Survey Says: Facebook Is for Friends and Family

The FTC's survey expert, Michal Malkiewicz, backs up the government's case with data. But first, current WhatsApp head Will Cathcart, and the start and end of two more exams on Day 15.

Purple Reign

Day 15 of the FTC’s trial to break up Meta got underway with additional testimony from of the purple-haired FTC expert on “integrity” we saw on Day 14, Damon McCoy. This blog described “integrity” as related to “safety.” But the FTC’s first substantive question of the day was, “[A]re safety and security the same as integrity?” McCoy answered that they are not. “There’s overlap, and then there’s areas that are distinct between those concepts.” That led to a more specific point that, whatever Meta invested in integrity or safety, there wasn’t a way to “tell how much was allocated to Instagram.” This was a response to Meta’s expert, V.S. Subrahmanian, who we’ll see later in trial and will probably opine on Meta’s investments in Instagram.

The FTC also tried to rehabilitate McCoy on questions that Meta raised about his methodology during cross-examination, with a callback to Instagram co-founder Kevin Systrom’s statement that Meta’s integrity systems were more sophisticated than Instagram’s. That didn’t change McCoy’s opinion because he “used very rigorous qualitative methodologies to analyze the data,” like what he would “use in peer-reviewed publications.” It was an answer that sounded tailor-made to survive a “Daubert challenge” to the admissibility of expert testimony. A Daubert challenge asks, among other things, if the methods the expert is using to reach his opinions are the same methods the expert would use in the classroom or field. It’s not clear though that McCoy’s approach, relying on his “expertise, experience, and the documents provided,” as he testified yesterday, meets that standard by providing something beyond a weighing of the other evidence, which Chief Judge Boasberg can do on his own.

Chatty Cathcart

The FTC’s second witness of Day 15 was Will Cathcart, the current head of WhatsApp at Meta. He started as a product manager on the News Feed team in 2011, becoming vice president for product management for Facebook and then vice president of the Facebook app, before taking his present gig leading WhatsApp in 2019.

Will Cathcart. Credit: Meta.

The first section of the FTC’s examination covered the competitive threat from Google+ as a rival social network to Facebook in the early 2010s:

“Q You understood that Facebook viewed the launch of Google+ as a threat, correct?

A Yeah. I mean, again, I was new, but you know, it was a new option from a very large company with a lot of resources. I certainly think we took it seriously.

[. . .]

Q By taking it very seriously, you saw it as a threat?

A Sure.”

There were several exhibits on this topic, but that exchange was perhaps the best for the FTC’s case.

Why focus on Google? We’ve seen and heard throughout trial that Google tried to acquire WhatsApp in the past, and that Facebook was worried Google might make a move in the future. The combination of WhatsApp and Google+, which was Google’s own social networking service, could overtake Facebook.

The FTC’s theory is coming into view: While one way to show the competitive threat from WhatsApp is to focus on its potential as a standalone chat app that could expand into a social network in the style of Kakao or LINE, another, easier case to prove is that WhatsApp would have become a social network when joined with Google+. It’s that latter case that was on display with Cathcart.

Another email from Cathcart neatly summed up the FTC’s view that Facebook and Instagram’s core use cases are friends and family sharing, part of its attempt to define a personal social networking services (“PSN”) market:

“Hey, everyone. We just spent the week going over drafts of our H2 plans across the FB app with Mark. I wanted to share some of the high-level discussion and feedback we received on the plans and what we were doing over the next few weeks in response to that feedback. . . . But I believe that we've got meaningfully less effective at two critical use cases for people on Facebook:

• I have a few minutes to spare, and I would love to use that time to check on what my friends and family are up to and talking about.

• I want to share what I’m doing, thinking, or interested in with the people I care about.

[. . . .] I believe we are getting less and less effective at helping people share what they want about their life. And as this happens, we are becoming less relevant as a place for people to see what's going on with the people they care about most. . . .

When people check Facebook with a few minutes they have to spare, they only have a brief window of time. The median session is barely over a minute. And if Facebook is less reliably the place to see what your friends are sharing, it's a less attractive option, way, for people to spend those brief moments. Additionally, we have created a huge tax for ourselves with the slow performance and b[u]gginess of our app.”

Another email had similar language (emphasis added):

“My sense of where we are is that our product strategy is to build more and more valuable services into the Facebook app that are built on top of the unique connections people have on Facebook (it’s the one place where people connect with everyone they know in their life).”

But this time, shown the above email, Cathcart tried adding new caveats. Here are the second and third attempts to get him to agree to what he wrote (emphasis added):

Q. You thought that the marketing strategy would be effective because Facebook was the only place where people connect with everyone they know in real life, right?

A. Well . . . you only market because you are trying to convince people to try you over alternatives. When you market, you try to tell a story that's powerful. You try to differentiate -- you know, BMW tries to convince you that they're the ultimate driving machine, even though there are others cars you can buy. I was trying to lean into something broader, but they would resonate with people as something they liked. They like connecting with people they know in their life.

Q. What you wrote here was, “it's the one place where people connect with everyone they know in their life”--

A. I thought that was an effective tag line or could have been.

The ducking and diving was much the same in a subsequent line of questioning about an email Cathcart wrote in 2018 saying that the company had “multiple apps” pointed “at the same main use case (friends and family sharing)”. Asked about it, Cathcart called friends and family sharing “a” main use case, not the main use case. After some valiant attempts to get the question answered, the court sustained an objection, and the FTC moved on.

Finally, the conversation turned to WhatsApp. Over 90% of messages on WhatsApp are “sent” by 1 user to 1 user, Cathcart agreed. This is unlike the “broadcast” sharing of 1 to many that we saw FTC expert Cliff Lampe describe in support of the PSN market, and it helps put WhatsApp and (more importantly for the FTC’s case) iMessage outside that market. But Cathcart clarified that, from “the perspective of a sender, about 10 percent of what’s sent goes to an individual. From the perspective of the receiver, it’s a little over two-thirds [that] are received from groups.” That’s “because when you send to multiple people, more people receive it.”

The FTC’s direct exam then moved into a confidential portion. When trial resumed publicly, Leslie Pope of Kellogg Hansen was at the stand for Meta, continuing her cross examination of Cathcart that had started in the sealed session, which sounded more like a direct. Cathcart explained that he decided to take the job as head of WhatsApp because he was disagreeing with his wife on the pronunciation of their daughter’s name. He “realized all the discussions about naming her we typed with our thumbs because we were messaging each other,” so it dawned on him that messaging “was where the world was going, and . . . it would be exciting to be a part of that.” It was kind of a weird spin on apparently not speaking with your spouse in person enough about naming your child.

The broader point Meta was trying to establish was that friends and family sharing has moved to messaging, a trend that the pandemic accelerated. This helps Meta show that the market should include messaging apps like iMessage.

Pope also walked Cathcart through demonstratives showing how to use WhatsApp. Standard group chats can have up to 1,024 participants, which sounds more like broadcasting than something intimate. Indeed, WhatsApp has added a feature called an “announcement group,” which is a 1 to many broadcasting function; think schools sending out snow day notices to parents that don’t want to be inundated with replies. As Chief Judge Boasberg mentioned, “it sounds like a thousand reply-alls” could be “incredibly intrusive” and “inefficient.” And now there are “Channels,” too: a pure broadcast feature that “can have millions of followers.”

Cathcart gave the DC-appropriate example of World Central Kitchen, the aid group founded by celebrity chef José Andrés, as an organization that uses WhatsApp group chats “when they go into a disaster zone or a crisis zone.” And Cathcart shouted out the lawyers, who “use it to send documents around courts informally.” I can’t say I’ve ever sent a filing to a court on WhatsApp. But it’s certainly true that in the WhatsApp group chat for my law school study group, which continues a decade later, we’ll talk about current legal events by sharing the PDF or a screenshot of a filing or order.

There was also a look at “Communities,” a feature that I didn’t even know existed on WhatsApp, but appears similar to a Discord channel in that it can serve as a topical compendium of chat threads. Cathcart said WhatsApp introduced this feature in response to the “competitive dynamic,” citing Telegram’s larger groups and channel features.

Cathcart testified that he considers Telegram, Signal, Snapchat, WeChat, LINE, Kakao, and Viber as competitors to WhatsApp, plus the messaging function on TikTok. Asked about competitors in the U.S., he narrowed the list to iMessage, Google Messages, Snapchat, Signal, Telegram, and TikTok messaging. Messaging apps have a social graph; they use your contact list, but also have group chats and channels that include people whose phone numbers you don’t have. Driving the point home, Pope asked whether Cathcart agreed with what sounded like a statement from Lampe that “WhatsApp and other messaging apps are not used to maintain low-cost, distant relationships with people”: Cathcart disagreed, citing the number of random Happy New Year’s messages he gets on WhatsApp.

The cross ended with an important point for Meta’s defense about switching behavior: Looking at data from Peru, Cathcart explained that WhatsApp users shifted time to TikTok. And in the case of TikTok being down or banned in certain countries, usage shifted to WhatsApp. In Meta’s view, this makes those apps competitors.

On re-direct, Cathcart was confronted with a statement in an exhibit that “public broadcast and private messaging create major flywheels”. He agreed that they could be mutually reinforcing: “Beyonce may not read it, but someone I know might.”

Near the end of the exam, Chief Judge Boasberg followed up on a point that was discussed earlier by the FTC’s financial expert, Kevin Hearle: how is advertising revenue allocated to WhatsApp? The issue is that the “click to chat” feature used by small businesses might not be incremental ad spend but simply ad dollars that might have been spent placing ads elsewhere in the Meta ecosystem. Cathcart didn’t know and said that finding out wasn’t that important for Meta.

Being Michal Malkiewicz

The next witness on Day 15 was Michal Malkiewicz, the FTC’s survey expert. His name is pronounced like this:

Malkiewicz has had an interesting career. He served as an “economic research assistant” to Judge Richard Posner of the U.S. Court of Appeals for the Seventh Circuit, a towering figure in “law and economics” as well as antitrust. Malkiewicz has an undergrad degree in economics and environmental studies, plus an MBA and a master’s degree in economics, but he also took some graduate coursework on survey design and worked for a time at the National Opinion Research Center, which is considered the gold standard for survey design. He’s an adjunct professor at the University of Southern California, but sounds like he’s primarily a professional expert witness, now at Charles River Associates.

In this case, Malkiewicz conducted a survey to find out what users think of Meta’s apps and potential competitors. The headline is that “when asked about their most important reason for using these services, to keep up with their friends’ and families’ lives in one place was selected more than any other reason for Facebook, Instagram, and Snapchat and not for any other services.” That wasn’t the most common choice for responses about X, YouTube, and TikTok. Malkiewicz also reviewed “a universe of surveys available through discovery” and public research, and found that the results of those surveys were consistent with his.

Malkiewicz’s survey appeared to be drawn from a preexisting pool of potential respondents. It’s a lot less labor intensive to ask questions to a group that has already been vetted as representative of the U.S. population than to try to build a statistically representative pool to draw from in the first instance. Respondents are asked some top level questions to clarify if they qualify for the survey. For example, Malkiewicz excluded respondents under the age of 18—something that Meta made an issue of on cross given that many users of social media, especially on TikTok, are minors. In all, there were around 3,500 respondents—or 3,228, as Meta specified on cross.

Malkiewicz found that around 60% of respondents said the “most important reason why” they use Facebook is to “keep up with my friends’ and family’s lives in one place,” vs. single digit percentages for entertainment and news. Fewer respondents said the same about Instagram (36.7% called friends and family most important), but that was still the leading response. For Snapchat (34.1%), friends and family statistically tied “communicating privately” as the most important use. No more than 5% of respondents answered “keep up with my friends’ and family’s lives” for YouTube, TikTok, LinkedIn, Pinterest, Reddit, or Nextdoor. More respondents gave that answer for messaging apps like iMessage and WhatsApp—around 25%—but private communication eclipsed that answer as most popular.

Former college debate star Ana Nikolic Paul took the reins on cross for Meta by asking the inverse: isn’t it true that the majority of Instagram users didn’t report that friends and family connection was the most important use-case for them? Yes, because 36.7% is less than 50%, but it was still the plurality response. 26.1% of respondents selected entertainment as the most important use of Instagram.

The next part was more effective. Recall that the FTC’s PSN market includes Instagram and Snapchat but not other messaging apps like WhatsApp or iMessage. Paul showed that keeping up with friends and family ranked higher as a response for apps not in the proposed market like WhatsApp and Facebook Messenger than MeWe, another app in the FTC’s proposed market:

• Facebook Messenger: 25.7%

• WhatsApp: 25.2%

• Text messaging/SMS: 25.0%

• MeWe: 25.9%

• Instagram: 36.7%

• Snapchat 34.1%

Paul also showed a YouGov survey from 2020 that Malkiewicz didn’t cite, which found that in response to the question, “What do you typically use TikTok for?”, 51% answered “to watch videos posted by my friends” and 44% answered “to have fun with friends by creating a funny video using TikTok”—about 10 times more than responded in a similar way in Malkiewicz’s survey. Malkiewicz countered that the comparison was apples and oranges: the 2020 survey didn’t ask respondents to select the single most important use-case as was apparent from the percentages of the different responses, which totaled to more than 100%.

The cross also covered some potential methodological problems with the survey design. As a gut check, Paul showed a demonstrative comparing what people say is their favorite Taylor Swift song (Shake It Off) vs. the Taylor Swift song that’s streamed the most (Cruel Summer). An interesting FYI on a demonstrative, but not really an admission admissible as evidence.

Some other purported problems: Malkiewicz didn’t do a pre-test of his survey. A pre-test is when you ask some questions to a smaller pool just to see how people understand the questions and answers. From that process, you might uncover problems with how you’ve written the survey. Malkiewicz said he didn’t think there was a need to do this given the abundance of existing surveys on this topic. Paul’s cross highlighted one potential problem that might have become clear from pre-testing: the “other” response asking if individuals used “an internet service other than” the dozen or so listed as answers. Users could type in their own answer, and as some pointed out, the question was potentially limitless and it would be impossible to recount all internet services used in the past 12 months.

The sample itself drew scrutiny, as it was a “non-probability opt-in sample,” rather than a random sample. This led to some composition problems, in addition to the age-related limitation. For instance, 61.6% of respondents in Malkiewicz’s survey had a bachelor’s degree vs. 35.1% of the adult U.S. population according to the census. But he pushed back, finding differences not meaningful and that the “disposition gives large enough sub-samples for each of those categories” such that another survey expert could re-weigh the results based on the composition. “I saw no reasons to do that and no evidence that the results would actually change,” he said.

Due Up

Meta’s Alex Schultz took the stand last. He’s been sitting at Meta’s counsel table for most of trial as their designated corporate representative. Typically, fact witnesses aren’t allowed to see or hear the testimony of other fact witnesses; they must be “sequestered” from trial until they’ve testified. The corporate representative concept creates an exception so that there can be at least one non-lawyer executive at the company who can hear and see the evidence.

We’ll cover Schultz’s highly anticipated testimony when it concludes.