) Determine whether Google’s challenged conduct harms consumers and developers
Now, he’s explaining what concepts as simple as “competition” and “monopoly power” mean. I’ll let you know when we move further — he says he’ll get specific.
Epic’s lawyer, making the point that all of Google’s other platforms and ecosystems products are unprofitable — save Google Play. Skinner agrees it’s the one exception.
But Google gets a chance to ask one last series of questions before Skinner leaves the stand, one that takes us back to the A1 Camera Company. Yes, Google Play is the largest source of revenue, but Android is the largest source of costs, Skinner testifies, and he believes it makes sense to keep those on separate P&Ls.
Skinner is done, and we’re being introduced to Stanford economics professor and former economics chair Doug Bernheim, Epic’s economics expert, who says he has antitrust experience as well.
Google’s accounting professor waffles a bit, starts coming up with explanations, and Judge Donato makes him answer “yes, no, or I don’t know.”
“Yes, I think some of them should be,” Skinner finally replies. But then Epic’s attorney shows he replied differently in an old deposition: “I’m not necessarily opining that these Android costs should be allocated.”
Epic gets him to admit that he has not determined which costs should be allocated. “Well, I think I’ve made clear that at least the portion of the costs of these for the ecosystem,” he begins, before getting cut off.
The jury definitely looks refreshed today, by the way. Still saw a couple of yawns, but frankly, I’m yawning, too.
Google pointed out Epic’s accountant never publishes expert research. Epic is now pointing out Google’s accounting professor has never practiced as a professional accountant.
Both point out that the other’s expert basically has a business giving testimony in court. “About 15-20 times, yes?” asks Epic of Google’s witness. Skinner says that sounds about right.
And, like Google, Epic is attacking what the expert was hired to calculate and what they did not. This time, Epic is pointing out that Google’s expert didn’t bother trying to actually calculate Google Play’s app store profit margins — only that of Google’s larger platforms and ecosystems group, which includes Android.
That’s the expert opinion of Google’s accounting professor, who says it “just makes sense if you want to track the costs of a particular activity, like R&D” in a separate profit and loss statement (P&L).
“The other reason is these costs benefit the whole ecosystem... we call this a joint cost instead of a shared cost because it benefits all the different products simultaneously,” says Skinner.
He says it’s not unusual for companies to allocate their costs this way. He believes the operating margin of Google’s platforms and ecosystems group in 2020 would be 19 percent if you account for this — not the 66 percent that Epic’s expert calculated for Google Play specifically.
Skinner and Google are giving us a theoretical example of the “A1 Camera Company.” Skinner bids us assume its P&L shows $1 million in revenue and $600K in costs for a $400K operating profit — an operating margin of 40 percent.
But “R&D costs are now tracked in a separate cost center,” Skinner explains.
You could track $200K worth of R&D costs on the original P&L, or you could split out a new P&L for an “R&D cost center.” If you did, it would look like the camera company only had $400K in costs and, thus, had a 600K operating profit for a 60 percent operating margin, while another part of the company had costs of $200K and made no profit at all.
(But is this what Google actually did? We haven’t gotten there yet — Google’s just educating the jury. In fact, Google is literally asking Skinner to tell the jury what this means.)
We just saw a Google org chart with a straight line down through the middle — from parent company Alphabet down through Google Services, to Google’s platform and ecosystems group, and finally, individual components like Google Play and Android that exist inside that group.
“This structure provides the context within which we can understand how Google Play operates,” says Skinner.
Now, Google is pointing to CEO Sundar Pichai’s simple quotes about how Play wouldn’t be possible without the investment in Android.
Skinner begins by saying that “Epic overstates the profitability of Google Play,” because it “excludes the costs of the investments Google has made over the years in Android.”
Google spent a good bit of time talking up his credibility before submitting him as a witness. He was a professor at the Roth School of Business from 1989 through 2005 before joining the University of Chicago and says he chairs an investment group that manages around $45 billion in assets.
He’s talking about how he teaches “managerial accounting,” a field which, I assume, will help Google make its case that it didn’t need to include Android costs in Play’s P&L.
Google’s lawyer is arguing that the company spent over $40 billion on Android and that Barnes didn’t account for that.
“Those costs were not allocated to Google Play based on the decisionmaking that Google determined,” says Barnes.
Google’s last point before passing the witness back to Epic: Google also generated $92 billion for app developers as of 2020 — something that also shows up on different financial statements than Google’s. Barnes supposes yes, that money would show up on the financial statements of the developers that Google supports.
Epic’s final question for Barnes was basically: “Is it safe to assume costs on one P&L can just be moved to another P&L?” Barnes says no, it would be improper.
Google’s lawyer just pointed out that Amazon spent $1.4 billion on its Appstore over seven years, and yet, Barnes found Google spent $2.4 billion in a single year.
It sounds impressive, but it’s also not quite relevant to what Barnes found. “I see what you put on the slide, but you haven’t compared it to revenue,” he said, nodding to the fact that it doesn’t speak to Google’s profit margin at all.
And now we’re in virgin territory: we’re seeing an “Alternative Profit Calculation” where Google Play would have lost money from 2011 through 2017 if it charged just a 12 percent revenue share — roughly $3 billion in aggregate before turning profits in 2018 and beyond — assuming Google still spent as much as it did and everything else stayed the same. Barnes is visibly skeptical it would.
“That’s just math, that’s not proper financial analysis,” says Barnes.
Google began by showing that Barnes is primarily an expert who consults in court and primarily looks into financial fraud — then got him to agree there’s no allegation of fraud in this case and that Google’s financial documents seem credible.
“That was the result of my inquiry, yes,” says Barnes.
Now, Google is trying to justify its high profit margins, pointing out Barnes isn’t aware if Google has higher profit margins than any other app store — and even if it did, getting him to admit that companies with strong brands and strong products could have higher profit margins than others.
Barnes is definitely on the back foot! But how relevant is this? I guess we’ll find out after Epic’s lawyer gets another chance with him.
It allegedly hit 39 percent in 2015, 43 percent in 2016, 49 percent in 2017, 59 percent in 2018, 63 percent in 2019, 66 percent in 2020, and finally, 71 percent in 2021 — generating over $12 billion in operating profit that year.
Barnes continues to say he looked through documents to try to find any missing costs Google would have incurred to operate Google Play but didn’t find them.
“Is there some missing bucket of costs that could be added to Play to reduce its margins?” asked Epic’s lawyer.
“It would surprise me if in fact that were the case,” Barnes said.
It’s Google’s turn with Barnes. Google is beginning by attacking his credibility.
Everything I just told you about Google Play’s P&Ls was pretty opaque in the courtroom — but Epic’s first expert witness of the day is here to argue they clearly show Google knows full well that Google Play is a profit center.
Ned Barnes is a CPA with Berkeley Research Center, and he says that the costs of Google Play were already accounted for in Google’s various internal financial statements.
“How did you know they were not under the direct control of Google Play management?” Epic’s lawyer asked.
“My group found they were under the direct control of Ruth Porat, CFO of Google,” Barnes replied.
We just saw Google’s platform and ecosystems planning P&L for the Google Play app store in 2020 with forecasts for 2021. It showed revenue of $14.66 billion in 2020 and gross profit of $11.44 billion that year.
Headcount for Google Play was 1,256 in 2020 and was projected to go up to 1,278 employees in 2021.
In a separate 2018 document, we saw revenue of $6.85 billion.
We also saw a financial document for the Android business, not Google Play, that showed Google did at one time at least account for its costs of building Android (including employees, infrastructure, marketing, and the “Google Play service”) separately.
Okay, now we’re getting into it — apparently Google Play’s P&L explicitly doesn’t include Google’s costs of developing Android, and Epic’s pushing Cramer to explain why. He basically says it’s because Play is part of Google’s broader Android ecosystem, not its own business — despite having its own P&L.
“If Play Store was an independent company you would incur other costs that are not reflected on the Play P&L today and don’t need to because Play is not a standalone company,” he says.
But “Android development costs don’t just benefit Google Play, right?” asks a lawyer.
“I don’t want to speculate...” Cramer begins, “I think it’s fair to say the Android investments, are, yeah, connected to broader ecosystem.”