What this covers
Robert Cantwell and Jesse Pujji, who buys ads on Facebook and Instagram, examine the ad business from both sides of the auction. Robert gives the company-level numbers: about 10 million active advertisers spending roughly $11,000 a year each, and capex of about $21 billion, third in the S&P 500. He also describes a slowing flagship US app offset by fast Instagram growth. Jesse gives the operator's view.
The mechanics get most of the space. Jesse traces Facebook's move from a Google-style second-price auction to a VCG auction around 2013, and explains why ads in the newsfeed beat right-rail placements. Advertisers, he says, start from a budget and an acquisition target rather than a CPM, and he reads rising CPMs as a sign that ads are reaching the right people. Robert covers lookalike audiences, which acquire customers at about $60 to $70 each. Jesse explains pixel-built behavioral segments and insists Facebook sells aggregate segments, not individual data. Everlane serves as a case: about a quarter of its acquisition spend went to Facebook.
The contested material comes last. Jesse argues that Apple's IDFA change has made cross-app attribution harder. Robert argues Facebook bought Instagram and WhatsApp because Apple and Google became mobile gatekeepers, treats payments as the next step for a super app, and doubts that decentralization has weakened aggregators. On risk they differ. Robert's chief worry is the next computing platform, and he expects government rules on social advertising. Jesse thinks regulation unlikely and counts user flight as the more serious threat.
Facebook's advertising business is a targeted-media network whose advantage comes from a self-reinforcing data flywheel and scale, and its real risks are platform shifts and regulation rather than advertiser defection.
- Scale in users and advertisers feeds targeting data, which raises ad effectiveness, pulls in more ad budget, and rebuilds the data advantage
- Advertisers keep spending because Facebook's targeting and auction mechanics convert budget into customers more efficiently than alternatives
- The most serious threats are the next computing platform after mobile and regulation of targeted advertising, not user or advertiser flight
Facebook's network flywheel and self-serve targeting make it a superior ad platform.
- Before the web, advertisers could only target by show or channel and time of day, so a six-year-old saw the same beer ad as a 45-year-old watching the Super Bowl, and targeted advertising effectively began with the web.
“I'd argue that targeted advertising never existed until the web came along.”
- A social network is designed to be a targeted advertising platform because it knows who people interact with and what they like, whereas a paid media channel like YouTube only knows what a viewer watched, so its targeting is cruder.
“A social network, by definition, is designed to be a targeted advertising platform, as opposed to a paid media channel”
- Facebook's self-serve ad platform is so frictionless and works for any goal that it has invested more in it than any competitor, and that investment helped it reach a position where its capex is the third largest in the S&P 500.
“They've invested more in their self-serve ad platform than any of their competitors.”
- Facebook's core business is a media network that aggregates viewer attention and sells that time to advertisers, which is a traditional media model, but digital tools make the targeting far more powerful than broadcast television.
“The business of Facebook is a media network. It aggregates viewers and attention, and it sells that time to advertisers.”
- Facebook's advantage is a flywheel: the bigger the network, the more data it has, the better targeting it can offer, the more effective its ads, the more ad budget it wins, and the more data it gets back into the system.
“the bigger you are, the more data you have. The more data you have, the better targeting you can offer.”
- Facebook's lookalike audiences let a brand upload an existing customer list, have Facebook find similar users who have not encountered the brand, and acquire customers at about $60 to $70 each, comparable in quality to existing customers.
“Over that whole period, we paid about 60 to $70 per new customer. These customers were just as good as former customers that we had.”
Facebook's large base of users and advertisers sustains steady ad revenue growth across regions and apps.
- Facebook has about 10 million businesses actively advertising, each spending on average about $11,000 a year, which implies roughly $110 billion in revenue for the year, with the advertiser count growing by about a million per year.
“Ten million advertisers spending $11,000 each, and Facebook will do about $110 billion in revenue this year.”
- Facebook holds more than 80% share of social network ad spend, LinkedIn and Twitter together hold about 10% and are shrinking, and Pinterest and Snap hold about 5% and are growing slightly.
“Facebook has more than 80% share. There's a second tier of players. You've got LinkedIn and Twitter. They've got about 10% share”
- Instagram generates a little more than 20% of Facebook's advertising revenue, and a little more than half of total revenue comes from outside the United States.
“a little more than 20% of that is happening on Instagram”
- Facebook's US flagship app is slowing, while Instagram is growing so quickly that combined US growth still looks strong, and outside the US Facebook Blue still grows while Instagram is growing everywhere.
“Facebook in the United States, the traditional Facebook blue app.com, is a slowing asset. However, Instagram is growing so quickly”
- Elevation Partners became one of the first non-venture-capital investors in Facebook around 2009, part of the pre-IPO private investing category that has since grown large.
“we became the first non-VC investors in Facebook. This was back in 2009 or so”
- Facebook's roughly 3 billion users outside China will grow to about 4 billion within a few years, and there are only more people using more Facebook services over time than there are people leaving, regardless of media narratives.
“Within the next handful of years, that number's going to be four billion. So no matter what you read in what piece of media, there are only more people using more services from Facebook over time then there are people defecting from them.”
Customer acquisition stays centralized on Facebook even as other industries decentralize.
- When COVID hit, travel advertising dollars dried up but e-commerce dollars flowed in quickly, keeping Facebook's headline revenue steady, while Google was more exposed to travel and less able to absorb new industries.
“Well, travel ad dollars dried up, but e-commerce ad dollars came rushing in shortly thereafter”
- Decentralization is better understood as a cyclical reality rather than a technological improvement, and even though merchants have cheap tools, they still earn lower margins outside marketplaces like Etsy that aggregate demand, so decentralization has not significantly reduced the importance of aggregators.
“our experience with decentralization is not so much a technological improvement as a cyclical reality.”
- Even as finance decentralizes into neobanks, crypto, and tokenized assets, customer acquisition remains centralized in Facebook, so the industries being disrupted send their customer acquisition spend to Facebook and its properties.
“What's happening with all of these other industries that are becoming decentralized is it's sending them to Facebook and Facebook's various properties as a customer acquisition channel.”
- Across all channels where Everlane could acquire customers, about a quarter of acquisition spend went to Facebook, and a little under half of revenue growth from about 15 million to about 100 million came through performance advertising.
“And of all the money that was spent across all of the different places on the internet where you could acquire customers, about a quarter of that went to Facebook.”
- Facebook counts about 200 million businesses using some of its free tools, each with under $10 million in revenue, and some of them will break out and need to scale, and Robert sees no better place to allocate customer-acquisition capital.
“Facebook counts about 200 million business users that are using some combination of their free tools.”
Facebook's multi-app super-app strategy should extend into payments and commerce monetization.
- After desktop, Apple and Google became gatekeepers of mobile screens, so Facebook concluded it needed multiple apps, and it bought Instagram and WhatsApp to replicate its desktop position across three apps.
“And they bought Instagram, they bought WhatsApp. And so in order to have a similar position that they had on the desktop browser, they now needed to have three different apps”
- WhatsApp, with about 50 million business users, is likely to be monetized as an enterprise and super-app services platform with payments and customer service rather than through ads, and Facebook could run it at cost as a high-quality funnel into its ad platform.
“They have previewed that they have 50 million business users on it. So I think they're teasing that it may become more of an enterprise like product”
- Facebook and Instagram already operate as super apps in the Western sense, with newsfeed, search, marketplace, and shopping, and the open question is how far they can extend beyond that, with payments the most obvious next addition.
“So when you think about whether or not Instagram and Facebook are super apps, I think they are already super apps.”
- Peer-to-peer payments such as Venmo and Cash App are the most obvious missing icon for Facebook and Instagram, and Robert expects that sending cash or crypto inside the apps is only a matter of time despite past hiccups.
“I think the most obvious admission is payments. CashApp. Venmo. Why are these social peer-to-peer payments things happening on other platforms outside of Facebook and Instagram?”
- Facebook Marketplace, Shops, and Instagram commerce are ways to further monetize the strength of the existing network rather than build new audiences, extending both platforms' long growth runways.
“So some of the new things that Facebook is doing is both strengthening and monetizing those platforms further. So Facebook Marketplace, shops, Instagram commerce, those are ways to further monetize the strength of their existing network.”
Long-term platform health and next-platform infrastructure justify investment over short-term revenue.
- Facebook is investing heavily in Oculus, which Robert calls the leading physical VR hardware, and lowering its price, because even without monetizing a third platform directly, being critical infrastructure there preserves competitiveness.
“Even if they don't find a way to monetize this third platform, so long as they are a critical part of the infrastructure of the platform that comes after mobile, they will be able to ensure that Facebook's position within it remains competitive.”
- Facebook shut down the gaming promotions in its newsfeed that had been crowding out user experience, even though it hurt revenue, because it judged long-term platform health more important.
“And they shut it down. And they said, 'You're not allowed to run these types of promotions for these types of games anymore.' And it's going to suck for our revenue for a bit.”
- Facebook's experiment with paying creators for short-form video, to compete with TikTok, could change the look and feel of Instagram and Facebook more than any other initiative, though it remains an experiment.
“Facebook is putting a toe in the water of paying creators for content. I think that is the angle that could change the look and feel and experience of the Instagram and Facebook platforms the most”
- The biggest existential risk to Facebook is failing to become a critical piece of the infrastructure of the next computing platform after mobile, because losing that shift would reduce its strategic priority within the platform owner's ecosystem.
“whatever the platform, if one shows up, after mobile, that shift and how Facebook navigates that shift is the most existential threat to the company”
Founder leadership gives Facebook vision and candor despite scale-driven missteps.
- Because Facebook is the largest founder-led company, the founder's continuing leadership provides consistency of vision, execution reliability, and moments of candor that boards or politicized management teams rarely produce.
“The good part about it is that the same person that founded and created the company is the same person that's leading it today, which gives you consistency of vision.”
- Facebook's political and misinformation problems are a far more complicated version of the earlier ad-quality problem, and government will impose rules on social network advertising, with a negotiation among public, government, and company that is still underway.
“So I think there's a great negotiation that is taking place between the public, the government and the company.”
- Robert believes Facebook's intentions are not to harm the world, and he sees its missteps as a byproduct of scale and influence.
“But I have no reason to believe that Facebook's intentions are to ruin the world.”
- Facebook's original move-fast-and-break-things culture persists in practice because Zuckerberg still runs the company, even though the company publicly walked away from the slogan.
“even though they've stopped saying that and have publicly said they've walked away from it, the same guy is still running the place.”
- Robert speculated that Mark Zuckerberg already knows what the third platform will be, before Robert himself figures it out, because the founder thinks about the question constantly.
“I'm willing to bet that Mark Zuckerberg knows what the third platform's going to be before I ever figure it out.”
Search and streaming are distinct categories, so Netflix is not a fair comparison to Facebook.
- Search advertising is high-intent behavior, which Robert places in a separate category from social networks, with Google and Amazon as the main players in that marketplace.
“And then the third is simply search, which is high intent based behavior, which is primarily Google and Amazon”
- Netflix is not directly comparable to Facebook because it is a paid content platform that monetizes through subscription, while Facebook monetizes attention by selling advertising and does not pay users to post.
“Netflix is a paid media company. Facebook is not paying users to post”
- Digital advertising splits into social networking, streaming media, and search: streaming (Netflix, YouTube, TikTok) competes for paid content, while search (Google, Amazon) captures high-intent behavior, so Netflix is not a natural comparison to Facebook.
“Netflix is a paid media company. Facebook is not paying users to post”