Instagram Archive

Presented with a side-by-side screenshot of the three short-form video products on the same Saturday Night Live video, Hemphill takes a long pause to identify the apps before he’s saying he’s not sure which is which.

Meta’s lead attorney Mark Hansen is using cross-examination to attack Hemphill’s credibility. Hansen calls him the “very definition” of a witness with preconceived notions about Meta’s liability and brings up what a press report called a “roadshow” he embarked on in 2019 with former Biden administration official Tim Wu and Facebook co-founder Chris Hughes to encourage state and federal enforcers to look into Meta’s potentially anticompetitive behavior. Hansen charges that Hemphill hasn’t adequately disclosed this activity, which the Meta attorney characterizes as advocacy for a cause. Hemphill says he wasn’t advocating that they bring a case, but to probe and find out if a valid case existed.

After Hemphill describes how Meta’s increased rate of showing users ads is negative for consumers, Boasberg asks him to square that notion with his claim that WhatsApp and Instagram would have had to make money eventually, whether or not they were part of Meta. Hemphill says that even if both served ads to users as independent apps, it might have been to a lesser degree than they do now, or they might have offered users other benefits in exchange.

This is the message Boasberg says he’s received throughout trial, and he’s trying to square it with Hemphill’s assertion that users still want to see posts from their friends, even as they spend more and more time watching videos from people they’re not connected with. “Shouldn’t we assume that’s what they want, and therefore, what you’re terming underinvestment is just a shift that follows where users want to be?” Boasberg asks. Hemphill says Meta, not its users, determine the makeup of posts they see in their feeds, and even if friends content is declining as a percentage of the app, it’s still very large in absolute terms. And even if Meta makes less money from ads in Reels, a lot of content from people users don’t know now shows up in their feeds, too.

Boasberg pushes back on Hemphill’s framing of the Instagram acquisition as an effort to maintain a competitive moat. Hemphill references a 2012 email where Zuckerberg discussed the importance of keeping Instagram running after buying it to prevent another app from filling the vacuum. But Boasberg says this seems contrary to what actually played out: Meta invested in the app and it experienced massive growth. “This competitive moat seems a minor factor in its decision making,” he says. Hemphill says the moat was Meta’s intent at the time, even if the app ultimately exceeded expectations.

Boasberg brings up Meta CMO Alex Schultz’s testimony of “explosive growth” for Instagram and says that it would be “pretty hard” for Instagram to grow much larger than it already has in the US, since it already has most eligible US users on its platform. Hemphill says there’s other measures on which this whole slice of the social media market might have been better for consumers if Meta never bought Instagram, like consumer welfare and the quality of apps.

Meta paid a $6.5 billion to $10.9 billion premium on its acquisition of WhatsApp, depending on which valuation you look at, Hemphill says. This can’t be explained away with standard business reasons for paying more for an acquisition, like synergies that would be particularly beneficial, because he says there’s no evidence Meta analyzed this. Instead, he says, the premium “reinforces the conclusion that the project was an anticompetitive project.”

And the most likely way it would have done this without Meta would be building its own personal social networking service that competed with Facebook and Instagram, Hemphill says, since this is how messaging apps in other countries monetized. Boasberg brought up testimony we’ve heard that WhatsApp’s founders were resistant to this kind of pivot, but Hemphill says that they still had investors and employees to answer to who would eventually want a payout. Plus, he says, Meta eventually was able to convince them to monetize the app somehow.

Hemphill compares Meta’s claims that it’s power is constrained by newer apps like TikTok to Microsoft’s claims of being constrained by the PalmPilot and AOL at the time it was found to be an illegal monopolist. He says Microsoft sustained its dominance even in the face of competition “more or less on the margins.” He also says that what Meta calls “headwinds” from TikTok didn’t stop the company from beating its own revenue projection in 2024, which already forecast 78 percent growth from 2020.

Meta has argued that users’ reports about how much they think the company cares about them is mostly a reflection of media reporting, rather than actual changes to Meta’s products. Boasberg asks Hemphill whether changes in user sentiment really are about the actual products or perceptions of them, and whether that matters. Hemphill says that even in the case of Cambridge Analytica, where reporting came out long after the privacy incident at the heart of the story, “it’s not that the Facebook business in fact changed with the revelation, but what did happen was a sort of scales falling from your eyes effect, where users became aware of this possibility.” He adds that Meta pays attention to user sentiment “in a way that is not just managing their reputation.”

Meta shows fewer ads to what they call “needy users” whose engagement drops more significantly when they’re shown more ads. Hemphill calls this a “discount” for those users — and a sign of price discrimination by an illegal monopolist.

Creating a social graph of users’ friends, likes, and interests isn’t as meaningful for social apps as it used to be, Schultz says. That’s thanks to advances in AI, which has made it easier to predict and serve content to users they might not have even known they’d like. The FTC says Meta’s network effects and extensive mapping of users’ connections makes it difficult for upstarts to dislodge it.

Had Meta never bought the app, Schultz says Instagram would have continued to copy Twitter. “I think Twitter had a less effective growth team than us,” he says.

After a brief dip in Instagram’s growth rate in 2018 due to changes in how Facebook promoted it, the app’s growth rate recovered and continued to go up and to the right, a chart shown in court shows. Schultz says the results mean the change wasn’t as dire as Systrom might have believed.

For all the tension that came from Facebook’s decision to promote Instagram less from its app, it wasn’t all that consequential for Instagram’s growth in the end, Schultz testifies. Instagram’s growth rate dipped 14 percent, but it was still growing users overall. Ultimately, Schultz says, it was a “blip.”