Meta CMO: How Advertising Works Under the Hood

I’m done with the doom around online ads—they bankroll the open internet, let small teams test and scale ideas, and they’re the real business under the hood.

Antonio praised Alex’s new book as the go-to on growth and performance, so we dug into why ads fund almost everything online, how data really works, and where AI takes targeting, creative, and even transactions. We also hit Instagram’s alternate timeline, WhatsApp as a stealth super app, and Meta’s shift from a friends feed to AI-ranked content.

Antonio, what grabbed you about the book and why bring this conversation together?

It bridges boardroom strategy and the gnarly details of broken pipes and skewed analytics, which is the truth of marketing: you need both the funnel theory and the hands-dirty debugging in one mental model.

Alex, what were you trying to say that people still miss?

There wasn’t a practical, end-to-end playbook I could recommend, and I wanted to counter the reflexive cynicism about ads with a clear case for why they create access, growth, and opportunity.

Before defending ads, let’s steelman the critiques—what were people worried about?

People felt spied on and misunderstood how targeting works, which fueled myths like phones listening in; the reality is first-party signals and lookalikes, not microphones, and the shadiest behavior was mostly third-party data brokers, not platforms like Facebook selling data.

The severe critique frames it as surveillance capitalism—users as product and marketers as manipulators—and adds that ad incentives sour content itself.

Historically, ads often pushed news toward broad appeal, while pure subscriptions can pull outlets toward serving a specific tribe; incentives matter in every model.

Big picture, growth lifts living standards and pays for public goods; you still need guardrails, but treating advertising as the villain misses how often it just helps people find products they actually want.

Antonio, if you updated your earlier Facebook era take, what’s materially changed?

The hot zone moved from ad rules to organic content moderation, and how much platforms should weigh in on truth versus being open forums.

The bigger shift is product: feeds moved from mostly people you follow to mostly unconnected content driven by modern AI ranking and short video, with TikTok forcing everyone to compete in that quadrant.

Define connected and unconnected for listeners.

Connected is from accounts you chose—friends, pages, groups—while unconnected is anything you didn’t follow; each platform started in a different quadrant of public or private, connected or not, and now they all play across the grid.

Given that shift, how do you think about AI and engagement moving from friends to feeds, and maybe to chatbots?

Ranking has always been machine learning at scale, and now we lean into algorithmic campaigns and focus creativity on the data we feed them; AI content will also become a thing people seek, so we have to be great at both ranking it and creating with it.

You sketch a future of hyper‑personal media—what do AI ads look like when the feed is built just for me, and maybe the browser fades into a bot?

Search could split between paid AI assistants and a freemium model that inherits search ad principles, while creative will merge with targeting so an ad talks to you, sizes to you, and lets you buy inside the thread; agents will also advertise to agents, and the core performance rules still apply.

Do we head toward a world where more of life is free if you take ads everywhere?

Paying people for their data is a dead-end because the raw data isn’t worth much on its own and invites fraud.

Think of data as sand that only becomes valuable when refined into useful systems, and remember we already enjoy world-class products free because ads fund them; just be careful with incentives, because misaligned payouts create abuse.

Why didn’t Meta ship a single super app like WeChat?

WhatsApp is closer than people think: its Status sharing is bigger than Instagram or Facebook globally, payments in India are growing fast, and WeChat’s breakout was driven by cultural and banking quirks like viral red envelopes that are hard to replicate elsewhere.

Outside the U.S., WhatsApp is the business communications layer and social glue, and many of tech’s societal ripples come from putting a networked computer in every hand, not just from feed ranking.

Retail media is another big shift, with ads inside retailers using first-party data and native checkout; it’s powerful for performance and dovetails with privacy law.

It also lets platforms double-charge brands with placement and transaction fees, but it’s really a digital replay of paid shelf space and catalog placement, and first-party data plus algorithms give the giants a head start.

If you’re charitable, privacy rules nudged the market toward first-party relationships where incentives to respect users are stronger.

Any other concrete platform shifts you’d flag?

AI agents will need personal context to be truly helpful, which means thorny privacy work, but personalization will surpass what we have now even as cookies keep hanging on.

Direct mail keeps morphing from letter to email to SMS to push and now into a chat thread, and in many Asian markets AI chat commerce and AI support are already at scale and changing the game.

Identity keys keep changing—from mailing address to cookie, device ID, platform ID, and now a wallet address—but the fundamentals of addressability and attribution stay the same.

Alternate history: if Instagram had stayed independent, what happens?

It likely lands closer to Snap or Twitter scale, not zero but not today’s Instagram, and early on we clashed over building a true growth product org and balancing public figures with friend connections before aligning.

What if Libra had been allowed to fully launch?

The climate now is far friendlier to crypto, with rails moving under the hood so people can hold dollars or equities on chain without friction, which opens real financial access.

There’s also a national security angle: it’s better if trusted, Western messaging platforms carry compliant payments rather than ceding that layer to opaque networks.

Will AI crown the incumbents or upstarts?

It’s genuinely open—some incumbents like Microsoft are moving well, but it’s unclear if the lasting experience is a consumer app or a thin API layer embedded everywhere.

Leadership teams either reinvent like Google and Amazon or fade like Yahoo and eBay, and moments like this decide which path you take.

Meta’s talent model seems tightly aligned to output; will that become the norm?

The best companies already judge outcomes over hours.

AI is a force multiplier that makes top individual contributors even more valuable, shifts work up the stack toward logic and creativity, and likely widens pay for the highest-impact people.

Do companies end up with fewer employees, or more?

Some existing work gets automated, entirely new work appears, and previously uneconomic work becomes worthwhile, so headcount depends on how those three forces net out.

Over time, technology creates more roles than it destroys, and as intelligence becomes a utility, the human touch and problem framing matter even more.

Five-plus years after the metaverse bet, where are we and where is it going?

The vision mixed AI assistants with audio and visual glasses, and that’s arriving faster on the AI side than on AR optics, so expect everyday glasses with always-on AI and useful overlays to feel normal within the decade.

Some tech waves take decades to mature, and VR may simply be earlier in that arc even if it ultimately lands at scale.

Could anyone else have owned mobile the way Apple and Google did, or VR now?

Timing waves is hard; people declared the year of mobile for ages and still missed the real inflection, so probably not.

As we close, what’s misunderstood about Meta right now that you want to set straight?

Ads are useful when done right and people prefer personalization to generic messages, and my book is meant to teach the fundamentals so builders can measure, target, and create responsibly and effectively.

The book is Click Here; this was a great conversation—thanks for joining us.

Thank you.

Thanks for listening to the a16z Podcast; follow and rate the show, share it with a friend, and find more on YouTube, Apple Podcasts, and Spotify; you can also follow us on X and subscribe to our Substack. This conversation is for information only and not legal, business, tax, or investment advice, and a16z or its affiliates may hold positions in companies mentioned—see a16z dot com slash disclosures for details.