← All-In

DOJ tries to break up Google, vaccine questions, Ukraine escalation & more

Episode 1132023-01-27EnglishMembers only
Listen to this episode00:00 / 1:38:30
Reading settings
Reading preferences
Text size
Download in-depth read
Read offline

English · In-depth read · Current version

Downloads are not available yet.

Text language

In-depth read · Key ideas, reasoning, and examples, organized by topic.

§01Cold open: a Nikki Haley riff and a moderator vote

The episode opens with banter. One speaker asks how much momentum Nikki Haley has this week, and the reply turns it into a running joke about "trying to make fetch happen", implying she is not gaining traction. The riff is comic and offers no data. It also drifts into a gag about a spreadsheet of "long Google, short Facebook" from the previous year.

A programming note follows. A speaker says a Twitter survey picked Sacks as the person the audience most wanted to moderate next, and a mock-announcer hands him the show. The host who normally moderates is teased for being annoyed, and then says he will "never ever ever" moderate this particular segment. The turn is played for laughs.

§02A scripted intro roasts the hosts

The stand-in host opens by citing the doomsday clock, which the speaker says has never been closer to midnight, and jokes that his co-hosts would rather discuss stock portfolios. Ukraine, he says, is being saved for later in the show. The remark is a setup for the roast, not an analysis.

The rest of the intro is a series of jokes about the other hosts: one is imagined being entombed with his money and sweaters, another is cast as a servant in the afterlife for flattering tech founders, and a third is described as just paranoid enough to survive a bunker. The others laugh, ask whether it cost them, and tease the reader for reading the scripted notes aloud on an otherwise unscripted show. One speaker recalls the intro was cut once and then put back.

§03Issue 1: the DOJ moves to break up Google's ad business

The first topic is a lawsuit the host says the Justice Department and eight states filed on Tuesday. As described on the show, it alleges Google abuses its role as one of the largest brokers and auctioneers of ads placed on websites and mobile apps, and it seeks to break up that business. The host says the filing promises a protracted court battle with large implications for digital advertising.

Google's response, as summarized by the host, is that the DOJ's request to unwind two acquisitions from a decade ago tries to rewrite history, that the DOJ mischaracterizes its products, and that people use them because they are effective. The company also points to competitors such as Microsoft, Amazon, Apple and TikTok. The host then asks Chamath whether the DOJ has a case, whether Google has a monopoly in online advertising, and whether breakup is the right remedy.

§04Chamath: the market share data does not look like a monopoly

Chamath calls the lawsuit "totally ill founded" and argues it reflects personal enmity toward entrepreneurs in parts of the US government. His definition of a monopoly is a stagnant market where the monopolist has pricing power and control. He cites a Bloomberg chart, which he says shows Google at 26.5 percent of the market, a diverse group of others at 43.4 percent, Meta at 18.4 percent and Amazon at 11.7 percent. These are his figures from an article he shared on the show and were not checked here.

He contrasts that spread with the 1980s case in which Mobil was broken up, saying that pie chart would have shown close to 100 percent share. He adds that Amazon was barely present years ago, now holds almost 12 percent, and could reach 15 to 20 percent in a few years on his forecast. In his view the DOJ is trying to punish American companies rather than be logical.

He says a better case would target search, which he calls a monopoly for Google, and then try to show that it created pricing power in ads. He predicts the current approach will "get deconstructed" and fail, noting that the Texas version of the lawsuit was already thrown out of federal court, a claim made on the show and not verified here.

§05The government's narrower market definition

The moderating host agrees Chamath makes a strong point, since Google's share of digital advertising overall is about a quarter. He then presents the government's counter-framing: the relevant market is brokered advertising for third-party sites and apps. On the lawsuit's chart as he describes it, Google has 90 percent or more on the sell side because of the DoubleClick acquisition, between 40 and 80 percent on the advertiser side, and over 50 percent of the ad exchange. He asks Friedberg which lens is right.

§06Friedberg: why publishers choose Google's ad network

Friedberg says he has worked with ad networks since about 2002, was close to the DoubleClick acquisition and was a business product manager on AdWords, though he is no longer directly involved. He confirms the chart's basic picture: a network connects buyers of ad inventory with sellers, and publishers keep selling through whichever network pays them the most.

He gives two reasons Google won. First, it ran its network as an auction, with advertisers competing and quality signals such as click-through and bounce-back feeding in, which he says yields the best price for publishers and a better ad for the user. Second, it offered the highest revenue share. He recalls earlier networks paying publishers around 40 percent and says Google now pays roughly 70 to 80 percent, a figure he says he would need to check.

Because the system is an auction and opt-in, he argues, a publisher can leave if the price is not the highest. Publishers once built their own sales forces, but using a network turned out to be more effective, and only a few dozen bidders are needed for strong competition. He concludes that Google's lock-in is simply that it pays publishers the most, and that breaking it up would likely leave publishers earning less. This is his own assessment.

The moderator restates the point as the auction preventing monopoly rents. Friedberg agrees and adds that Google already pays the highest share on its network, so it is unclear what higher figure, such as 90 percent, would be the "right" number. He begins to add two further points before being interrupted.

§07A counterview: right target, too late

The moderator turns to another host and asks whether he agrees with Friedberg. The reply is that monopoly behavior is going on at Google, obviously, with search and its own content and services put up top, and that search is probably the easier target. On the ad case, he says the DOJ looks about ten years behind and should have blocked the 2008 DoubleClick acquisition.

He adds that publishers would answer Friedberg by saying that selling directly to large brands brings a much higher CPM. He estimates those direct relationships cost 20 to 30 percent, about the cut Google takes. He also notes that Amazon has built a large ad business, Netflix and Disney now have ad tiers, and ChatGPT is challenging search supremacy, so the move looks late, as with Microsoft. These are his estimates and framing.

The moderator asks whether the DOJ is acting as a "rug puller" by trying to unwind ten-year-old acquisitions. A speaker says it does not make much sense to him to unwind deals from a decade ago.

Keep reading as a member.

You've read the free preview. Become a member to read the rest, with side-by-side reading and downloads where available.