Antitrust

The Justice Department is investigating whether Google crafted its agreement to skirt regulatory scrutiny, Bloomberg reports. The deal brought Character.AI’s co-founders back to Google and didn’t technically involve an exchange of shares, though investors were set to receive a payout, The Verge previously reported. Google spokesperson Peter Schottenfels told Bloomberg that Google is “always happy to answer any questions from regulators,” and added that Character.AI remains separate, with no ownership stake by Google.

After about four days of witness testimony in its case-in-chief, and on the sixth week of trial, Meta concludes its defense. The FTC plans to put on its rebuttal case on Tuesday, where it plans to call back its economic expert Scott Hemphill to respond to Meta’s experts.

Even though Carlton says Instagram has been such a “grand slam home run” that it’s hard to imagine it doing even better without Meta, the FTC points to past markets where output increased even while plagued by a monopolist. It pulls up a slide used by the Justice Department in its antitrust case against Google, showing that global PC shipments increased while Microsoft dominated, with similar trends for long distance calls during AT&T’s monopoly, and global crude oil production at the height of Standard Oil’s power. Carlton says that anything is possible and “the moon could fall out of the sky tomorrow, but if you’re asking me if its likely that output would be higher in the but-for world, the answer would be no.”

If market players knew what Instagram would come to be worth in the years after Meta’s acquisition, the deal could have fetched a price at least 40 times more than the $1 billion it got at the time, Carlton testifies. “I just see no basis for making this assumption that Instagram in the but-for world would have been even better,” he says.

Carlton performed a regression analysis to show that the FTC’s theory that Meta charges users interested in friend content a higher price through ads does not bear out. There’s “no systematic positive relationship between ad load and number of friends,” he concludes. In fact, he found, younger users who spend a larger percentage of time on Facebook ad Instagram on friends and family sharing receive fewer ads than older users who spend a smaller portion of their time on such content.

In the past two years, Carlton says the amount of time US users spend on “friend” content on Instagram has declined from 11 percent to 7 percent. As of January, US Instagram users spent 51 percent of their time on the app on Reels. When Instagram has experimented with reducing users’ access to Reels, he says, time spent on the app goes down. Those who had full access to the short-form videos overall spent less time on Feed and Stories, which the FTC has said is associated with friend content that doesn’t compete with TikTok and YouTube.

Competition from TikTok and YouTube is really what keeps it at bay, Carlton testifies. He points to a chart Meta showed in its opening arguments demonstrating that a greater portion of Meta users went to those two apps than Snapchat during its 2021 outage. Ignoring this reality, Carlton says, “that’s just missing the boat. It’s just understating the importance of these very important influences on Meta’s behavior.”

Carlton is baffled how FTC expert Hemphill could say users would have been better off had Meta never bought Instagram in WhatsApp. Instagram, he says, “exceeded the expectations that people had at the time of acquisition, so to say that it would have been even better — how can you say that?” With WhatsApp, he adds, Meta actually lowered the price users had to pay for the service, from a dollar a year in some cases, to zero.

University of Chicago professor Dennis Carlton is up next. Carlton’s testimony will attempt to take down the analysis by the FTC’s expert witness Scott Hemphill by pulling apart his theory about the relevant market and competitors to Meta’s alleged dominance. Carlton calls Hemphill’s claims “confusing” and says he generally prefers to see “what people actually do than what people hypothesize they could do” when it comes to consumer behavior.

Boasberg ruled from the bench Tuesday denying Meta’s motion asking him to rule against the FTC even before it launched its defense case. It was something of a long-shot since the trial is already so close to the end, and Boasberg had already said he wouldn’t pause the case while weighing the motion. But the filing outlined some of the key reasons Meta thinks it should win the case, and we’re hearing these in more detail as it continues with its case-in-chief.

The FTC demonstrates that Instagram hit 200 million monthly active users, and launched its video and messaging features all before migrating from AWS to Meta’s internal infrastructure. Meta has used its infrastructure improvements for Instagram to defend against the FTC’s claims that its acquisition was anticompetitive. Shortway also concedes on cross-examination that his start date at Instagram was technically a few days after the acquisition closed, and a lot of the early issues he mentioned were ones he heard from speaking with co-founder Mike Krieger during the hiring process.

Prior to the acquisition, Shortway testifies that infrastructure was one of the “major causes of our pain” and the AWS hardware Instagram relied on “was not able to handle the workloads that we were throwing at them.” The small startup also didn’t have anyone on staff with expertise in fighting spam or conducting content moderation. The acquisition changed some of that within days, he says.

Meta production engineering director Nicholas Shortway, who worked as an infrastructure engineer at Instagram prior to the acquisition, is now testifying about what it was like keeping the app running in the early days. “Everything was on fire,” he says, and since so few people on the team knew how to get the app back up and running during outages or other crises, Shortway recalls carrying a computer with him wherever he went, and fixing Instagram from grocery stores and movie theatres.

Those would have been worth $4 billion today, FTC attorney Noel Miller says. “Please don’t say that,” Acton responds, laughing. “After taxes, it’s two.” Acton departed in 2017 before his stocks vested, shortly after a disagreement with then-Facebook COO Sheryl Sandberg, who wanted to put advertising on WhatsApp.

Acton charges that despite his opposition, Meta ultimately followed through with a business version of WhatsApp that he says diluted its end-to-end encryption. He emphasizes that the launch happened after he’d left the company.

That’s the message a WhatsApp employee relayed in November 2014, after it became a part of Facebook. “Overall, Facebook would first like to push WhatsApp in countries where its obvious that WhatsApp is or can be the leader,” the employee wrote to WhatsApp’s founders. “There was some reservation in countries where Messenger is the leader (and WhatsApp has less reach/engagement but where lies greater opportunity).”

The FTC shows a series of emails between Acton and and a potential job candidate in the two months before WhatsApp’s deal with Facebook was announced. The candidate, whose name was redacted in the public document, identified himself as a senior member of Facebook’s finance team and reached out to pitch himself as a chief financial officer for the startup. Acton agreed to take the meeting and after coordinated meetings with other team members and asked for references. The exchange suggests WhatsApp had plans to build a sustainable business absent a deal. But before any hiring took place, Facebook bought WhatsApp, and it no longer needed its own financial leader.

On cross examination, Acton testifies that the startup added group messaging, video sharing, profile pictures, and many other features prior to the 2014 acquisition. While he and co-founder Jan Koum vehemently resisted making WhatsApp into a social media platform, the FTC seems to be suggesting that it was far from a static product limited solely to text messaging, absent Meta’s ownership.

Acton testifies that the startup had already broken even on the roughly $8 million it raised and had only begun charging its $1 per year subscription in seven countries, meaning there was plenty of room to grow. As its pricing power grew, he theorizes, WhatsApp could have eventually bumped the price to $5 or $10 per year. While the FTC has theorized that WhatsApp likely would have had to run ads eventually to satisfy investors — with or without Meta — Meta says the founders had no interest or need to shift to an ad model without the acquisition.

Acton says the messaging app company had talked with Google about a potential acquisition back in 2011, but the startup decided against it. At one point, Google founder Larry Page had asked to be given the opportunity to bid on the app if they were eventually open to selling, Acton testifies. But WhatsApp never went back to Google to solicit an offer once Meta agreed to buy it for about $19 billion in 2014, he says, adding that Google was not as good of a fit at the time.

Acton reinforces a notion we’ve heard about the WhatsApp founders throughout trial: that they had absolutely no interest in building an ad-supported product or a social media feed. He testifies he wanted to focus on building the communications product and limiting distractions. “We had no ambition to build Facebook-like functionality like a feed,” he testifies. That’s why he wrote this infamous note in the early days of WhatsApp as a reminder of its guiding principles.

Before Meta bought it, WhatsApp was moving toward a model where users would pay $1 per year for the messaging service after a free first year. Though the company was cashflow positive in early 2014, WhatsApp didn’t spend money on marketing in the US, since the wide availability of cheap or free messaging there at the time would have made it a “waste of money,” Acton says.

That’s the message from Meta’s next witness, Catherine Tucker, an MIT management and marketing professor with expertise in digital monetization strategies. Tucker argues it’s important to account for this kind of competition because it’s the crux of how Meta makes money.

Government attorney Mitchell London points out that List’s “natural experiment” about how consumers behaved when India banned TikTok leaves out some potentially important context. He notes that 58 other apps, including popular messaging and social app WeChat, were banned alongside TikTok — meaning there were fewer alternative apps for consumers to turn to. List says TikTok made up the vast majority of usage prior to the app bans. But London also points to a Meta document that says, “In emerging markets, especially India, Facebook often serves a completely different purpose.”

On cross examination, the FTC interrogates List’s finding that Google Chrome is the top app users go to when they’re incentivized to spend less time on Facebook. List says he didn’t focus on the app as much as YouTube and TikTok since “it’s a little like off-device time has an infinite number of things you can do.” The FTC seems to be suggesting that this is exactly the point — and why evaluating Facebook’s relevant competitors based on where users shift their time is an imperfect measure.

Had the two never merged, List testifies that each company would have likely given in more to advertiser demand for more ads, not less, as the FTC has claimed. He says that over the long run, advertiser-side incentives would win out over incentivizing user engagement if the two remained separate, since one platform couldn’t recoup lower revenue from the other. He claims this is “the direct opposite result” from what the FTC’s expert found.

List uses the “natural experiment” that booted 200 million TikTok users from the app to argue that consumers see it as a fitting substitute for Meta’s apps. Within about two weeks of the ban, he says, Facebook and Instagram saw 20 percent increases in time spent on their apps. This means, according to his analysis, that TikTok should be considered a relevant competitor to Meta.

In List’s experiment, Instagram users who were incentivized to reduce their usage of the app diverted their time to YouTube more than other apps he tracked. The video app saw an 18.9 percent diversion rate, while Snapchat, which the FTC says Instagram directly competes with, sees a 2.2 percent diversion rate. Facebook users incentivized to lower their usage diverted the greatest share of their time to Google Chrome at a rate of 9.3 percent.

List makes this argument using a pricing experiment he ran where he paid a treatment group $4 for each hour they reduced their usage of Facebook and Instagram. The FTC has argued these apps have unique features that users greatly value to help them connect with friends and family. But List found that participants decreased their usage of Facebook and Instagram’s friends and family features about as much as all its other features – showing users don’t particularly value those features more than other ones the apps have to offer.

University of Chicago professor John List, who previously served as chief economist at Uber and Lyft, is now on the stand to try to dismantle the FTC’s market definition of personal social networking services, and dispute its claim that “friends and family sharing” is a core use case for Facebook and Instagram. He tells the court that MeWe, a small competitor in the market Meta allegedly monopolizes, is “economically inconsequential.”

Alison testifies that he’s concerned about short-form video content being “commoditized,” since creators can post across many different apps. Still, he sees building Reels as a huge engineering undertaking and investment that was existential for the future of the business.“If we didn’t invest in Reels, then long term, our entire business was probably going to go down significantly,” he says. “We were really believing that this was going to be the future of our business.”

As the app grew early in the pandemic, Alison says, “we were very surprised by how much time people were spending on TikTok.” Meta found that number was roughly 120 minutes per user, per day.

When Matheson points out that Facebook still prompts users to log in or sign up to connect with friends, family, and people they know, Alison cautions that “just because something is on our website doesn’t mean that it’s completely up to date,” since they’re working to update how they describe the brand. He adds during cross-examination that “people are coming to Facebook for several other things besides friends” and it’s in the middle of an evolutions of how to describe the app.

Matheson describes at a high level a publicly-redacted experiment from a 2021 presentation. Facebook found the experiment increased the amount of original content users shared to their feeds, but time spent on the app went down. Matheson says this would be bad for Facebook because if users see all their friends’ posts and don’t come back, it can’t serve them as many ads. Alison says that the number of times a day users opened the app also declined, as did its “meaningful social interaction” score — so another interpretation is that users would miss content they care even more about, like that from a support group.

That’s how Boasberg interprets the new feature that consolidates friends’ posts into one feed. “This arguably could also enable you on the feed to diminish further the number of friend posts because to the extent people say, ‘hey, I want to see more friend posts,’ the answer is, ‘just go to the tab, it’s all there,” Boasberg says. “And then if in fact the demand is really for unconnected content, then this sort of lets you have your cake and eat it too.” Alison says the decline in friend content is mostly due to users posting less to their feed, not Facebook’s own decisions.

Alison describes it as “a nice feature” for Facebook users who want it but no longer the “main character.”

That’s why Facebook created the friends tab to consolidate posts from users’ connections in one place. As Facebook now recommends more posts and videos from content creators and other accounts users aren’t connected with, Alison says that if a user really wanted to see every post from their friends in the regular feed, “you might have to scroll through 10,000 posts,” even if there’s only 10-20 new friend posts to see.

Facebook recently introduced a friends tab to try to recreate the feeling of scrolling your News Feed circa 2006. Alison says they think “there are some features from the very early days of Facebook and social networking that could become more interesting again” as apps including Facebook move toward public, algorithmically-recommended content. He describes it as an “experiment or a bet that we’re making to almost bring a little bit of nostalgia to Facebook as the core experience goes away from friends.”