Google 第二部：Alphabet（音频）

Are you intentionally wearing a black turtleneck for this one?

No. It is actually going to be one of my carve-outs, though.

Amazing.

What? You think I dress up like Steve Jobs for a Google episode?

Well, I thought because of the war between Android and—

I walk in and there's this smirk on your face.

In the late 1990s, Google built the best search engine for the rapidly growing internet. With a breakthrough search algorithm, low-cost servers based on commodity hardware, and the best business model of all time—search ads—they turned that search engine into a cash-gushing business and took it public in 2004.

But then, curiously, they started doing some things that weren't related to search. They launched a breakthrough email service in your browser with Gmail, maps that were far superior to the current state of the art, and Docs and spreadsheets with real-time collaboration for the first time. Of course, there was YouTube, then Android, and their own web browser with Chrome.

Astonishingly, today Google has 15 products with over half a billion users. Seven of those have over 2 billion users. David, that means over 25% of humans use 7 of Google's products.

Just unreal. I can't wait to tell all of these stories today.

Yes. And they've also launched some colossal failures. Plus, to try to compete with Facebook, there was Google Wave, Buzz, and about half a dozen messaging apps—maybe a dozen messaging apps over the years. Hot-air balloons to provide wireless internet. And, of course—

Oh, man, I forgot about the hot-air balloons.

Google Glass.

Can't forget about that one, unfortunately.

So why did they do all this? As a business, Google was and still is the company that makes the vast majority of its money from ads on search results on the web. So today, we tell the story of Google as the innovation factory of the 2000s, its reorganization into the parent company Alphabet, how all these different products cleverly serve different business purposes, and how it all feeds into Google's original core mission to organize the world's information. And we'll end this episode's story right at the dawn of the AI era.

Oh, you're giving away the end.

Oh, spoilers. Sorry. So is Google a search engine? Is it the platform company of the web era? Or is it an incubator that just happens to have struck gold with search and perhaps AI? Today, we dive in.

David, where are we starting this Alphabet story?

Oh, I have a very, very fun beginning for you, Ben. I want to start with a quote from Russ Hanneman—

The fictional character—

From the HBO show Silicon Valley.

Awesome.

And the quote is: “If you show revenue, people will ask how much, and it will never be enough. The company that was the 100x-er, the 1,000x-er, is suddenly the 2x dog. But if you have no revenue, you can say you're pre-revenue. You're a potential pure play. It's not about how much you earn; it's about what you're worth. And who's worth the most? Companies that lose money.”

Immortal words of wisdom for the technology world. God, that show was so good. Why do I bring this up? Why do I start here?

Why are you talking about this? Google is a cash-gushing machine. Revenue is obviously not the problem for Google. But what was the problem in 2004, 2005, and 2006 was being viewed as, in Russ's terms, a pure play.

When Google went public in fall 2004, the stock shot up, basically doubling in 2 months. Wall Street loved Google. AdWords, the search business model—everybody had to own shares. Google had cracked the code on monetizing the internet. The more people use the internet, the more they search. The more they search, the more money Google makes. Simple, easy, pure play, you might say.

Yep.

That is, until Google announced its fourth-quarter 2005 earnings. Full-year 2005 revenue was $6.1 billion. That's almost double the $3.1 billion it had in 2004, the first year it was public. But earnings are flat. Profitability is down.

Google's now investing in all these new products and services: Gmail, Maps, and the forthcoming Google Docs. Later that year, in 2006, they would buy YouTube for $1.6 billion. Wall Street hates this. Hates it.

This is a huge amount of their cash they're putting back on the table and betting for the future.

So this is January 2006. The stock falls 27%. Wall Street's like, “God, these guys, what are they doing? They're messing it up.”

Steven Levy writes in In the Plex that the perception of Google's ventures beyond search at the time was that the company was “tossing balls into the air like a drunken juggler.” They were a pure play in investors' eyes, and now they're messing it up. They're adding all this other stuff. They don't want the other stuff.

Yeah. So then, as you teed up in the intro, the question is why did they do all this? I think the way to answer it is to start by telling the stories of all the individual products.

Let's do it. Strap in. I will say, Ben, doing the research took me way back to the early Acquired grading acquisitions. This is the cornucopia of hits of iconic product launches in tech history.

So the first, and probably the most important here because it sets the stage for everything else, the first major non-search product was on April 1st, April Fool's Day, 2004: Gmail. The most famous, infamous non-joke April Fool's Day announcement of all time.

Yes, but it sure sounded like a joke.

Here's the announcement: in 2004, entirely web-based email in your browser. You can log in and access it anywhere, on any device. Google search is built in. You don't need to spend all this time sorting your mail into folders anymore. And 1 gigabyte of storage, free. No need to delete your mail. No need to clean up your inbox. No need to do anything ever. And the whole thing is free.

Yep. Of course, this sounds like a joke. This is too good to be true.

The universe at the time is Microsoft selling enterprise-grade mail for a lot of money, or there are all these free web-based services popping up, like Hotmail, which Microsoft would end up buying, Yahoo Mail, and AOL. You get 5 megabytes of storage.

Yeah, not even. At the time, Hotmail, which, as you said, Microsoft owns, had 2 megabytes of free storage, and Yahoo Mail had 4 megabytes.

There's another great story from In the Plex. Steven Levy is interviewing Bill Gates at the Newsweek headquarters office in New York shortly after Gmail comes out, and they start talking about Gmail. Bill can't believe it. He's offended by Gmail because he thinks that giving people all this storage is just wasteful. “You're doing email wrong. It's morally repugnant to leave all of this email sitting right on the servers.”

I was thinking about it. Until Gmail, the paradigm for email was that people treated it like regular physical mail. You sort it. You file away the important stuff. You throw out the pieces you don't need anymore. I mean, even Bill Gates operates this way.

Yes.

So Gmail—this is radical. This is a radical notion of how email should work, and it was also correct. If you sat and thought about it in, say, 2001 or so, when Gmail started getting worked on within Google, and you thought about the combination of the growth of the internet, which obviously Google has a front-row seat to, and Moore's law, you would logically come to this conclusion: the cost of sending, storing, and searching email would asymptotically go to zero. And thus, as that happened, a whole lot more email was going to be sent in the world.

So can I tell you my understanding of where this story starts, in 1996?

Oh, I was going to go back to 1999, but yeah, go for it.

All right. So I know you're about to bring up the name Paul Buchheit. Is that right?

Of course. Yeah.

Paul was kind enough to speak with me before recording this episode. Paul, famously, was the inventor of Gmail. In 1996, Paul was a student at Case Western Reserve University in Cleveland, which you may also know, David, famously was one of Ohio's first campuses to have broadband internet.

Yes, one of the first campuses in the nation to have broadband internet in the dorms and all over campus.

Oh, okay. I knew about Paul's fascination with webmail starting in college, but I didn't realize that Case Western had broadband. So when you're living in the universe of broadband everywhere, he was living like 15 years in the future temporarily, for 4 years in college.

Yeah. 1996.

Yes. So he realizes email is kind of a bummer if it's a thing that you download and that lives on your computer.

The information should just exist at my fingertips all the time. Bits are becoming free to move around. So he gets obsessed with this idea in college that email should exist on the web, in a browser, without ever having to download it. And he builds a prototype for webmail when he's in college.

Wow.

In 2001, famously, pre-IPO at Google, Larry Page feels like Google is moving a little bit too slowly and gets rid of all engineering managers. So Larry and Wayne Rosing, who is leading engineering, go and meet with each engineer individually to talk about ideas that they could work on. This tells you so much about Googleyness, but it also tells you a lot about the caliber of the engineers they were hiring at the time. They would just approach them and say, “What ideas are you thinking about? Here are some ideas we have. Can you just full-stack own this product entirely yourself?” In Paul’s meeting, they knew about his previous interest in email and web-based mail, and they sort of floated this amorphous idea to him. That’s where it comes from.

Ah, so Larry and Wayne suggested it to him. Interesting.

Here’s some other stuff that Paul said. Part of the motivation was that they were looking to make something that would make Google stickier. You’d have this ongoing relationship, so if there was a next Google after Google, there was some reason why you would still have a relationship.

Which, obviously, Yahoo would have for many, many years, even though there was a next Yahoo after Yahoo in Google.

We still get emails from people with Yahoo Mail.

Do you know how Paul found out about Google in 1999?

Oh, no.

Slashdot.

Really? That’s awesome.

And he sends an email to jobs@google.com.

Unbelievable. Fitting that he gets hired with an email. Heyo.

In 2001, Paul gets to work with encouragement from Larry and Wayne. Do you know what the original seed of the code is?

Oh, no. Go for it.

Google had just bought a company called Deja News, their first acquisition. It was the corpus of all the old Usenet posts.

Oh, yeah. Then this becomes Google Groups, right?

That’s exactly right. And Paul’s working on that. Part of that was a feature to do real-time indexing of all the posts that would allow you to search the whole corpus. So Paul just applies that to his own personal inbox. The first instantiation of this is just a search box to search his personal Unix mail directory as if it were the old Usenet posts that they had just bought. That’s the first version of Gmail.

Amazing.

As he’s building on that, the first thing he needs is a web front end, an interface. Hotmail’s out there, Yahoo Mail’s out there, webmail’s out there. It sucks. It sucks for a lot of reasons. There’s got to be a way to make it better, make it more performant, and make it better to use as a web page. So he’s playing around with JavaScript and what he can do with JavaScript to make this web application of email better.

The history of JavaScript is fascinating. Brendan Eich created it at Netscape back in 1995. We did a whole episode with Brendan years ago about this. The idea behind JavaScript was to include a programming language as part of web browsers so that people could make dynamic web pages instead of just static HTML documents. The problem was that it was a casualty of the browser wars with Microsoft and Internet Explorer and everything that killed Netscape. Up until this time, in 2001, JavaScript existed, but it wasn’t super popular.

It wasn’t very powerful. You could do weird stuff like animate something on the page, but I would describe it as toy-like and not a real programming language, for sure.

Yep. And for what the web was up until that point in time, you didn’t really need it. Static web pages are fine for most of what’s happening. Even google.com was static. You type a search into the search box, Google’s servers process the query, and they send you a whole new static web page with the results. But you’d imagine that for doing something like email on the web, or any application on the web, you don’t want the site to reload every time you open a new email, create a draft, or move something around in folders.

You might want to move from a website to a world of web applications.

Yeah. But this is how Hotmail and Yahoo Mail worked. Every time you took an action, it reloaded the page, and so they were super slow. Paul’s thinking, “Maybe I can use JavaScript to make this better.” He’s working on it, and he discovers a little-known feature of JavaScript called XMLHttpRequest, which lets a web page automatically fetch new XML data from a server without reloading the page. Paul’s like, “Oh my God, this is gold.” This is the birth of Ajax: asynchronous JavaScript and XML.

So, David, I assumed you were going to go here. I thought you had it all laid up. You’ve been letting me go. You’ve just been feeding me a rope the whole time. You’re trying to tell me that Gmail is the first Ajax application.

Well, the first widely adopted one around the world.

That’s fair to say. That sort of set the bar for what dynamic Web 2.0 websites could be.

Yes. The origin of XMLHttpRequest is a part of Internet Explorer, first implemented by Microsoft and used in this part of Outlook called Outlook Web Access.

I think I did know this. When I worked for my high school, I could log in on any computer into my Outlook through their web access, and that thing used Ajax. I think it only worked in Internet Explorer. So that is the origin of why this API exists in the first place, ironically, for another mail client.

It’s so deeply ironic that this originated for a Microsoft mail client.

Yes.

We’re going to get deep into that in just a minute here.

Yes. When Paul discovers this, it’s almost like Google Search all over again, when people realize what you can do to create something that looks and feels like—and has all the functionality of—an application that heretofore would have been a program that you installed on your personal computer, or an app on your Mac that maybe you downloaded from the internet, but more likely you went to a retail shop like CompUSA and installed on your computer. You can now just do this in a web browser. This is incredible.

The web is the platform of the future.

Yep. So Paul builds the prototype and shows it to Larry and Sergey. They’re super jazzed. Supposedly, Larry and Sergey become the first beta users of Gmail. They are the seed Gmail users, and they start using it exclusively as their mail service within Google. By the time it launches publicly, all of Google is on Gmail, using it and addicted to it. It wasn’t called this at the time, but it’s in the cloud. You don’t have to have your mail stored on your machine or a specific server. You can log in and access it anywhere, on any network, on any device.

All this stuff sounds so boring, but it was completely breakthrough.

Larry and Sergey are jazzed first because of the incredible nature of this product. Larry especially is a product person, and his view is, if we can build a better product and it’s on the web, then it’s good for Google and we should do it. That is a huge part of the motivation underlying Gmail and everything we’re going to talk about. But there’s also another reason, and that’s Microsoft.

Google was doing great, printing money with AdWords and search—the greatest product, the greatest business of all time. But they’ve got a big risk, which is that everything about Google, everything about the web right now, flows through Microsoft, flows through Internet Explorer.

Yeah. Google’s entire money-printing machine was built on top of Microsoft’s, and at 2 layers. At this point, over 90% of Google search queries were done on Windows PCs, and 90% were done in Internet Explorer running on those PCs. So Google’s got the killer app for the web in search, and the thing under them is a browser owned by Microsoft. The thing under that is an operating system owned by Microsoft.

Yes. They exist at the pleasure of Microsoft at this point in history.

And Microsoft has a different business model.

Google’s business model—the greatest of all time—is that people use Google Search. They discover more of the web. They spend more time online on these new sites and services that they’re discovering. As they’re spending more time online, they search more. Searching more leads them to discover even more new sites and services. The cycle repeats itself, and Google just monetizes the whole thing.

Yes. Web usage isn’t bad for Microsoft, but if the platform of the next generation becomes the web and people are writing web applications instead of Windows applications, that makes Microsoft’s platform a lot less valuable versus other operating systems, like the Mac, or a future where we change away from desktop computers altogether.

At a minimum, Microsoft, business-model-wise, doesn’t care about the web because they don’t monetize the web. Microsoft makes money by OEMs selling PCs that have Windows on them, and then Microsoft sells software that goes on those PCs. At a minimum, they don’t care. At a maximum, like you’re saying, web apps are an existential risk to Microsoft. There’s a future application platform that just doesn’t really require their participation, other than the fact that they control Internet Explorer. At least for now, that’s really important.

And most of Microsoft hasn’t realized this yet. Thank God for Google. Microsoft’s distracted with the albatross that was Longhorn, which would become Windows Vista.

Yes.

A few people in Microsoft realize this, but Google for sure realizes it, too. Eric Schmidt for sure realizes it because he was the CEO of Novell before coming to Google. And who is Novell's competitor? Microsoft. And Microsoft crushed them.

So why is Google so jazzed about Gmail? They need to build up leverage with consumers, with users who will demand rich web applications, so that if Microsoft ever tries to disadvantage Google or disadvantage web apps and things moving to the web, really, the only defense against that is if consumers have already adopted this stuff and love it and would revolt. And so this is what Gmail is.

Yes.

So Gmail development is trucking along through 2001, 2002, and 2003. This is hard to remember now. It took 3 years to develop Gmail.

Long development cycle. Yeah.

To be ready to release publicly, and then it was in beta for like 10 years.

Yeah.

I think the reason it took so long was that this was all new. There wasn't a lot of depth of knowledge out there about JavaScript. Certainly not about Ajax and XML dynamic refreshing.

It was really hard to program. Today, you've got all these nice abstraction layers, these frameworks that people have built to do web development that really didn't exist for making Ajax applications.

Yes. Okay, so Google is finally getting ready to launch it. We're in 2004. There are a couple of questions. One, the service, for all the reasons we just described—Google, Larry, Sergey, Eric—they want it to be so compelling that consumers demand it. It takes off like wildfire. It builds this strategic moat against Microsoft, but it will cost money.

There's a reason other people don't do this.

Yeah, there's a reason that a gigabyte of free storage seems a little crazy. Even if you assume—and I think this is probably directionally correct—that because of Google's commodity infrastructure advantage, they could launch Gmail at like one-tenth the cost that anybody else could, remember that there's no public cloud at this point in time.

So you'd have to go build your own data center to do this.

You can't just launch on AWS. There is no AWS. But even assuming that Google has a 90% cost advantage on the infrastructure side, the state of the art is that other competitors are offering 4 megabytes of free storage. Google's going to offer a gigabyte. Sure, knock that down by 90%, but the effective cost is still 100 megabytes. So how do you get around being flooded with cost and infrastructure demand when you launch it? They come up with the invite system.

Yes.

And this is so brilliant. I actually don't know if it was designed as this prestigious growth-strategy thing that it became.

Anyone got any Gmail invites? Please, I'll do anything.

Yeah. Yeah. Please, please, please. Or if it was truly because of the infrastructure cost. Either way, it's just brilliant. When they launched it on April 1, 2004, they sent out 1,000 seed invites to Gmail. It's a private, invite-only internet service. They sent them out to influencers. The term didn't exist back in the day, but influential people and journalists. And then each user had a set number of invites that they could give to other users to invite their friends.

And it was low. It was like 5 or something. And then it wasn't clear when they would top back up. But you'd give out your 5, and then at some point you'd come in and you'd have 5 more. You'd have 3 more. It was super dynamic and very clearly whatever Google felt like they could give away from their servers at the moment.

Yep. But it was so brilliant. It made it feel like you're in this special world of people in the know, with super-incentivized viral word-of-mouth growth, because I'm telling you it's a gigabyte of free storage. It's this incredible service. They were selling on eBay for 150 bucks. There was a monetary value to these things. Yes, yes, they were trading on eBay for an average price of 150 bucks in the early days. And so I'm giving you this gift.

Incredible.

And look, everybody wants this, but you need to have the product quality that cashes the check.

Yes. It needs to be a real gift.

Right? And it was just better. It wasn't just something I'd sign up for and then churn and be like, “Cool, I locked in my username or whatever.” It was something that you actually used every day. Or, in the words of Larry Page, it passed the toothbrush test. It was a part of your daily habit, something you do once or twice a day.

I wish I could only refresh Gmail once or twice a day.

So, David, was this the first software that used a waitlist like this? Because obviously, it's become very popular since.

I think so. So that's how they take care of the cost side of the equation so it doesn't run out of control: the invite strategy.

Well, still not making any money, though.

That's question number 2. How are we going to make money from this thing? Because, yeah, okay, there are all these strategic reasons to do it. It'll increase traffic on the web, increase time spent, and make people search more. We'll make more money indirectly, but they still don't really know that. So they think, okay, we need a monetization strategy baked into the product itself.

Yes.

Well, how do you make money from anything at Google?

This actually came up during development. So even in the prototyping phase, Paul logs into the database of ads. It's just funny that, at that point in time, Google has this big database of ads.

Right? Yeah. I'm just going to access the ads database—all of them.

Yes. And these are the ads that would run when you searched and landed on a search results page. And so he decided to do content matching against your inbox and just show those ads on the page next to your email. And even though they weren't meant for that, it actually turned out that these search ads were pretty relevant. It actually was a decent ad to be showing you while you're looking at your inbox about similar topics.

So he just rolls this out. Even though all these people in Google are actually using it as their mail client at the time, people were pissed. People were like, “Are you looking at my emails?” All the things that would then come later in public actually happened inside Google first.

But Larry and Sergey loved it. They were like, “Oh, this is so obviously the answer.”

Interestingly, this experiment predates AdSense. So Google has the display-ad offering for website publishers that's called AdSense. That's different from AdWords, which is the keyword advertisements on a search results page. AdSense hasn't launched yet. And there are multiple versions of history here: How much credit for AdSense does Gmail get in discovering this?

But it is safe to say that the idea of display ads that are content-matched against your Gmail did contribute to the idea for the first version of AdSense, which was essentially the same thing: content-matched ads just on a publisher website instead of in the content of your inbox.

So the product launches publicly in April 2004. As you'd expect, people go nuts. It is truly a revolutionary product. And Gmail grows over the next 20 years from that 1,000-user initial public-beta seed base to over 2 billion today. And it's still by far the best email service. Even if you use another front end for your email, for your Gmail, like Superhuman or whatnot today, you still want Gmail on the back end, at least as a consumer.

Yes. So once Gmail starts to take off, Larry and Sergey and Eric see this and they're like, “Wow, we should do this a lot. Let's go.” Let's build as many web applications as we possibly can imagine.

What else can go into the browser that we didn't think was possible before? This fires on every single cylinder for us. Most importantly, grow the web.

Grow usage.

You grow the web, you grow the time that people spend in web browsers. They will search more. We will make more money. And beyond that, with some of these products, like Gmail, we can monetize the products themselves. Great. Two, we are building our strategic moat against Microsoft. The faster that we get the internet-using public to fall in love with and use web applications, the less and less leverage Microsoft has over us.

To use Ben Thompson speak, Google realizes the web can become the point of integration. Maybe the OS isn't what the whole universe has to target: the hardware makers, the OEMs, the application makers, the users. If applications start living in the browser, then the web can become the point of integration. Users just need a browser, and OEMs just need an operating system that can access the browser.

And what's so great for Google because of their business model? Sure, it's great when they build and own and operate and run and monetize web applications themselves, like they do with Gmail, like they'll do with Maps, like they'll do with Docs, like they'll do with YouTube that we're about to talk about. But if they don't, it doesn't matter as long as anybody does it.

Right? They just need to be the wind at the back of web adoption.

Yes. So that leads to a whole flood of Google web products and services to come.

All right, David. So, Gmail: we've got our existence proof of an Ajax-based web app. It's going viral. People love it. We can really build web applications now. Let's go nuts.

Yes. So the next big web apps following Gmail were Maps, Docs, and Spreadsheets. All absolutely incredible.

And it was not clear that these things were possible with web technologies.

These required incredible technical and product vision. So, first, Maps. We actually did a whole Acquired episode back in the day just about Google Maps.

The 3 companies they acquired.

Yeah. It starts in 2003, so even before the Gmail launch, when a young associate product manager—APM—at Google named Bret Taylor—

Of course, of ACQ2 fame, Bret Taylor.

Recent ACQ2 guest Bret Taylor. Also, FriendFeed founder, Facebook CTO, co-CEO of Salesforce, chairman of OpenAI—

Former chairman of Twitter.

Yeah, yeah. That Bret Taylor starts his career out of Stanford in 2003 as an associate product manager at Google. He ends up going to Larry and is like, “We're missing out here. AOL has MapQuest, which they've just bought for $1 billion. And I'm hearing through the grapevine that Yahoo is about to make a big push and launch Yahoo Maps.”

And so, as you would expect, Larry's like, “Oh, yeah. Is this a web product?” “Yes, of course.” “Go do this.”

For all these things that we're studying here, there's a business rationale, which might be extremely indirect, but it's there. This idea of increasing web use increases Google Search, which increases the money printer. But then there's also an abstract rationale, which is, our mission is to organize the world's information and make it universally accessible and useful. And Maps is squarely in the middle of that.

Yeah. Now, the thing was, as big as MapQuest and Yahoo Maps were about to become at the time—and they were big. I remember using them. My parents used them. Everybody on the internet used these services—they weren't what you think of as Google Maps today. They were static web pages.

Yep.

They didn't use Ajax. And the whole point was to get driving directions—

That you could print out.

Exactly. And the business model for these services was that, on the printed piece of paper that people would print out, you would put ads.

Yep.

It was like a Trojan horse newspaper business—

Right?

So Brett and Larry and Marissa are looking at this like, “I think we can do better than this.” So they go out and buy a little company in Australia called Where 2 Technologies, which was started by these 2 brothers, Lars and Jens Rasmussen—

Who were incredible engineers, and they had built a real-time interactive maps application, except it was an installed desktop app.

And so they're meeting with them, and Larry's like, “Okay, this is what we want, but we need it on the web.” I think actually the quote was, “We like the web at Google.”

And this is how good of engineers the Rasmussens were: they go off and, in I think 3 weeks, they rewrite and rearchitect the entire application to run as a web app, and they basically independently discover and implement a lot of the JavaScript and Ajax features that Google was working on internally for Gmail. Gmail still hadn't launched yet.

Amazing.

So Google ends up buying Where 2. That becomes the core of Google Maps. Around the same time, they also acquired 2 other companies: ZipDash, which did traffic data, and Keyhole, which would become Google Earth. Now, Google Earth was an installed desktop application. Ultimately, everything that Google Earth was building would get folded back into Maps later.

It's actually not true. I thought that, and just last night I realized you can still go to earth.google.com and get a completely different 3D experience than Google Maps.

Oh, no way.

It's all on the web now. It's unbelievably powerful. Oh, so it is a web app, but it's separate from Maps.

Yes.

Oh, I didn't know that. Oh, I have to check that out.

It's amazing.

That's awesome. Yes. Keyhole and Google Earth, I think, is my favorite part of our first Google episode earlier this year: that the whole thing ended up just being a Trojan horse downloader to get Google Toolbar installed on Internet Explorer on people's systems. It was organizing the world's information and making it universally accessible and useful, but it came with Google Toolbar.

Yeah. The greatest distribution hack for Google Search of all time.

Yes.

Anyway, back to Google Maps and Where 2. February 2005, Google Maps launches. People go nuts. It was a live-mapping, dynamic web application.

Do you want to know my favorite Easter egg from the launch day of Google Maps? I don't know if you know this. When you loaded up maps.google.com, do you know what you visually saw?

I have no recollection.

You saw a great big ocean and North America, and then, floating in the middle of the Atlantic Ocean, you saw the UK, and then there was nothing past it.

They hadn't built it yet.

They hadn't built it yet. Europe, Asia, Africa—not included. It's not even like it's off-limits. It looks like there's an ocean where Europe should be.

How do you decide what the MVP is, or the minimum viable product, to ship on the map? That's amazing. All right, there's one more really important piece of Maps, which is that the next year, in 2006, they released the API.

And this is what really kicks off the Web 2.0 era. Gmail, JavaScript, and Ajax had inspired developers out there, for sure, to make richer web apps, and people were doing that. When Google releases the Maps API, this thing called mashups starts happening. You remember this?

Absolutely.

It's now super easy to grab Google Maps and build stuff on top of it. And it's really hot, and this enables startups. So, like Zillow, Uber, eventually DoorDash, and Airbnb. Think about all the companies that just couldn't exist without the Google Maps API.

There was that whole web of geo-related companies too. Remember that era of mobile, social, local—SoLoMo?

Oh, yeah. Foursquare and Gowalla and, yeah, all those—

All this existed because Google Maps existed.

So, back to Google's overall strategy here and adoption of web apps and sort of building this moat and defense against Microsoft. This is just incredible. I mean, here's Maps itself as a first-class, rich web application that tens, eventually hundreds, today billions—2 billion-plus users—use and love every day.

And now here's this API that's making it really easy for other startups and other companies to go build great web apps, too. The lock-in just keeps getting deeper and deeper and deeper for the web.

Yep. And at first, the API was notoriously free or very inexpensive at very high limits for a long time. That's different now. But for the longest time, it was just, “This is a part of the mission, so we're doing it, and we'll figure out the business later.” It was a very founder-driven thing.

Now, it's popular to create maps. I mean, Apple at some point flipped into doing it, and there are these other third-party companies, and there's OpenStreetMap and all this stuff. For the first 5 to maybe 8 years, Google was kind of the only one that had a passion for this and a willingness to spend into the giant hole that you need to create maps of the whole world.

I mean, it's an incredibly hard data and engineering problem. And they had to go draw all their own maps from scratch, acquire the data, figure out how to get fresh data all the time, and create a crowdsourced thing among Google—was it Google Maps explorers or something like that? All the people that would update these things.

This is an extremely Googley problem and a founder bet: “Nope, we're gonna go spend hundreds of millions of dollars, billions of dollars on this, drive cars around taking pictures of everything, figure out how to not overshare personal information on this, and do it dynamically because you're capturing a huge amount.”

I mean, it's just a wacky, wacky engineering problem that's daunting, and they took it on.

Yep. And we're not going to talk about this today, but put a pin in it for the next episode: Maps is one of the most incredibly strategically valuable data assets for the AI era, and specifically for self-driving cars. Yes. But today, Maps has over 2 billion active users this year. They don't break out revenue, but estimates are that Maps does well over $5 billion in revenue, maybe even $10 billion in revenue.

The larger part of that is ads. You see recommended places to go around you all the time whenever you open Google Maps now that are sponsored ads, just like on Google Search. The smaller part is from the API licensing that you were talking about, David. But this is a real business for Google today.

Yep. All right, the next ones that we've got to talk about are Docs and spreadsheets. These aren't the biggest Google apps out there today. I think if you lump them all together into Workspace and Drive, it is over 1 billion users.

That whole suite is among their most-used products.

That whole—you call it an office suite? Is that what you would call it? Yeah.

Sounds like an office-type suite.

It's a good idea. Someone should do that.

So Docs and spreadsheets hit Microsoft right where it hurts: Office. People have tried both before Google and after Google to compete with Microsoft in productivity forever.

WordPerfect, Lotus Notes, Lotus 1-2-3.

We talked all about that on our Microsoft episode.

By the way, WordPerfect was acquired by and run by Novell. Who was the CEO of Novell? Eric Schmidt. Eric knows all about this.

But here's what I will say, David: if you were starting with the goal of competing with Microsoft or trying to build a word processor or trying to build a spreadsheet, you would be doomed to failure. What Google was doing was saying there is something that is uniquely possible with web applications and Ajax in this Web 2.0 era for the first time, and that thing is real-time collaboration—real-time, multi-user collaboration.

These were, as I've tried to rack my brain—I talked to Sam Schillace, the founder of Writely, which Google acquired and which became Google Docs—he believes these were the first real-time, multi-user collaborative pieces of software in history. It just wasn't possible before the web.

Yeah. Jonathan Rochelle, the founder of the company that would be acquired and would become Google Spreadsheets, basically said the same thing. His comment was, "We actually didn't know if it was possible to do this on the web." Google said, "Based on the success we're seeing with Gmail, I bet we could do actual spreadsheets in the browser with real-time collaboration."

When the Sheets team came in, it was truly an open question: can we make it so you and another person can work on the same, very basic spreadsheet at the same time?

Interesting. The Docs team—Docs was an acquisition. It was a company called Writely that was founded by Sam and his 2 co-founders, who were great programmers. They'd worked together for many years.

I used Writely before it became Google.

No way.

You were one of very few people who did that.

Yeah.

Because it was not an independent company for long.

The product launched in August 2005. Google bought the company in March 2006, so you had about a 6-month window.

Wow. But yeah, they built real-time collaborative word processing as a web app, inspired by Gmail and everything that was going on at Google. The whole company started as, "For our next project, let's explore what we can do with JavaScript and Ajax. What would it be like if we put a word processor on the web?"

They weren't actually even thinking about collaboration at first, but then, as they were working on it together, they naturally started collaborating and thought, "Oh, this is the killer feature."

That's funny. That's different from the spreadsheets team. Their whole thing at first was, "We're not going to make a better spreadsheet than Excel. So if we put it on the web, it has to be about sharing and collaboration."

Yep. And so, to your earlier point, nobody can compete with Microsoft in productivity software. One, because they'd been doing it so long, they had this feature wall of so many features that people needed—

Two, proprietary file formats.

They had a network effect of the file format.

You built your big model in Excel. Good luck.

Other people need to be able to run it on their installed desktop applications. Good luck getting somebody to try downloading or buying a new piece of software and installing it on their machine.

But three—I mean, the biggest by far—the enterprise agreement. This is Microsoft's whole entire business model.

Right? You don't have to be best in breed in any specific thing. You just have to be a platform with everything.

Yep. And IT departments will buy it. Especially for productivity software, really all the money is in B2B and work applications. If IT departments are buying the Microsoft enterprise agreement, they're getting everything. Good luck unseating Microsoft Office.

And I'm not sure you could do this as an independent business, because think about how long Google went with these things before they were adopted by bigger companies. For the longest time, it was, "Oh, a Google Doc—that's a thing for either your personal life, or maybe a startup would use it." But even a medium-sized company? You can't be serious. Get out of here with that.

Google was basically able to subsidize it because they had a giant existing business.

You are so right. Nobody except Google could do this for a whole bunch of reasons. One, you talk about subsidizing. Imagine trying to build this software as an independent company, or really even as any other company. It would require a lot of infrastructure. Real-time, multi-user collaboration in a web app—gosh, that seems like really complicated server and backend infrastructure.

For Google, it's what they do. Running Docs and Sheets, the incremental load to Google's infrastructure was trivial compared to Search. They already had it built out. It was super cheap.

Yeah.

Two, they don't need to make money from it. This is the big reason why nobody else could compete. Microsoft has all the dollars completely on lockdown because of the enterprise agreement.

Big dollars. These small and medium-sized businesses would, of course, pay for something, but those dollars don't add up to be nearly as big.

Right. Exactly. Google, though, that's fine. Microsoft can keep all the dollars. All we care about is people using the web. In this instance, particularly with Office and productivity, this is really about putting the screws to Microsoft a little bit and distracting them.

From Google's point of view, this is a cheap distraction. If this gets Microsoft all spun up, Microsoft is now all of a sudden getting asked all the time, "What's your web strategy for Office? When are you going to add collaboration to Word and Excel?" They don't have any answers.

I literally worked on this. My internship was at Microsoft, and I worked on adding headers and footers to the Microsoft Word web app. We were porting the Windows code to have perfect document fidelity to the web. When you looked on the web and then printed from the web, the document would be laid out pixel for pixel, character for character, exactly how it would look on the printed page.

When you have that requirement, that is a hard, hard engineering task, and it's still not as good as Google Docs.

Right? I love it that this launched your technology career.

Yes.

Amazing. But yeah, from Google's perspective, this is amazing. Microsoft is now forced to bring their crown jewels to the web, which they don't want to do. And Ben, to your point, because they have to make it look, feel, and function exactly like the installed desktop apps, this is going to take them a long time and be a big investment. Fantastic.

And no matter what, it's going to be more complicated, because with Google, it's install-free. There's no licensing. Someone just shares a Google Doc with you. If you have permission to view it, you view it.

With Microsoft, I remember at first it was sort of antithetical. It was like, "But what if I haven't bought Word? Can I just use Word for free on the web then?"

Right?

Is Microsoft okay with that? Am I going to hit some weird usage tier? What? So it's confusing for users. It forces the company to think about pricing and packaging. It was a masterstroke by Google.

Yep. So fast-forward to today. It's hard to get real, actual apples-to-apples data on Google Workspace versus Microsoft Office users. But basically, the way to think about the market is that Google has the vast majority of users and usage of productivity software, and Microsoft still has the vast majority of dollars. And that's fine. Google's super happy about that.

Is that true, that there are more active users of Google Workspace than there are of Office?

Yeah. I mean, I think if you look at users of Docs or Sheets or Slides, it's in the billion-ish, 500-million-to-billion range for each of those. Office, I think, has a couple hundred million users worldwide.

Whoa. Yeah, that's crazy. I didn't realize that.

Pretty wild, right? But to my point about the dollars, Microsoft's Productivity and Business Processes segment—which is mostly Office; I think LinkedIn is now part of this too—last year generated over $120 billion in revenue.

Google reports Workspace as part of the Cloud segment. All of Google Cloud—inclusive of infrastructure as a service, all the AI infrastructure, all that—the whole Cloud segment for Google last year did about $50 billion in revenue, less than $50 billion.

And that's high-margin revenue.

High-margin revenue. Google's Office products are some small portion of a $50 billion revenue segment. So yeah, Microsoft still has all the money.

Google's got all the users, and everybody's happy.

But you're so right. Everyone is happy. This is exactly what Google wants.

Yeah. And ultimately today, Microsoft is fine with this arrangement too. The ultimate fun fact, though, is Sam Schillace, founder of Writely. He would go on to manage all of Docs and Sheets, and I think he actually managed Maps at some point, too. He is now the deputy CTO of Microsoft.

Careers are long.

Amazing. The interesting thing, reflecting on Google's actual business here and comparing it against all the things that we're talking about, is that Google essentially won search by the mid-to-late 2000s. I mean, I know Bing hasn't even launched yet, and we'll get to that, but search was going to continue becoming a more and more giant market.

And so all this stuff they're doing, it's like, oh, we've won, and this market is naturally going to become large. I guess let's just fuel it getting larger and try to do a bunch of stuff under the umbrella of our mission. But what do we really need to do?

And the slightly more altruistic answer, I suspect, if Larry Page was sitting next to us, he would say, "What is the goal of a company?" The goal of a company isn't necessarily to build the largest business. It's to fulfill its mission. And yeah, we got a money-printing machine from search, and we're investing a lot of money still in search and making that better. But all these things fulfill our mission, too.

Yep. And I think these things are all true.

Yes.

So on the back of the success of Maps, Docs, and spreadsheets, this really starts to inform Google's strategy here. Specifically, they've seen, hey, we can acquire these web app, Web 2.0 companies, bring them into Google, turbocharge them, and offer these magical experiences to consumers. We get all this strategic value out of them, both on the offensive and defensive front.

We can operate these things at a fraction of the expense that it would cost anyone else to do so as a standalone company or as part of other big companies.

And some of the things we could buy actually fit into our core ads business quite well.

Yeah. What if we went big with this? Like, really big.

Like something super expensive to run that requires storage of massive videos, bandwidth for streaming these massive videos, and lawsuit protection.

Yep. It probably also costs a lot to buy because it's well-funded by Sequoia. That leads us to YouTube.

All right, David. The YouTube story.

The big kahuna.

The big kahuna. Ah, the most embarrassing thing in Acquired history was our early episode on YouTube.

All right, I have got a proposal for you.

Okay, I'm ready for it. You want to take it out of the feed? Delete it today.

We're setting the record straight. When we finish this section, we are regrading YouTube. We are updating the Acquired canon. It's happening.

Oh, let's do it. We're bringing grading back, baby.

Great. I'm glad you're into it.

I love it. I love it.

Awesome. All right, YouTube, 2003. Same time frame as everything we're talking about here. Gmail hasn't even launched yet. Google starts working on Google Video. The idea is that there's a lot of information in video, and thus it fits Google's mission, Ben, as you were saying earlier.

And also, there's just so much more advertising dollars in TV than anywhere else in the global economy.

To this point in time, TV was the bulk of ad spend.

Yep. If you go look at some of the old Mary Meeker Internet Trends decks from this time period and look at the share of global ad dollars spent on TV versus any other category, it's just so much bigger than anything else.

David, I am so glad you did this. We are brothers. I did the exact same thing to try to tee this up.

Amazing.

I have the stats in front of me. For listeners, digital advertising—you know, Google's universe—would not eclipse TV until 2017 or 2018.

Wow. So, almost 15 years in the future from when we're talking about here in 2003.

Yes. That is the wildest thing, that TV was bigger than digital for that long. Mary Meeker famously had this point that she made every single year: the attention was all in the digital economy, but there was this gap, and the ad monetization hadn't caught up yet. It took all the way until 2018 for the flip to finally happen, where digital overtook television.

Thanks to YouTube.

Yes.

And Facebook, Meta, TikTok, and et cetera.

And the rest of Google, too.

I know, I know. So this Google Video project actually came out of the ads? It didn't come out of engineering and the rest of the Google product?

Of course this fits the mission. There's a lot of information in video. We should totally do this.

Here's how Larry describes it: Google Video was first launched in 2005 as a search service for television content.

Yes. Because TV closed captioning made search possible, and user-generated video had yet to take off. But it subsequently evolved into a site where individuals and corporations alike could post their own videos.

They were digitizing TV because the transcription wasn't as good as it is today, so they needed the closed-captioning data to make it searchable. They were almost like meta-searching: they were looking for other websites that allowed people to upload video and including that in the search results, also.

Yep. Sure. You can see how this conceivably could be a product vision you could have at the time, but Google Video was the wrong product. The problem was, one, you couldn't actually watch the video. It was just search that then directed you, just like Google's main search business model, off of Google Video to go consume it somewhere else.

In the beginning, it didn't even have a player.

Whoa. I didn't realize that.

Yeah, crazy. And the bigger problem, though—another big problem, shall we say—was that the focus was on traditionally produced, head-content kind of content, not long-tail, not user-generated content. It was really tied to TV.

There was a press release that said that they could search the content of TV programs, find programs containing the content they're looking for, and discover when and where the program would next air.

Yeah. So meanwhile, obviously, here we are in 2004, 2005, 2006, and consumer-generated digital video is becoming a thing, either via standalone new devices like the Flip Cam—

Flip was a startup, right? And then Cisco bought it.

Yeah, my other internship employer bought Flip while I was there. This is like Ben Gilbert personal history. But more commonly, there were dedicated devices like the Flip Cam, but digital point-and-shoot cameras had gotten so good by this point in time. This is going to come back up later in the episode.

People thought this was the big consumer-electronic-device vector before smartphones. People were really, really excited about how good and how universally adopted digital cameras were. All of a sudden, in the mid-2000s, for the first time, anybody could make a video at any time.

And iMovie was just becoming a thing. So you could shoot it on your point-and-shoot, and you could edit it on your computer.

That's right. So YouTube, in early 2005, 3 PayPal employees—the PayPal Mafia, actually fairly junior employees at PayPal—Chad Hurley, Jawed Karim, and Steve Chen, leave PayPal and create YouTube.

Okay, Ben, I have 2 deep-cut YouTube corporate-history trivia items for you. Number 1: do you know what YouTube's original tagline was? The name of the company was YouTube. What was the tagline and the value prop?

I have no idea.

Tune in, hook up.

Really?

It was a video dating service.

I did know that. They actually posted Craigslist ads in the Bay Area for attractive women to make videos to post as profiles on the site.

Unbelievable.

They got no responses, as you would expect. Thank goodness for them—and Google—because then they pivoted into a general-purpose video-uploading site where anybody could upload anything. That made YouTube.

So that's trivia question number 1.

Okay.

Trivia question number 2: Do you know who Chad Hurley's father-in-law was at the time? Chad was the CEO.

Oh, no. I have no idea.

Jim Clark, of Silicon Graphics and Netscape.

Didn't know that.

Yeah. So not only were they part of the PayPal team and the PayPal Mafia, they had the best adviser of all time.

Wow. To navigate the Silicon Valley ecosystem and the internet ecosystem with Jim Clark.

The brilliance of YouTube—and it really was absolutely brilliant—was threefold. One, it was super easy for anyone to upload a video, so they had a killer content-acquisition model: anybody, anytime, anything.

As soon as the servers processed it, they would put it live. There were no copyright checks. Unlike Google Video, which would take 1–2 days for humans to pore over the video, make sure it was all good, bless it, and then put it live—which, of course, wouldn't scale in the UGC era—YouTube was just like, “Whatever, upload it.”

Second, it was super easy for anyone to watch a video. You needed a really good viewer in the web app to view the videos, and Google Video didn't have that at the beginning. So YouTube had a killer content-consumption model: go to YouTube.com, find something, or find a link.

The third brilliant thing about YouTube was seeing a YouTube video embedded on another website. Boom, you're watching the video. That was a killer growth and distribution model. YouTube also had great search pretty much from the beginning. You could search YouTube and find the videos you were looking for.

Pretty quickly, YouTube became—and still is, as Google talks about all the time—the 2nd-largest search engine on Earth behind Google.

It's amazing.

Searches were happening on YouTube.

That happened quickly. I always thought that was a more recent, last-10-years phenomenon. I think that happened very quickly. YouTube traffic scaled so fast and so big.

You can see how YouTube wasn't only the correct video platform for the web, doing it much better than Google was doing it with Google Video. There's also some version of the world where they might become a real competitor to Google's core business. If all these searches are happening, they could add search for other things on YouTube, too.

Right? I don't think they had any plans to do that, but it's the same rationale as Mark Zuckerberg saying, “Uh-oh, everyone's using WhatsApp for messaging.” Whether or not they put in a social media feed stream, they always could. It's really dangerous to me for them to be out there aggregating all the users, attention, and habits when they always could do something like that.

Exactly. Same dynamic. Whereas in the previous categories of apps that we talked about, Google had the advantage of uniquely being able to do it as Google in a way that startups couldn't. Here, it's a little bit the opposite. YouTube, as a small startup, had the advantage of saying, “There are copyright rules and laws, but I don't know. We're just a platform. We're just a startup. Anybody can upload anything.”

Google, by this point in time, was a public company. There was no way they could behave like this.

Well, it's funny. They could, but they wouldn't. They actually could do it and stay in business, whereas YouTube could say, “Eh, whatever.” But then it would go out of business because it would get sued out of business.

It's this really interesting catch-22: this is the way to start and get all the users because this is the best user experience, and at the same time, it will not work as a resource-constrained small company. Once it started, it needed to be part of Google.

Right. Yes, obviously we're going to get to that, but in the beginning, oh my gosh. The embeds were a beautiful distribution-growth mechanic for YouTube, but people were just uploading copyrighted videos that people could watch for free.

It's almost like Gmail. It is so unbelievably compelling to a consumer when your friend tells you about YouTube, sends you a link, or you see an embedded page and think, “Whoa, I can go watch ‘Lazy Sunday’ from The Lonely Island and Saturday Night Live in my web browser anytime I want for free, with no commercials.”

Yes, I want that.

In fact, when users started uploading “Lazy Sunday,” the Lonely Island skit from Saturday Night Live, to YouTube, this was in that brief phase where YouTube was an ascendant startup and not yet part of Google. That one skit increased YouTube traffic by 83%.

Wow.

Unbelievable.

And so they very quickly raised money from Sequoia. Is that right?

Yep. It was basically incubated at Sequoia when the 3 founders left PayPal. Sequoia invested right away. I think it was Roelof Botha's first investment when he joined, because Roelof knew them from PayPal. He's also part of the PayPal Mafia.

Exactly.

Sequoia led another round pretty quickly thereafter because the infrastructure costs started scaling astronomically, as you would imagine.

There were 3 things that were very expensive, 2 of which were ongoing. One was a 1-time cost, but it was still expensive: encoding the video. The video might eventually play on multiple types of devices and multiple browsers, so there's a lot of encoding that has to happen.

The other 2 were big, ongoing variable costs. You have to store all this video, and the biggest of all is networking. Bandwidth becomes extremely expensive and costs you every single time someone plays the video. Your biggest cost driver scales with minutes watched, so that is eventually going to kill you unless you have an aligned business model.

Yep. By the way, it would also be really nice if whoever owns and operates this had its own really good, really cheap infrastructure with all of these things built into it.

So, pretty quickly—within a little over a year of launching—YouTube was in way over its head: the content issues, the copyright issues, and the infrastructure-scaling issues.

It's all exactly what they wanted. It's going as well as they could have hoped, and it isn't way over its head.

Yes. If this had happened today, you could probably raise enough capital from the private markets to address this and scale up as a company fast enough. Especially with public cloud, you could probably build this as a standalone company.

Yeah. Today, you can go raise billions of dollars as a Series A startup if you're in the right space, doing the most interesting things with a big market.

2005 and 2006 weren't the same kind of private-capital environment. Of course, there was no way the company could go public with all these issues or anything.

Right. In particular, there was a giant suit from Viacom.

Yes. Because of these things, YouTube ended up basically putting itself up for sale. It had no leverage in content negotiations with rights holders, and the infrastructure was killing it.

So, in November 2006, less than 18 months after the product launched, Google bought YouTube for $1.65 billion in stock.

In stock. I'm glad you caught that, too.

Stock. Yes. We heard in the research that after this deal, Patrick Pichette—I think he was the CFO of Google at the time—said, “Never again.”

He said, “Never again.” This was our biggest mistake.

He said, “Never again. This is the last stock deal that we ever do.”

Google's market cap has increased 20x since the day that this deal closed. If it had paid in cash, Google would have made an extra 20x multiple on whatever you already think the multiple is on its purchase of YouTube.

Yep. The thing is, though—we will correct the Acquired record at the end of this section—either way, even if Google paid 20 times $1.65 billion for this, it got a screaming deal. YouTube is so valuable.

All right. I have some of the numbers from the first few years that I was able to cobble together, and then I want to talk about some of the product evolution over the years.

Yes. Great.

All right. Google bought it for $1.65 billion. Interestingly, Shishir Mehrotra went on the Grit podcast, the Kleiner Perkins podcast, this week and laid out a bunch of data on this. I didn't have a chance to reach out to Shishir yet because it just came out, but a lot of this is from that conversation.

Shishir was the head of product and, basically, the CPO/CTO at YouTube—not right after the acquisition, but within a year. He came in for 4 or 5 years. After the acquisition, he said YouTube was doing about $30 million in revenue.

Okay.

They did have revenue. I believe, to foreshadow our next chapter, that it was in the form of programmatic advertising on the DoubleClick Ad Exchange that they were using to make money.

They were losing about $1 billion a year run rate on $30 million in revenue. The amount of money they lost was almost exactly equal to a penny per view. Just imagine that every time you loaded YouTube in those years, Google would flush a penny down the drain. They had to figure out something to do about this.

For the first couple of years, the CFO at the time was terrified of it scaling. He was saying, “Please don't scale in its current state.”

But of course, there's nothing they can do. The cat's out of the bag; it's scaling. And the CFO was exploring, “Hey, can we sell this to one of the other companies who was bidding on it?”

That's right. Because Yahoo and the media companies also wanted to buy YouTube.

Yes. So Shishir says, “We were broadly known as Google's first mistake.”

Well, back to my thesis up in the intro about being a pure play, investors didn't like this for a long time. This was a huge knock. I mean, gosh, when we did our episode 10 years ago about YouTube, we said it was a terrible acquisition.

Yes. The thing we haven't talked about: music licensing was really expensive. They were one of the top revenue sources for the music industry for a long time, maybe even still one of the top few to the music industry.

Yeah, right up there with Spotify.

Yep. So, on the product side of things early on, as you were saying, the way that you found YouTube was you would see it embedded on a different site. You would click through, and then you might stick around to watch something after, but then you'd leave, and your entry point to YouTube again was another embed. Most sessions did not start on YouTube.com.

So you weren't going to YouTube with the idea that they'd recommend something to you. And even the people who did go to YouTube.com in this 4- or 5-year period after the acquisition, 90% of that traffic was there to search, and they just ignored anything that you recommended to them. It takes a long time, A, to build habits, and B, to build out the technology to make any sort of recommendation or browse or anything good.

Yep. First with related videos and then, ultimately, the feed. And just for a sense of scale, there was a report that estimated that YouTube, that year in 2007, consumed as much bandwidth as the entire internet did in the year 2000—just 7 years before. I have an extremely similar stat from Shishir, which is from a later period. It's 2014, but it's apples to apples rather than comparing 2007 to 2000. He said in 2014, YouTube was 20% of the bits on the internet. Wow.

I mean, this stat, but especially your stat, illustrates just how much this thing took off and also just how much more bandwidth video took up than any other media type on the internet.

Yeah. But the long-term play here obviously is the money-maker slide: yes, video—the reason that it gets consumed so much is that this is what humans want—

Right?

And you can advertise against it.

And Google realized this. So I think they were very smart to, rather than trying to continue investing in Google Video, basically say: They got the lightning in a bottle, they have the consumer brand, they have the attention. Let's just go buy that thing. And on an expected-value basis, if you're making a bet, sure, you could build it on your own cheaper, but your chance of succeeding is so unlikely relative to buying that thing that it's actually a deal to get it for $1.65 billion, plus the billion that we'll need to invest every year for a few years to run it in the red.

So, 2009 is the year where the business really starts working. Google actually discloses nothing about profitability, but the ad revenue tripled in 1 year—in 2009. In 2010 and 2011, they turned profitable. There was a report that in 2012 they were estimated to make about $4 billion in revenue but roughly break even. Then, in 2012, 2013, and 2014, I think they were small—profitable, but profitable. Then, from 2013 to 2015, on the product side, that's when things really changed.

The north star really became: users should go to YouTube to be entertained for 15 minutes, and it's our job to do whatever we need to do to make that true. A few things really helped with this. One was the shift to mobile. In mobile—and remember, they were a launch partner on the iPhone—

Oh, we're going to get into it.

Okay. There were a lot more low-intent sessions. So, people who opened the app rather than clicking through from an embedded page.

Low intent meaning low intent to watch a specific thing.

Yes. It's a beautiful thing on mobile that you can sort of say, “I've got something to recommend to you.” And obviously short-form vertical video like TikTok and YouTube Shorts and all that these days is that on steroids. Mobile also made it the case that any given user was more likely to be logged in. That way, all the personalization, all the algorithm stuff works well.

They also adjusted their core metric internally away from views and to watch time. And YouTube was very early to the concept of creator monetization. For a long time, it was the only place on the internet where creators could make money.

They share revenue with creators.

And in our old episode, we sort of knocked them. We said, “Look, this business has to give its first 50% off the top of any revenue it makes to the creator of the video. That's a way worse business than, say, Google search ads or Facebook, where—you know—Facebook has influencers on their platforms too, all the Meta platforms, Instagram, and their rev share, if it's anything, sure isn't 50%. It's probably closer to zero.”

And YouTube, right early on, said, “You're a 50-ish% partner,” which takes you a decade longer to get profitable but helps you build that base.

It just creates amazing incentives for people to build businesses and careers on this. I mean, YouTube is the ultimate instantiation of the internet to me and the power that it can provide to individuals to make a living. It abstracts the need at all to create or run a business. It really just simplifies it down to: make content that people watch, and you will get money for it. You don't have to do anything else in between. There's a little sleight of hand that you did there, David, which is that people watch.

Well, yes.

So, YouTube internally went back and forth for years on this, and I think we're sort of in this no-man's-land that we've landed in today. Camp 1 is, “Hey, the way to make people most engaged is by getting them to follow creators, and they curate the information sources they want.” Camp 2 is, “In algorithms we trust.” It turns out Camp 2 is actually correct, which is unfortunate. It's a messed-up incentive. Most of the time, if you show someone something that they're subscribed to or you show someone something that the very smart computers have figured out, you will watch and then watch another video after that. Usually, the algorithmic approach is right. And so there is sort of this internal conflict there where they say, “Yeah, of course you should subscribe, but your views are only loosely related to how many subscribers you have.”

Yep. This is the dark side to the YouTube economy.

Yes. But putting that aside, just the sheer concept of anybody and everybody in the world who has a video camera today can create something, and if it's good and people watch it—and the definition of good being the algorithm likes it—you will make money—

With no other steps in between, that can only happen on the internet.

It's pretty interesting because it kind of has these 2 business models in core Google land. They have the AdWords business model, where they're the first-party media site. Each search result page is a form of media, and they run ads on that, and then they keep approximately 100% of the revenue generated from that ad. The advertiser pays them, and they share some in the form of traffic acquisition costs that we'll talk about later, but it's largely a first-party ad.

And then they have this other form, AdSense and the Google Content Network, when they show display ads on other people's websites, where they share like 70% of the revenue—most of the revenue—out to the—

The publisher.

The content owner.

Right. The content owner actually is the reason why there's an ad there in the first place. And YouTube was sort of an interesting mix between the two. They were comfortable and familiar with the idea that we can manage a platform where we actually share a lot of the revenue with those producing the content, which is interesting. Like, if they had never gone into the AdSense world and they were purely a search engine, I think it would have probably been more of a fight to try to do this 50% split with creators.

All right. So, there's a thing that I mentioned earlier: this notion of people on mobile are more likely to be logged in than people who just hit a web page on desktop. Logged-in-ness is essential for YouTube's success. That is actually new to Google. Logged-in-ness is not essential to the effectiveness of a search engine or even the monetization of a search engine. We've sort of flirted with this idea—you can kind of hear it through our episode—of things like Gmail are good because then you're logged into Google, but there's not a giant lift.

It's sort of like a nice-to-have.

Yes. Search, especially on the advertising side, is already so bottom-of-funnel—

Right? The intent is right there. I don't need to know what your demographics are. I know what your intent is—

Right? You search for a shovel, I'm going to sell you a shovel.

Yeah.

That's a stark contrast to YouTube, where the whole YouTube flywheel really only works with logged-in users—

Right? Not just for serving videos and content for you to watch, but also for advertising.

This is how the television advertising ecosystem works. It's about targeting. Why does Chevrolet advertise on football games? You need demographic data—

Right? All right. So, Ben, are you ready to regrade the acquisition of YouTube by Google?

Yes. We've got to set the context of how big it is today.

All right. So, back in 2016, we had 2 knocks on YouTube. One is that it wasn't a destination site. Narrative number 2 was they only get to keep 50% of their total gross revenue, and then they've got these crazy infrastructure costs, and they'll never be able to outrun them.

Well, they sure have solved both of these issues. Clearly, YouTube is a destination site. Actually, more than anything, it’s a destination app. You open YouTube and get an algorithm we trust.

It’s what I do every night before I go to bed.

Yes. And it’s some mix of things you’re subscribed to and things you’re not, but things YouTube believes will grip your attention at that moment.

On the infrastructure costs, let’s just start by unpacking their financials. Last year, in 2024, YouTube ads alone did $36 billion in revenue. Half goes to creators, so they have $18 billion left to play with, and they’ve had 2 decades to figure out how to get their variable costs down for video hosting, bandwidth, compute, encoding, music licensing, and all that stuff.

They now do insane feats of engineering, including their own custom silicon for video encoding. They also have a whole bunch of crazy things that they do, like changing the video encoding that’s used depending on how many views the episode has.

Interesting.

They do vanilla H.264 when you first upload it, and then when it hits some number of views, it switches to a format that’s more computationally expensive to encode but smaller to distribute. Then when you hit another threshold—when you have, like, 5 million views or something—they do it yet again. They re-encode the video and make the file size even smaller.

So they’ve figured out all these little optimizations to make any given stream as inexpensive as possible on a variable-cost basis.

Brilliant.

On the revenue side, they have gotten much better at selling ads, and most estimates are that YouTube is actually quite profitable. On top of the $36 billion in advertising revenue, Google reported that, if you include subscription revenue—things like YouTube Premium, YouTube Music, and NFL Sunday Ticket—they’re now doing over $50 billion in revenue.

Wow. So, David, this now makes YouTube the 2nd-largest media company by revenue, after only Disney.

And Disney has so many other things contributing to that revenue: theme parks, cruise ships, merchandise, et cetera.

Yes. So YouTube is already bigger than Disney’s media business, and this year will likely become bigger than Disney’s entire business.

The question is, how does that revenue figure compare to Netflix?

I’m glad you asked. Netflix’s annual revenue for 2024 was $39 billion, so they’ve already eclipsed Netflix.

Wow.

So there you go. Google doesn’t release usage data for YouTube, but I’m pretty sure that YouTube is the biggest single property on the internet in terms of minutes spent by humans on it. It’s not—

I believe that—

—the biggest number of users on the internet. Both the Facebook blue app and WhatsApp are bigger in terms of total number of users, but I think YouTube probably dwarfs them in terms of time spent by users on the app. I think it is the biggest. I could see that—

—the human attention-time sink known to man.

So then the question becomes: How profitable is the $50 billion in revenue? Officially, we don’t know, but there are these great things called research firms out there that make our jobs at Acquired much easier. Storied firm MoffettNathanson published a report earlier this year estimating that YouTube does about $8 billion in operating income—$8 billion a year. So I want to compare that against their total investment into—

Oh, yeah. Okay, great. This is the way to grade it. Okay, here we go.

Yes. Now we’re getting into grading here. We’re landing the YouTube plane.

The definitive Acquired regrading of YouTube.

So, as mentioned earlier, I don’t think they ever lost much more than $1 billion a year, and I think they got to break-even within, conservatively, 5 years. So after the $1.7 billion purchase price and the $5 billion in additional costs, Google paid $6.7 billion. I’d bet it’s closer to $5.5 billion to $6 billion to own something that spits off $8 billion a year in profit today, and revenues are growing 10% to 15% every year.

By the way, I also think the theme of this whole episode is the dual bottom line to Google of everything they’re doing: both revenue and profits and strategic insulation versus other large tech companies.

Oh, but David, you’re forgetting the 3rd: organizing the world’s information. Okay, the triple bottom line. There we go—the triple bottom line.

Google itself, not including YouTube, pretty much whiffed on social. Really, really strategically good for them that they own YouTube, isn’t it? Now that you know Meta and TikTok exist.

Well, here’s the crazy thing: They whiffed on social, and then what ended up happening—

Was social became YouTube.

Yes.

It’s the craziest thing. We don’t open apps anymore to look at what our friends are posting, a place where Google has no presence. But you open Meta’s most important property, Instagram, and you look at Instagram’s most-used thing, Reels, or you look at TikTok, and what do you see? You see videos from people you don’t know.

I mean, it’s crazy that the rest of social media, or almost like user-generated media, pivoted into Google’s space. This was the big denouement to our Meta episode last fall: Social networking, as the conception of it existed in the mid-2000s and 2010s, is dead. It’s gone. It bifurcated into private messaging and public media.

Yes. The sort of middle ground of a wide group of people you kind of know is effectively dead. It’s close friends, and it’s, “I don’t really care where it came from, but it’s entertaining.”

Yep. So you could do a discounted cash flow on this thing that I just gave you—call it $6 billion of investment and now $8 billion growing at 10% to 15% every year—but there’s additional strategic value, too. In addition to this thing you just said, this becoming the winner in the short-form era, they have the largest corpus of video to train on for the AI era.

Yeah, let’s go. MoffettNathanson estimated that, if this were publicly traded, it would be worth about $500 billion as a standalone company. Even conservatively, if you take media-company comps and do a revenue multiple and discount all the strategic future value, it’s still like $200 billion.

So this is officially one of the best acquisitions of all time, and I am raising my grade from a C to an A+.

I am obviously right there with you. This isn’t an A++.

Now, it’s not really fair to say that it’s like turning $6 billion into $500 billion. That initial $1.7 billion was largely Google stock that they traded, so that had real opportunity cost. But it’s still ridiculous.

Like I said earlier, a screaming deal either way.

Yes.

All right, there we go. We have revised history. Corrected the record.

Yes. All right. Well, for our next chapter, I motion that we go back closer to Google’s core business of advertising on the web—

—and maybe also stay closer to Acquired’s original raison d’être of discussing the greatest acquisitions of all time.

So, DoubleClick. Well, if buying YouTube in October 2006 for $1.65 billion was a lot, Google decided to basically double that a few months later, in April 2007, when they bought DoubleClick for $3.1 billion in cash this time, not stock.

And this is on the display-ad side of the house. So Google’s got 2 advertising businesses at this point. There’s AdWords when you search and get the ads that show up above the blue links, and then there are the off-property, or Google Network, ads. At this point in time, Google is just operating something called AdSense, which is this ad network that they’ve started.

Yeah.

So DoubleClick actually has a fascinating company history before Google that I did not know.

Yeah. Not a hot, rising startup like YouTube that they bought for a couple billion dollars—

Though it once was.

Yes.

All right. So, here's the DoubleClick story. And huge thank you: there's a new book that just came out by Ari Paparo. The book is called Yield. DoubleClick was originally founded in 1995.

So, before Google—

Before Google, the founders were Kevin O'Connor and Dwight Merriman, and their headquarters were in New York City. The original idea was twofold. One, build software that could let advertisers serve ads across websites. This is called an ad server. And two, build the network of websites and media where the advertisements would run. When people talk about paid media, it's the advertisements themselves that would run.

Over the next 5 years, they end up building and acquiring their way to being the leading display ad network and ad server.

And they went public during this time, right?

Yep. 1998. A shining success of the dot-com industry. However, the dot-com crash happens. 70% of DoubleClick's customers not only churn but go out of business. A huge number of DoubleClick's advertisers were actually VC-backed startups. Brand dollars hadn't really spread to the web yet. Like we talked about, digital advertising was so early and so nascent.

Yeah. It was Pets.com that was advertising on other dot-com properties.

Exactly. They're almost levered to the bubble, which is probably the right way to think about it. So, easy come, easy go. In 2002, after they're sort of limping along for a while, they sell that ad network division off for under $15 million—with an M.

Wow.

So now all they've got left is the software, the sort of ad-server part of the business. So fast-forward to 2004: they're this kind of sleepy, slow-growth company with a shrinking market cap. The ad server, their software, was still widely used, but digital marketing on the web just wasn't actually having that much spend flow through it.

They decided to put themselves up for sale. Google actually took a meeting to look at it to see if they wanted to buy it. They decided not to, and eventually they sold it to private equity. Two different firms, Hellman & Friedman and JMI Equity, bought it in 2005 for about $1 billion. The IPO-day valuation was double the final price tag at which they would sell it to private equity. And in many stories, this is kind of the end of the story. This is the start.

Yeah, it's sort of crazy, given the fact pattern that you just told us, that 2 years later Google's going to buy this thing for $3 billion.

Yes. So, David Rosenblatt becomes CEO, and he has a very familiar person on his team whom all of you will probably recognize: Neil Mohan.

Yes. The head of product and strategy at the company was Neil Mohan. Neil, of course, is the CEO of YouTube today.

Yep.

So, from DoubleClick originally—

Many would argue the best thing that Google got in the DoubleClick acquisition.

You could argue that. Now, here's the amazing thing. What happens under the private-equity ownership is that they launch a completely different product: this new thing called an ad exchange. This is when the concept of an ad exchange is first invented.

Remember, it was very straightforward before this. There was just an ad network and some software called an ad server. The ad exchange is this sort of brilliant idea that we can cross-route demand between ad networks. At first, what this is sort of used for is the remnant or unsold inventory. Oh, we've got some page loads. We don't currently have a buyer in our ad network for them, so throw them up on this exchange and see if, programmatically, some people will bid on it and we'll get more dollars this month for the same number of page views.

But technically, what was going on is it was really sophisticated, and it allowed for some crazy stuff to get done. You could bid in real time, including against the publisher's direct-sold ads. You know, let's say the New York Times has done a specific deal with Ford, then in a real-time basis—

Right? If somebody else is willing to top Ford, then you can displace them. Yep, yep.

Exactly.

Gosh, this sounds a lot like Google, doesn't it?

It really created modern programmatic display advertising, for better or for worse. That's basically what happened here. And as a publisher, when you start working with an ad exchange, you can incorporate multiple different networks, agency trading desks—because this became a big thing with ad agencies. You can stand up these complex rules engines.

Effectively, what happened is you sort of jumped in front of the ad networks. You almost disintermediated them. You're the lowest-level building block that everything else has to integrate with. And eventually, what started as this ad exchange that just became used for remnant and unsold ultimately becomes the primary way that digital media is bought by the biggest advertisers with the biggest publishers, and all, of course, bought and sold through these big agencies.

Google is running this little thing called AdSense. It's kind of for smaller publishers, and it's very DIY, self-serve. It's almost like a techie utopian's version of how to run ads on websites. Whereas this ad exchange is, let's acknowledge all the complex realities that exist in all these business relationships and all these purchasing decisions—the way Madison Avenue has evolved from the Mad Men era to this moment in time in the early 2000s.

And let's essentially construct fat pipes for money to flow—

Through all this. And what I mean by that is direct integrations into ad agencies' financial systems, and the ad agencies control the budgets for all the big brands and all the big dollars that are flowing.

That's exactly right. So, if only Google had a way of unlocking and now participating in these deeply integrated money flows. Google had a few other problems. The way DoubleClick worked, it performed a lot of really fancy stuff, like frequency capping to make sure you don't see the same ad 46 times. It used third-party cookies. Google was philosophically opposed to using third-party cookies, so they couldn't do stuff like that, but DoubleClick could.

Google didn't have a lot of these big sales relationships, since at the time, again, they're very obsessed with self-serve web pages. Advertisers just log in and upload and transact. So Google ends up kind of locked out of the best ad inventory. Advertisers on Google could really only be placed on the long tail of websites, which meant advertisers were willing to pay less to appear on those websites.

Again, we're all in the AdSense part of the world, not search ads. Yes, there are all sorts of things that make them not enterprise-grade here. So, Google decided they'd like to buy DoubleClick.

Yes. Well, that's sort of the story out there. The reality is, think back to how you started and when I interjected and I said, “Gosh, a lot of what DoubleClick is doing really sounds a lot like what Google is doing,” right?

We were talking to Tim Armstrong for research for this episode. Tim, of course, was head of sales at Google for many years, and we were asking Tim about DoubleClick. He was like, “Well, I was close with the DoubleClick guys, and I wanted to meet with them here in late 2006 or early 2007. It just so happened I was going to be in Seattle for some stuff, and I was emailing with them, and they were like, ‘Uh-oh, you're in Seattle? Actually, we're in Seattle too right now. We can get together here.’”

Tim immediately sounds the alarm inside Google to Eric and Larry.

They're in Seattle. Why are they in Seattle?

This is a New York-based company. There is only one reason why the DoubleClick guys are going to be in Seattle, and that would be if Microsoft is going to buy the company.

Yep. Now, back to everything we've been talking about all episode. What does Google absolutely not want to have happen here? Well, one was Microsoft kneecapping them by making changes to Internet Explorer or Windows or whatnot. They basically neutralized that through the whole web-app, Web 2.0 strategy.

Now the threat is, oh, Microsoft is finally going to wake up and do what they should have done 10 years ago and compete with us, build their own search engine—

Right? Be willing to be an ad-based business. Their DNA was, “Yeah, we'll do some ad stuff, and MSN kind of has to because it's a media business, but—”

We sell software.

We sell software. That's what we do primarily. And we would never trade our ability to sell software to make money on dirty ads.

Yep. Well—

Microsoft is realizing that for some set of users, Google's actually making more money on any given PC user than Microsoft is. And they're not happy with this. And they say, “Fine, we at least just need to be in that game too.” Yep.

So, the negotiations are kind of happening with Microsoft and DoubleClick. Tim told us this great anecdote where he's invited to present, and he still thinks it's an early-stage conversation in the negotiations. Somehow, he gets sent to the wrong floor. The person who is escorting him into the DoubleClick building sent him to a floor, and they sort of freak out when the door opens. They're trying to close the door, like, “Please go to the other floor.”

Tim is like, “What's going on here?” He steps out, runs down the hall, and sees a conference room full of all the Microsoft people and their accountants and their lawyers. And he's like, “Oh my God, oh my God, you guys, you're about to sign this deal with Microsoft. You've got to—”

So he gets them to hold off so he can kick the tires, do his diligence, and submit a counterbid. This is a crazy process that goes back and forth. Yahoo gets involved. There's a whole presentation series that happens where Yahoo, Microsoft, AOL, and Google are basically all getting the pitch, and DoubleClick is now...

I'm imagining they're all in an auditorium, and DoubleClick is presenting onstage.

Dude, there was a spreadsheet called YMAG.XLS. They created it to show, in each of these presentations, how much incremental money you'd make if you owned DoubleClick and tied it into your existing ad system. They were tweaking the numbers slightly for each one.

Google then submits its LOI for $3.1 billion, and it includes a clause saying they can't shop the deal around during this diligence period. The whole Google team goes to New York and rents out this big room at a hotel near DoubleClick's offices.

I'm going to read an excerpt from the book Yield:

The company's counsel—this is DoubleClick—checked her BlackBerry and held it up for David Rosenblatt to see. There was an incoming message from Microsoft's corporate development team. They were willing to match the offer for DoubleClick, and the message included an email from Steve Ballmer saying that he had opened the door for a much higher offer.

Ballmer wrote that if the offer match was not acceptable, DoubleClick should simply mark up the paper to meet its needs and then sign it, and Microsoft would review and rapidly countersign to close it with minimal negotiation required. Without saying so, Ballmer was communicating, “Here's a blank check. Tell me what closes the deal.”

Ultimately, a week goes by, and they're in this period where they can't really respond. They're supposed to just proceed with Google. A day before the LOI is set to expire, the DoubleClick team gets an updated term sheet from Google.

The financial terms of the deal are unchanged at $3.1 billion, but now the deal includes what they call a hell-or-high-water clause, which means that Google is committed to closing the deal without any substantive diligence or any other conditions. It's just money in the bank.

So there's no more diligence. DoubleClick just signs it: $3.1 billion. The private equity firm turns that $1 billion, which was leveraged—it was something like $300 million of equity and $700 million of debt—into a $3.1 billion sale to Google. Then it's over.

Not bad work if you can get it.

Nope.

So this was huge for Google. DoubleClick bringing it into Google really did help with those fat money pipes, as I was saying, of dollar flows from ad agencies.

Yep. But the biggest thing was that DoubleClick was the number-one player in the space. There was another public company called aQuantive that was the number-two player.

Which Microsoft then bought for twice as much. They were like, “We really wish we had gotten DoubleClick.” And then, within months—

It was the next month. Right away, Microsoft turned around and bought aQuantive for $6 billion, so twice the price. But Microsoft getting the number-two player versus the number-one player slowed them down.

We're heading right into Microsoft's search efforts with Yahoo and then, ultimately, Bing. Getting into the advertising business was worth every penny to Google, even for the sole reason of keeping the premier number-one player in the display ad space out of Microsoft's hands.

Yep. Okay. Hey, the one thing I will say here, David, is that unlike all those other Google products—Maps, Gmail, YouTube, organizing the world's information—this is not organizing the world's information and making it universally accessible. This is, “We're running an ad business, and we want to expand the ad business, so we're going to expand and protect our business interest by buying this.” It's a chess piece on the table, where having it in our hands versus other people's hands is better.

That's exactly right. There are ways that it systematically advantages you to own the exchange when you also own the network. This is only checking the box of strengthening our business without checking any of the other boxes.

You basically never heard Google executives get up onstage, inspiring people about the future of the company, and talk about basically anything DoubleClick was doing.

Yes, correct. And even fast-forward to today, unlike YouTube, it's not like this has become a world-dominating thing.

Right? If you're in the display ads world or you're a publisher, this feels like a huge deal. If you're Google, let's just look at the numbers today. Google in total in 2024 made $350 billion in revenue. About $200 billion of that is from Google Search. About $30 billion of that is from Google Network. This falls under Google Network.

Plenty of which existed before and would have existed anyway in AdSense, regardless.

I don't know about that. I don't know that Google would have become the dominant player in display ads—

Absent DoubleClick—

Without buying DoubleClick. Yeah, I don't think—

Yeah, yeah. But AdSense was probably doing $1 billion-plus in revenue at the time and would have kept scaling. So all that to say, in the context of Google, this isn't a YouTube.

It just doesn't matter that much. They make $200 billion in revenue from Search, where they get to keep 90%-ish of that after paying out traffic acquisition costs. With Google Network, they pay out 70%, and they only made $30 billion.

So if you start thinking about gross profit, it's comparing $9 billion to $175-ish billion.

Yep.

It's just not that consequential to the story.

All right. So, speaking of Search, catch us up on how the search business is doing during these years and why Microsoft finally said, “Okay, enough. We've got to enter this business ourselves.”

Yeah. We've been talking about the sideshows, like trying to add wind to the sails of the web, while Search is cranking on improvements to the core product and revenue is growing right along with it.

Here's a little timeline to catch us up from 2003 to 2008. They start updating the index more often. The index starts to feel not quite real-time, but it used to be that when you searched, you'd be getting results that were indexed three months ago. Now the web is feeling a little bit more—

Real-time-y.

Recent when you're searching it.

They launched Google Images, Google News, Google Books, and Google Scholar. They launched Google Suggest, which is when it starts autocompleting your searches. Later, they would launch something called Google Instant, which was very cool at the time. It's actually kind of gone away now.

Google Instant would run a completely new search based on every character you typed and show you the results page updated in real time with each next keystroke, which was pretty amazing.

I remember that being so cool when it launched.

Yeah. In 2005, they incorporated your search history into your results. This is when they started doing some personalization stuff with logged-in users.

In 2004, they had $3 billion in revenue. In 2005, they had $6 billion in revenue, so they doubled even at that scale. In 2007, they launched Universal Search across web, images, video, whatever—maps.

They tried to deconstruct your query and understand which of these things you were looking for. They used to basically build a completely separate search engine for each media type and leave it up to you to decide which thing to search.

That year, when they launched Universal Search, they did $16.5 billion in revenue. In 2007, this is when they became the largest seller of advertisements in the world—not just digital ads, ads. Digital ads would not overtake traditional media until 2018, as we talked about earlier.

Yeah, I was trying to square this. I guess that means the market share Google has of digital ads is so massive that it's bigger than even in the traditional space or the TV space what any one player has.

Yes. So every year for the last 18 years, Google has been the number-one seller of advertising of any kind in the world.

Wow. This, I think, helps you understand a little bit what's at stake in the era of AI. This is literally the trillion- or $5 trillion- or $10 trillion-dollar question: Can Google keep being the number-one seller of advertising in the world even through this sea change?

We should do a whole episode on that, probably. Maybe we'll do it next time.

But actually, there are some great corollaries with the mobile wave that we're about to talk about. To pull forward a few more search improvements that they would do later: In 2009, that's when they really did some real-time indexing of the web. In 2012, they launched the Knowledge Graph, so when you search about basically a thing with a Wikipedia page, you always get the kind of snapshot view on the right-hand side of that entity.

All along the way, they're tweaking the algorithm in an attempt to reduce spam. That's effectively the product changes on the people side of things.

They had really solidified themselves as the preeminent computer science research company at this point. If you were to refer in 2008 to a really smart programmer, you probably said, “Oh, they're like a Google-type engineer.”

They sort of took the mantle from Microsoft and had not yet relinquished it to Facebook, or later to Stripe, OpenAI, Anthropic, or any of the companies we would talk about in the future as this dense concentration of the best engineers.

They had pulled in a lot of the people from the big research labs that had been collapsing. So you had Jeff Dean and Sanjay Ghemawat coming from DEC. Ben, we did Sanjay a total disservice on the last episode. A lot of the stuff that Jeff did—and, of course, he became sort of a Google executive—Jeff and Sanjay pair-programmed together.

Yes, there's an amazing New Yorker article that was published long ago about their friendship and career partnership and everything that they accomplished together.

Yeah, we'll link to it in the show notes.

And basically, if you look at any big research paper about giant Google infrastructure stuff that was launched from, I don't know, 2002-ish, maybe even earlier, through the 2010s, Jeff and Sanjay are either the 2 authors or 2 of the 5 authors. It's amazing how much stuff these 2 guys invented.

They also got Bill Coughran and Rob Pike from Bell Labs. You had Xerox PARC and IBM's labs sort of losing prominence, and Google was just sucking in all these generational, heavy-hitter computer science, architecture, and systems programmers from all of those places. That's sort of how I would describe where a lot of the technical breakthroughs were really coming from, or at least the culture of technical breakthroughs.

We talked about these incredible products and incredible innovations, the development of the whole concept of a web application, but that was coming from these people who were coming into the company and who were, as you say, generational talents. Speaking of that, it was very convenient for a couple of things that they needed to start doing in 2008, namely launching their own web browser and then, shortly thereafter, launching their own mobile operating system.

It's astonishing that they did both of these things in the same year.

And this isn't like, “Oh, I'm going to start a browser,” the way that you can start a browser today. I mean, all these AI companies are launching browsers.

Yeah, they're using Chromium.

Right? This is a giant engineering undertaking. You need amazing architects. This is equivalent to Dave Cutler doing Windows NT. It was earth-shattering when Google launched Chrome.

Or everything Jeff Dean and Sanjay did in the early days of Google.

Yes. And when I say that people launching web browsers today are using Chromium, Chromium, of course, is the open-source version of Chrome.

Yeah.

That Google just gives away for free to anybody.

Yeah.

So, in February 2008, the shoe that Google had been fearing would drop for many, many years finally does drop: Microsoft is officially going to enter the search business. Microsoft makes a bid to buy Yahoo for $44 billion. The giant has finally woken up.

Fortunately for Google, they get a little bit of a reprieve because Jerry Yang turns it down.

So dumb.

It was one of the worst corporate decisions of all time because just 2 years later, after Microsoft launched Bing the next year, in June 2009, Bing would take over powering Yahoo Search in a deal that paid Yahoo $1 billion, versus the $44 billion that Microsoft had been willing to pay for the whole company just 2 years earlier. Then Yahoo would sell itself to Verizon for, like, $3 billion, I think—something like that.

Single-digit billions.

Yes. Google, though, knew this day was going to come eventually. Fortunately, by this time in 2008, Google had its competitive response all ready to go, which was Chrome.

And they had actually been working on improving the state of browsers for years.

Oh, yes, they had. The story of Chrome goes all the way back to 2001. Larry and Sergey wanted to build a web browser in 2001 for this very reason that we've been talking about the whole episode. All of Google rested on Internet Explorer.

I didn't realize that.

And also, I mean, it was Larry and Sergey. Of course they wanted to build a web browser. It's the most Google thing. Why wouldn't we build our own web browser?

Right?

But it was Eric who said in 2001, “No, we can't do this now. We can't poke the bear right now. Google is too young and too vulnerable.”

Calls like this are why Larry and Sergey brought in Eric.

Eric. Yes. The actual quote from Eric at the time—this is in In the Plex—is, “I don't want to moon the giant in 2001.”

It's a very Eric Schmidt quote.

But that doesn't mean that Google isn't preparing for this. Instead, what they do is decide that they are going to become the primary major benefactor for the new Mozilla Foundation and what would become Firefox. Mozilla was the nonprofit organization that was founded and spun off from Netscape when AOL bought Netscape.

Are they a funder? Are they actually just giving money?

Well, I think at first it probably was grants, giving money, because this was strategically important for Google. But then, once Firefox actually gets released by Mozilla and deployed out there, the way Google starts supporting Mozilla and Firefox is by paying traffic acquisition costs to them to be the default search engine in the Firefox browser. Spoiler alert: it's like what they do to Apple for Safari today, to the tune of, like, $20 billion a year.

Yes.

Firefox is where this all starts with Mozilla.

Right?

Actually, traffic acquisition costs originated before Mozilla because it's effectively the same thing that they were doing with software vendors to include Google Toolbar.

And so I think the mechanism of payment over time shifted to more of a rev share. My understanding now is that they share some of the revenue they generate from queries that originate—

Searches that happen in the browser.

Exactly. Which is why it ends up being kind of variable year to year. But yes, Google has a long history of paying for distribution of its search engine, and the new form that it is now taking is the Mozilla Firefox browser.

Yep. Starting with Toolbar.

And by the way, I should say Google becomes a giant contributor of source code to Firefox.

Well—

Oh, is that where you're going?

I'm going to get into this. So, for a couple of years, Google is paying Mozilla for default search in Firefox, basically funding Mozilla. After a little while, Google decides that it's going to hire some of the key Firefox engineers from Mozilla to come and work at Google directly. But they position this as, essentially, the same thing: “We are still funding Mozilla and Firefox. You're Mozilla; you can't give these employees, these engineers, stock options because you're a nonprofit. How about instead, they do the same thing that they're doing, which is working on Firefox? They'll just come and work here at Google, and we'll pay their salaries and they'll get Google stock options. Otherwise, they're going to get poached by all these tech companies.”

Fascinating.

You can see how this makes sense.

This team that comes over from Mozilla into Google becomes the core of a new, quote-unquote, product client group within Google.

Meaning products on clients, i.e., installed applications on PCs, not the web apps that the rest of new Google is doing.

And the leader of this group—Google hires him from McKinsey in 2004—is Sundar Pichai.

Oh, I did not realize that's where Sundar came from.

Yep. Well, again, all of this is very strategic because if you did someday want to build your own web browser—

Now you've got the bench of talent. They're employees.

So, there's a quote from Eric Schmidt in In the Plex: “This was very clever on Larry and Sergey's part,” because, of course, these people doing Firefox are perfectly capable of going and doing another great web browser. This group is sitting there within Google for a couple of years, almost like a latent sleeper cell within Google. They're just ready to activate as soon as the Microsoft threat becomes real.

And the way I heard it was that a lot of people were working on Google Gears, which is this browser extension that allows for offline functionality. They'd built the Google Web Toolkit to make web application development even more advanced and sophisticated. At some point, they kind of lost faith that Firefox was going to keep pace and stay as high-quality a browser as they needed it to be.

They also had some divergent technical ideas, like different architecture ideas for how a browser should function, that we'll talk about. I think all of these things are true. However, having your own browser when Microsoft launches Bing is hugely, hugely important.

Imagine if 90% of Google happened on Internet Explorer and, all of a sudden, Microsoft launches Bing.

Right?

There's no amount of money you could suddenly pay Microsoft where they would keep you as the default search engine, because they would just want all the traffic to go to Bing. They now have a great way to monetize.

Absolutely not. Bing as the default search engine—done, right?

And the thing that Microsoft would fail to realize with Bing is that you can't be second place in search.

Right?

The most liquid auction will always win, and Google has already run away with the search ads auction liquidity. Traffic on Google searches will forever be worth more than traffic on the second-place search engine.

Sure. It doesn't mean that the battle wouldn't be hugely damaging to Google if they didn't have their own web browser.

Yep.

And also, it's clear that web apps, JavaScript, and Ajax were very important, and Internet Explorer wasn't keeping up with the technology.

Yeah. So, there are 2 killer features—arguably, maybe 3—that they're going to bake into the Google browser.

Oh, I've got 6.

Oh.

So, I'm curious which ones you don't think are important.

Okay, I'll go through my 3, and then I'll see what else you have to add.

Okay, great.

Number 1, most important: it is going to have a super-fast, super-modern, super-performant JavaScript virtual machine called V8.

Yep.

That is going to run big web apps fast and stably.

We are the Ajax company. That's right. Two, web apps crashed a lot back in the day. They don't so much anymore, but they used to crash a lot.

Larry Page has this quote from when they're deciding that they should roll out Chrome, and he explains: “We have found the web-based service delivery model to have significant advantages.” You don't say. “But it also comes with its own set of challenges, primarily related to web browsers, which can be slow, unreliable, and unable to function offline.”

There you go. And so before Chrome, this is impossible to remember now, but if you had a tab or a window open and running a web app, and that web app crashed, it took down your whole browser.

Yep.

Everything that you had open was gone.

Tabs were not their own processes.

Nope. So each tab is going to be a separate process on your machine. If the web app running in one tab crashes, all it takes down is that one tab. And it made sense that before this they weren't their own processes because, one, tabs were kind of a new thing. But two, web applications were websites. The notion of web applications was only really 4-ish years old.

So those are my big two. I suspect one of yours is WebKit. I'm not including WebKit here because that was an Apple innovation.

Yeah, that they borrowed. I'll let you talk about that in a second.

I don't have anything more to say on that. It was the best rendering engine.

Yes. So let's say that's three. And then my sort of 3½ is the design. So, of course, the web browser ultimately comes to be called Chrome, which is ironic. Chrome is a reference to all the stuff in a web browser—the toolbars, the nav bar, et cetera—that take up space around the content. The idea with Chrome and the Google web browser is minimal Chrome, as little as possible. It's just about the content. Let the web and the web apps shine.

Yes. Okay, when you said UI, I thought you were going to say this. My fifth is the omnibox.

Ah, yes.

Originally, there was just the URL bar, and then when search became the killer app of the web, there was a second little input box for search on the right side. So we had that awkward teenage period where browsers had the URL bar on the left and then search on the right. And it's kind of clean to think about it on its own, because now that we understand that it was sponsored—I think for the longest time that was not in the public psyche—that whatever search engine appeared in that box in the right-hand corner was paying for that placement.

That was nice because you type in the URL bar and that's your organic typing. And then the other one is, “I'm willing to give a kickback to Google, probably, but it could also be Bing, could be Yahoo, could be whoever.” Google correctly understood that, from a user-experience perspective—but also, just thinking about their core business model—the right design for web browsers is that if you don't type in a URL, it should just search.

Just one bar. Why have two bars?

We imagine that generating a whole bunch more page views on search results pages creates a whole bunch more opportunities for our advertisers to reach your eyeballs. But I will say they were also correct from a user-experience perspective.

The fact that URLs ever leaked to the public is a mistake. That's letting an implementation detail of the technology—

It's an accident of history that consumers type—

HTTPS. Are you kidding me? Consumers never should have known the phrase HTTP.

They should just type “New York Times.”

Yes, which AOL tried to do. AOL Keywords.

Yeah.

So this is effectively leaning into that idea. You can use this box for typing in URLs, but really, what you use this box for is kicking off the Google search. So, brilliantly aligned with our business model.

All right. So that's what I got. What else do you have that's not on my list?

That's five. And then lastly, sandboxing. Each tab is a sandboxed environment. This prevented a ton of malware. This was a big breakthrough in computer security, where anything that was operating in that tab was in its own sandbox and couldn't be accessed maliciously.

Yeah, I guess, remembering back, before Chrome and modern web browsers, browsing the web was a security threat to your PC, right?

Yep, that's exactly right.

Great point. Okay, so they start work officially on this in 2006. They launch it in early September 2008, like a week before Lehman goes down. This is wild.

I remember that because I remember sitting in North Carolina at my Cisco office, and I remember reading the Chrome comic, which I actually just read a couple nights ago for this episode—this amazing web comic—at the same desk where I read the news about the great financial crisis and the world falling apart and Lehman Brothers collapsing.

Wow. So they launch it in early September 2008, and the way this is so Google: they hire the famous comic artist Scott McCloud to illustrate a digital comic book as an introduction to Chrome, explaining what it is, like a user manual in comic-book form. It's written for this weird half-user who's kind of technical, but you don't need to be a programmer necessarily. It's written for the tech enthusiast who can understand process independence, understand sandboxing, understand V8 and the JavaScript speedup, but it's not written for the general public.

Yep. But that was exactly the right seed-crystal user base to get Chrome—

Yeah—

—into.

Yeah. It's written for the Slashdot reader: the kind of people who are going to go home for Thanksgiving in a couple months and install it on all their family's computers and say, “You need to stop using Internet Explorer right now for all of these reasons, Ben, that you just listed.” Probably security being number one among them.

This actually was me back in the day. I was going to go home and install Chrome on my parents' computers so that they didn't get hacked and lose their financial information, et cetera.

Yep. Within 18 months, they got 40 million users. Then, let's see, they launched it in 2008. By 2010, they had 70 million users. Then, in 2012, they had 200 million users. Actually, what happened is it destroyed Firefox's market share. I think the launch of Chrome and the peak of Firefox were right around the same time. And then, after that, it really started eating away at Internet Explorer's market share.

And today, aside from mostly iPhones and Apple devices running Safari, it is the browser. To say it worked is the understatement of the century. It totally liberates Google from Internet Explorer and Microsoft.

When Chrome launched in 2008, Internet Explorer had almost 70% market share of browsers, and Firefox had most of the rest. Two years later, like you said, Chrome had passed 100 million users. By 2012, so 4 years after launch, Chrome and Internet Explorer were now tied for market share, with about 30% each. So Internet Explorer had gone from 70% down to 30%.

And this is both Chrome on the desktop side, but you're now also well into the rise of mobile. Apple's mobile Safari is becoming huge, and Google's Android, which we're going to talk about in a second, is becoming huge. Two years after that, in 2014, Chrome was now the clear leader with 40% market share. Internet Explorer was down to 15%.

It's over.

It's over.

And Internet Explorer is basically dead at this point.

2013, 2014.

Yeah, 2013, 2014. Today, it's not even close. Chrome has almost 70% market share, according to StatCounter, including iPhones, which all run Safari by default.

Right. Safari in aggregate across mobile and desktop—mobile is by far the biggest share of Safari's market share—is about 20%.

So Chrome has 70%, Safari has 20%, and there's 10% left. Microsoft, I don't know, has a couple of single-digit percentage points. Talk about flipping the tables. Chrome was massive.

And it was just better. It was so much better. And it really kicked off this amazing era for the web, with Apple needing to play catch-up and leapfrog. In a lot of years, it was actually faster than Chrome, and they would go back and forth. And it really spurred Apple, who was already a steward of WebKit. They had their own competitive response to Microsoft after Steve Jobs hated the fact that he had to keep shipping Internet Explorer as his best option on the Mac—

Because that was part of the Microsoft-Apple deal, right? When Microsoft saved Apple with the investment, Internet Explorer would become the default, right—

—on the Mac.

So Safari was created for that. But over the years, Apple's incentives, especially post-iPhone, were not to make it so web apps could be great. Apple's incentives were to make it so native mobile and desktop applications could be great. And so Google really pushing the envelope on the web's capabilities and what a modern browser could do forced this good-for-the-world race between Apple and Google to both make better browsers.

I don't think it is an exaggeration to say that Chrome kept the web alive—

Yeah.

—as a viable platform for applications.

Yep. I mean, Microsoft certainly didn't have an incentive to do it in the business model they were in at the time. And Apple doesn't.

They had every incentive not to—

Right?

I mean, who in the world least wants the web to be a viable application platform? Microsoft—

Right? At that point in time. Apple now, ironically. So there's one more amazing, delicious part of the Chrome story: do you remember Google Chrome Frame? Google Chrome Frame was a plugin for Internet Explorer that replaced IE's JavaScript engine.

Oh wow.

It pulled in Chrome’s V8 JavaScript engine and, I think, also WebKit for rendering. And so, for all the corporate users of America and the world who couldn’t install a new app, who were stuck with Internet Explorer, this was the only reason that IE hung on to market share for so long. It was just locked-down PCs. For all those poor souls, Google was there for you with the Google Chrome Frame plugin that let you run Chrome-quality web apps within Internet Explorer. Amazing.

All right, so I have a question for you on Chrome before we finish the story. Why make Chromium open source? The answer that I’ve read to that is mostly about the Google culture and trying not to be too evil about it.

Yeah, I would buy that.

Yeah, they’re like an open-source company. It’s in their bones to contribute to open source. There’s a thin business reason I can think of for why they would want to make it open source. The reason I can think of is that it doesn’t need to be closed source. They make their money from searches, right? It can’t hurt. Here’s the way it could help.

At first, I was thinking, well, wait, Google wants to own as many of the browsers that its searches originate from as they can, so they don’t have to pay out distribution costs in the form of traffic acquisition. So if someone takes Chromium and then builds a better browser, it’s kind of bad because now you have to pay that browser maker. But in practice, as long as it’s not Microsoft and as long as it’s more fragmented, that’s probably a trade they’re perfectly happy with. If they need to go and split some rev share and pay someone who makes some variant of Chrome based on Chromium, and that gets really big and it gets 30% of the market, great. Google’s delighted because it’s not Microsoft.

I mean, look, except for traffic acquisition cost, Google’s incremental gross margin on search revenue is like 98% or something like that. They’re still going to be an 87% gross-margin business.

Because the whole point of this was to prevent an existential risk. And so if they have to do some light rev sharing, even in their worst-case scenario where someone builds a successful thing based on their open-source project—oh, so horrible. We go down to 87% gross margin.

Right. Right. It’s still perfectly acceptable, which actually may be the way it plays out if any of these new AI browsers work out.

Yeah, maybe.

Well, I’m not sure any of these AI browsers would be willing to let Google pay them to be the default.

Big chess game to consider there.

Yes, that is the one catch: the owner of the browser still has to be willing to accept the payment from Google. I have one analogy for all of this that is a little far afield, but I think is actually the right way to think about this.

Okay. Lay it on me.

Walt Disney creates Disneyland. It goes great. It’s a very handcrafted, curated, aura-driven thing, but ultimately he has to play within the rules of things like city government. Okay, they go big and get a huge plot of land in Florida to build Walt Disney World. It would be nice if we controlled our underlying foundation a little bit more. So they build their own government district around the park, and they say, “We make the rules here.”

Right? It’s not totally dissimilar.

Yep. I’ve been reflecting a little bit on why, basically from 1998 onward, Google’s biggest threat was Microsoft—not because of Bing, not because of building advertising, but because of this kind of destabilizing thing. There’s sort of a fine point on it, which is that spiritually Microsoft was the platform of the PC era. And with this platform shift, it would be very convenient to just be like, “Oh, Google’s the platform of the web era.” But even though Google is the platform company of the web era, they aren’t necessarily the ones building the platform, right?

Yeah. Yeah. They still existed at Microsoft’s pleasure.

And no one owns the web as a platform. So there’s this kind of funky thing where Microsoft built and owned Windows and then dominated the PC era because of that. Google operates a search engine that generates advertising revenue. They don’t charge anyone for anything. They benefit from the web’s growth, and so they’re doing this strange indirection.

Yeah. It’s an ecosystem-building exercise.

Exactly. They’re trying to build the ecosystem. They’re trying to be the steward of the open web as a platform. And they put their finger on the scale where they need to and take a little bit more control and ownership, like with Chrome or like with some of these standards bodies, to push the web forward and make sure that the place where they live—their neighborhood, the web—is in good shape. But it’s not their platform in the way that it was Microsoft’s platform in that era.

All right. Maybe the Disney World analogy is even better than we gave it credit for a minute ago.

That’s sort of—I don’t know. It’s a little bit loose, but that’s what I’ve been kicking around.

I like it. I like it. Well, no doubt Chrome was a huge success. Hell, Sundar becomes the CEO of the company. No better sign of how successful it was than that, right? It shored up their future. We could do a classic Acquired “what would have happened otherwise” just for a minute or two here. What if Google didn’t launch Chrome? Let’s say Bing launches in 2009 and there is no Chrome, and Microsoft still has 70% share and will for a while.

Yep.

And they make Bing the default. Let’s see. Mobile would get big 3 or 4 years later, in the 2012-ish time frame.

Yep. Still small.

So they basically would have had 4 years of 70% of people using browsers on any device being defaulted to Bing. Now, obviously, a lot of people would still want Google. They were used to it. They’d switch back. But I don’t know, man. Defaults are powerful.

Defaults are powerful. I think you’re right.

This is why Google pays Apple $20 billion a year.

Yeah. To your point, maybe without Chrome, Bing would have been a serious competitor to Google. There is no more important distribution point for search than the web browser.

It is the way to monetize a browser. Basically, the single way to monetize a browser.

Yep.

Which, let’s make this relevant to today and stop dancing. If the DOJ’s ruling is that Google has to divest Chrome, there is one way that Chrome is a business, and that is getting paid by Google to drive traffic to Google as a search engine. It’s the only way to operate a business of a browser. Maybe in the AI era, it’s the AI company having it drive traffic, but it’s the same exact thing.

And so, one of two things has to be true. Google owns Chrome, or someone else owns Chrome and then Google pays them. But the thing definitely can’t be made illegal—or I don’t really understand what the goal is if you make it illegal—if you say they both can’t own Chrome and they can’t pay web browsers to drive traffic. Chrome has no potential to be a business if that’s the case.

Yep. So Chrome

huge win. In fact, it’s so much of a win that after a couple of years, Google starts thinking, well, gosh, maybe we should build Chrome into an operating system in and of itself. Let’s go attack Windows. Let’s take it to Microsoft where it really hurts: Chrome OS. It became successful in schools and education.

Yeah, Chromebooks. Chromebooks have major market share there, but it’s not a major player in the overall PC operating system market. It is wild how much the PC computer operating system market is still dominated by Windows. That has never changed. You and I live in this world where everybody uses a Mac. Mac has like 15% market share. Windows has like 70% market share of computer operating systems.

Yeah. Well, speaking of operating systems, I think it’s probably time to talk about Google’s big one, which is actually the biggest operating system in the world.

Over 3 billion active Android devices now.

Totally freaking wild. They bought it for $50 million.

Well, that’s a red herring. They’ve invested so much more. But just to make the point, it is hit after hit after hit. These things are not predicated on Google’s distribution. If you’re a company that launches a new widget and you can just distribute it with your old widget, it’s not that impressive when your new widget gets dominance.

But Google Chrome—I mean, they could do a little thing, and they did push it on Google search pages, but they managed to get a lot of distribution just by being a great product on the market, with viral adoption that everyone told their friends to use. And it was the David Rosenthals going home to Thanksgiving that were sort of the seed of it. And then within 3 or 4 years, they just ran the table. And it’s not just Chrome. Gmail was that way. Google Docs and spreadsheets were that way. I mean, that way, it’s everything—everything.

All these things are independent, great products that became dominant on their own merits, just like Google Search did.

Yes, I 100% agree, and also helped by the fact that they were all free.

Yes. Yes. Fair. And massively subsidized, at least in the early years before they were able to be businesses on their own, by the old money-printing machine in the basement of Google. Good old Uncle AdWords.

Yes.

All right, Android.

So Google's office spaces are legendary. The first one, of course, was Susan Wojcicki's garage in Mountain View, the company's first office. And then today, the Googleplex, the old SGI—Silicon Graphics—campus in Mountain View. In between, Google had another office for a couple of years in downtown Palo Alto, at 165 University Avenue, which would later also be the office where PayPal was started.

Oh, really?

Very lucky building. Yeah, in August 1999, when Google moved out of that office, do you know who moved in?

Based on the direction this is going, is it Danger?

Yes, it is.

Yes, Danger.

The company started by Andy Rubin. Andy, of course, had been an engineer at Apple and then left Apple with a group of rebels—I don't know, were they rebels?—who went to start General Magic. General Magic, of course, was a legendary failed startup in Silicon Valley in the early '90s, basically trying to create the iPhone 15 years too early.

After General Magic fell apart, he started Danger. Now, Andy's initial idea for Danger was that he wanted to make a wireless version of the CueCat scanner.

What is a CueCat scanner?

This was a device that plugged into your computer, looked like a cat, and scanned barcodes. Andy's idea was, “Okay, all this General Magic stuff we were trying to do was too far ahead. What if we think simpler and just make a wireless version of this to scan barcodes?” Not a big idea.

His first employee at Danger, a guy named Hiroshi Lockheimer, convinced him, “Hey, actually, a couple of years have gone by. Maybe we should revisit this General Magic stuff.”

Wait, Hiroshi was with him at Danger?

He was the first employee.

I did not know that.

Yep. Yes, he was.

I mean, he, of course, is instrumental in the Android story later. I did not realize the two of them were at Danger together, too.

I spoke to Hiroshi in researching this. Great, great guy. Hiroshi led Android and Chrome at Google for many, many years and would be the authority on this. So Hiroshi was like, “Hey, hey, maybe let's revisit this General Magic stuff.” And that led to Danger building the Sidekick and launching it in partnership with T-Mobile.

This thing was amazing.

That thing was so sick. I was jealous of all my friends who had one.

It was a messaging-focused, rich-application cell phone. I think it was, along with BlackBerry, the first vision of a cell phone where the primary thing you do on it is not talk to somebody; it's messaging. These things were freaking awesome. They were really big with celebrities. I think it was the plot of an Entourage episode at some point in time.

So they end up selling this company to Microsoft, right?

Yes. Microsoft did end up acquiring the company, but not until 2008, which is the same year that Android launched. Andy actually left Danger in 2003 and started a new company, Android, which in the earliest days was kind of like an open-source competitor to effectively BlackBerry software.

Yes. In its earliest, earliest days, the first version of Android—the company, remember I was talking about point-and-shoot cameras and digital cameras back in the YouTube section—was actually to build a cross-platform, open-source operating system for point-and-shoot digital cameras.

Oh, wow.

Yeah. That was Andy's vision: “Oh, hey, these point-and-shoot devices—hundreds of millions of consumers have them now. What if there were a powerful operating system? Could that be a Trojan horse to get an operating system?” You could sort of imagine it.

Right. If cameras became phones instead of phones becoming cameras, then yes.

Yep. Exactly. But pretty quickly, it became clear that phones were going to become cameras. The good thing, though, was that the software they were writing still worked just as well on phones. So Andy pivoted the company and shifted the delivery vector from cameras to smartphones.

At the time, the smartphone market, such as it existed—and it did exist—had a few players.

BlackBerry, Windows Mobile.

Well, yeah. So here were the players. Basically, phone companies were either full-stack, like Apple and the iPhone is today, where they made the phone and the operating system. That was Nokia. And then the big player in the smartphone market, at least, was BlackBerry. It made its own software and its own devices and was huge in the enterprise market.

Or you had OEMs, device makers who made devices and then bought an operating system, either from Palm—which made its own devices but then also started selling the operating system to other vendors—or the big player, Microsoft, with Windows Mobile. This was a licensing model, as we talked about in our Microsoft Part II. You paid Microsoft single-digit dollars, got an operating system, and then built the phone stuff on top of the operating system.

Exactly. And this was a good business for Microsoft. Obviously, it wasn't as big as the desktop market, but you can totally understand why this was their strategy. “We are the main desktop operating system provider. This is our business model there. Let's just do the same thing here.” It seemed to be working.

Yeah. And as far as the phone manufacturers, the OEMs, and the carriers were concerned, things were also pretty good. These phones that they were making couldn't really do that much, but because of that, they didn't actually cost that much to make. Meanwhile, consumers were paying through the nose for these things, because with a smartphone on a carrier contract, you were paying like $100 a month.

And they didn't consume that much data, either, because they weren't that capable.

Everybody was fat and happy.

Yeah.

So into this morass—which Steve Jobs was, of course, also looking at and saying, “This sucks”—entered Andy Rubin and Android. He went around pitching the phone manufacturers and the carriers: “Hey, stop buying an operating system from Microsoft or from Palm. I'll give you a great one for free. And, oh, by the way, it's going to be open source, and there'll be third-party applications that can be written for it. These devices will be super powerful.”

The ecosystem was like, “No, I don't want this one.” There was just no way in hell that AT&T or Verizon was going to work with a little rinky-dink startup that was valued at something like $10 million and had 8 employees. There were billions and billions of dollars at stake here.

But the other part of it, too, I think the reason that the smartphone market had stagnated for so long was this: everybody was happy. It's a nonpriority to upset the apple cart. It's almost like a version of enterprise software, right? The users don't like it, but the users aren't actually the customers here. It's the carriers who are the customers.

Yep.

So 2005 rolls around. Andy is now 2 years into the company, Android. He's managed to convince HTC, the Taiwanese manufacturer, to make a prototype with him. He's showing it to carriers and other OEMs. But for all the reasons we just discussed, it's tough sledding out there.

The company's running out of money. As Andy was going around trying to drum up investment for another round, he ended up meeting with Larry Page. Larry immediately said, “Forget raising another round. What if I buy you right now?”

So in July 2005, Google bought Android for $50 million. $50 million for Android. Oh, my goodness.

But, of course, that's a fallacy, because they would pour billions into development.

Yeah. They put billions in. But, to your point, this episode, Google is the hit factory here. This is the hit parade, right? The correct way to think about Android is that Google built it in-house, with a little kick in the pants from this startup that got far enough along with the idea that it forced them to do it now.

But they needed the kick in the pants.

Yes. Why was Larry so excited? Why did they buy Android right away? Eric, Larry, and Sergey all knew that they were late to mobile. Here we are: it’s now mid-2005.

We were 18 months away from the reveal of the iPhone. Apple and Google are very close.

Why do you think they knew that they were late?

I’m sure they were starting to get wind from Apple of what was going on.

That’s true. Eric’s on the board at this point, right, of Apple?

He’s not yet on the board, but he’s about to join the board. But the companies are very close.

Okay.

There’s that. But even let’s say they don’t know about the iPhone. BlackBerry is a thing. Yep.

Big adoption.

Yep.

Smartphones—and even Windows Mobile, as bad as it was—proved that there is demand. There’s clearly consumer demand for this.

Right? They had a version of Google.com for these devices to access, and they could see the traffic.

And they really knew it, especially from Maps. Google Maps on mobile devices—smartphones—was a killer application. Google was maintaining, I kid you not, different versions of Google Maps for mobile for the entire sea of phones out there. And so they know…

We have built our local government district on the desktop around our Disney World, and, uh-oh, it looks like mobile needs a district too.

Yeah. So, sure, you’re right. They would have done this anyway, but they were starting to feel already like, “Oh, shoot. We should have started this 2 years ago.” Buying Android kick-starts things. From the Google perspective, thank God they did, because we’re 18 months away from the iPhone launch. If they are starting from a cold start in January 2007, good luck. We’re not telling this story right now.

If they don’t buy Android and they don’t get started basically in the month that they did, this market belongs to Microsoft. Apple.

No, Microsoft.

Oh, why do you say Microsoft?

There are going to be 2 players in this market. I see what you’re saying. There’s going to be a fully integrated player, which Apple was going to be, and then there’s going to be an OEM plus licensed operating system. The model would have just been that Microsoft sells operating systems—a mobile operating system—to the OEMs who were freaking out that Apple was going to run away with it.

Great point.

To the OEMs—

Great point. So now back to Android and why Android was especially so attractive. Andy already had the right business model for Google. It’s just that, as Android the startup, OEMs and carriers are like, “Give it to me for free.” That makes you less attractive to me.

It’s funny how giving it away for free as a startup is a counter-signal. It makes you look desperate. But if you’re Google, it’s like, “Oh, they must have a really good plan here.”

Yeah. Exactly. So they start work on Android as part of Google in summer 2005. The plan initially is that there are going to be 2 versions of Android. There is a prototype and a device that will be more near-term to launch called the “Sooner,” sort of the more BlackBerry-like device, not a touchscreen device. And then there was a longer-term advanced research project, codenamed the “Dream,” for a touchscreen smartphone device.

In summer 2006, the next year, Eric Schmidt joins the Apple board.

He sees how far along and how good the iPhone is.

Uh-huh. And then, in January 2007, the iPhone is revealed in the greatest corporate presentation keynote of all time.

Yeah.

Eric is in the freaking keynote. Steve Jobs invites Eric Schmidt on stage.

And Android hasn’t been announced yet, right?

Nope. Nope. Nobody knows about Android.

This is in January 2007. And then July 2007 is when it shipped.

July 2007, yes, is when the iPhone shipped. Now, I believe Eric had disclosed the Sooner project to Steve because it was public that Google had acquired Android. I believe that Steve Jobs knew Google was working on a BlackBerry-style phone, but he did not know about the Dream prototype.

So Eric comes on stage. He makes a joke about merging the companies, that Apple and Google are so close they should merge. He says the company should be called “AppleGoo.” And then he jokes, “Well, but here’s the way with the iPhone that we can merge the companies without actually merging.”

He’s making these jokes, and the camera is focused on Steve Jobs, and he just has the ick.

I mean, that’s the best way to describe it. He’s trying to be a good sport and smile and be like, “Yeah,” but he has the ick.

It is unbelievable to watch this knowing everything that would happen over the next 10 or 15 years.

This incredibly close collaboration. There are 2 apps that launch in the very first version of the iPhone. Remember, it didn’t have an App Store. It was not open to third-party developers. There was a YouTube app and a Maps app, both of which were Google services.

Now, the apps are written by Apple. The icons are designed by Apple. They’re basically just consuming Google’s data as APIs. The only icons and apps on the phone are the ones that Apple puts there. And 2 of the…

Yeah, I don’t know how many there were—10, 12, 13 apps—are Google apps. It’s wild. By the way, the YouTube icon with the wood-grain TV was so awesome.

So awesome. I heard the YouTube team absolutely detested it.

They hated it. Yeah, they hated it. Well, because it wasn’t the YouTube logo, and they knew already—I mean, it was obvious this was not going to work. The YouTube team was like, “Apple didn’t put our logo on there. Of course, they’re going to start bringing in other video content over time.”

It was a little bit pre-algorithm, but the thinking was there: We have to make YouTube a destination and then control the experience when they’re in. And making the app icon reminiscent of an old-school CRT TV was also deeply antithetical to YouTube inventing the video of tomorrow.

Yes. Yes. It still looked great, though.

It fit in with that first iPhone for sure.

It totally did. Do you know who was the leader of the Google mobile teams that developed the backends for these apps?

Oh, no.

Vic Gundotra.

Really?

Yes. That was his first job, I think, within Google. First or second job within Google. Vic is going to come back up here in a minute.

So, the iPhone keynote: truly a world-changing, historic event. The Android team, of course, is watching this. And, yeah, that whole Sooner prototype—right in the trash.

The next day, right in the trash can. Directly in the trash can. The Dream is no longer a dream. It’s happening now.

Get in, kid. You’re the A team now.

Yep. Clearly, touchscreens are the future of mobile devices.

And a capacitive touchscreen at that. Yeah. So, remember, Eric Schmidt is on the Apple board. Once Steve Jobs finds out about what the Android team in Google is now doing, he goes ballistic.

Or perhaps, to use his word, thermonuclear.

Yes. Full-on classic Steve Jobs. Supposedly, at an Apple all-hands meeting—this is actually a little later—he’s overheard and quoted, leaked to the press, as saying, “We did not enter the search business. They entered the phone business. Make no mistake, Google wants to kill the iPhone. We won’t let them.”

Wow.

Which is, to this day, fair. Apple has been happy to just take a spiff off all the traffic that they send to Google and not compete in Google’s core business.

Yep.

Now, I will say I believe Apple reputation-launders a little bit. They get a lot of the value of being in the search business without having to do all the stuff that they demonize from a privacy, data-sharing, and all that ickiness perspective. But fine, whatever. It’s doing business.

That’s fair. Apple did not enter the search business.

So, in Walter Isaacson’s book, Steve Jobs says, “I’m going to destroy Android because it’s a stolen product. I’m willing to go thermonuclear war on this.”

Yes. He also says, “I will spend my last dying breath if I need to, and I will spend every penny of Apple’s $40 billion in the bank to right this wrong.”

He was pissed.

He was really pissed.

Now, interestingly, he doesn’t actually kick Eric Schmidt off the board until 2009.

Yeah, it’s interesting.

So, I think it took Steve a little while to realize what the Dream was within Google. And there’s also a reasonable argument back: Look, both companies took stuff from each other. A lot of the stuff that Apple touts—that they were the first company to ever do multitouch and that they own it—there were predecessor companies that did multitouch before them, too. The iPhone debuted a lot of technologies for the first time, and a lot of them were also just at the right time in history. I think Android arrived at a lot of similar conclusions at the same time.

True. It’s interesting you said multitouch. Multitouch actually becomes the battleground.

Because that’s the patent that they go to war over.

Those are the patents that Apple has. Steve Jobs threatens to sue Google over implementation of multitouch gestures. And so, as a result, Android for several years doesn’t have things like pinch-to-zoom or the sort of swipe operating-system UI navigation gestures. And I’m pretty sure if you remember early Android phones for the first couple of years, every single one of them had 4 physical buttons at the bottom of the phone to navigate the operating system.

I think this is why.

Huh. But let’s take Google’s side of this argument for a minute. When Android launches, they have the market, which today is the Play Store. Apple didn’t have an app store. Android had, when you swipe down, a notification center with all the notifications from each of your individual apps. You could—

It took Apple years to get that.

Drag to rearrange apps on the home screen. I mean, these are things that Apple then directly copied as well.

Right.

Yeah. All right. So, yeah, let’s get into the—

Great artists steal.

Exactly. Let’s get into the launch and the competition. So, November 2007. What’s that? 10 months after the iPhone reveal and 5 months after the launch. Remember, Android launches in 2008. Google announces the formation of the Open Handset Alliance.

That’s right.

And this is a partnership with HTC, Motorola, Samsung, LG, T-Mobile, Sprint, Qualcomm, Intel, Broadcom, and Texas Instruments. This was so confusing at the time. I, and everybody else, was like, “What does it mean? What is this? Is Google making a phone? Is Google not making a phone?”

So, then a whole year goes by with basically nothing. Then, in September 2008—a lot of things happened in September 2008: Chrome, Android, Lehman Brothers—Google announces the T-Mobile G1 phone, the Gphone. The T-Mobile G1 is manufactured by HTC. Remember, Andy Rubin and Android’s original partner in the prototype. The product name—the HTC product name for it—is the HTC Dream.

Heyo.

This is the Dream. This was what they were working on. In the U.S., it’s called the G1. It actually is a super interesting little device. I wrote an app for it. I had a class in college. It was a capstone class or something where I could pick my own project to do, and we had a 4-person team. One of the guys had a T-Mobile G1, and we wrote, I think, a Java thing for it. But he then founded the company DailyBooth after that.

Oh, wow. Yeah.

In fact, it may have even been a DailyBooth for Android app.

You had a lot of founders come out of your crew at Ohio State. Awesome.

So, it has a touchscreen on the front with the physical navigation buttons, like I was talking about. It has a slide-out horizontal QWERTY keyboard, sort of reminiscent of the Sidekick back in the day, unlike the iPhone. It has multitasking, so you can run multiple apps at once, and it has third-party applications.

Now, that’s a little bit unfair to the iPhone because, by the time the G1 actually launched, Apple had indeed just shipped the App Store. The event where they launch it is a T-Mobile event in New York City, in a commercial kitchen. It’s a haphazard, random launch. You can’t even find video of it today. There are little clips and still images you can find.

What is widely reported, and you can actually see in photos, is that Larry and Sergey do show up. They rollerblade into the building. They rollerblade onstage. There are all these T-Mobile and HTC executives there in suits. Here come Larry and Sergey on rollerblades onto the stage.

Yeah, it was haphazard, to say the least.

But the G1, or Dream, becomes a pretty decent success. It sells over 1 million units in the U.S., and just this one device—this one phone—gets 6% smartphone market share, which puts it roughly on par with Palm. The G1 kind of matches all of Palm in market share, but the smartphone market is still very small.

Yeah. It’s important to remember that mobile really wasn’t a thing until 2011. It was very obviously the next wave and the next computing paradigm.

Yep. But to be fair to Apple and the iPhone, it is starting to run away with the market. This is to the point of, man, if Google had not bought Android when it had, it would have been too late. Over the whole lifetime of the G1, they sell about 1 million units. The iPhone sold 11 million units in 2008 alone, and 20 million in 2009. Basically overnight, Apple and the iPhone went from not being in the smartphone market at all to over 50% market share of smartphones.

But as great as the iPhone was, it did have a few weaknesses.

No copy and paste. No copy and paste. Yep.

No multitasking.

As mentioned before, it didn’t multitask. It wasn’t very customizable. I think we’re still in the era when you can’t even change your wallpaper on the iPhone. I’m pretty sure we are.

I think it’s still just the black background.

Yep. You can’t put your own apps on it from anywhere but the App Store, even after it launches.

Yep. A big knock at the time. People loved that it was a touchscreen, but people really wanted the physical keyboards.

Yep. And the biggest problem with the iPhone, at least in the U.S., was that you could only get it on AT&T.

And you could only get it with the EDGE network. It was unusable.

That’s right. It didn’t have 3G.

It was so terrible. Eventually, the iPhone 3G came out within a year, but even that was really slow. The network had not caught up to what you wished the device could do for a few years.

Yep. So, that brings us to holiday 2009 and the Motorola Droid—

Changed everything.

It’s sort of funny to say now, like, “Oh, the Motorola Droid. This changed everything.” Yes, the Motorola Droid. I mean, when we interviewed Steve Ballmer a couple of months ago, he brought it up. When the Droid launched—it was holiday 2009—I think you and I were like, “Was it really that late? Wasn’t it early?” And he was like, “Nope. Christmas 2009. I will never forget it.”

That is when Android won the market. This was the moment—

And Google was really willing to put their brand second. Now, were they really putting their brand second? It’s Android versus Droid, so very convenient. But if you were to survey the American public in 2009, 2010, 2011, 2012, maybe even 2013, and say, “Do you know about Android, the mobile operating system?” “No.” “Do you know about Droid?” “Oh, yeah. I have a Droid phone.”

Then there were a couple of years after that where it was like, “Do you know about Google and Android?” “Yeah, maybe.” “Do you know about Samsung and Galaxy?” “Oh, yeah. I know about that.”

Yep. Exactly.

So, we’ll get into that in a second. The Droid—

Droid does, baby.

Verizon, at this point, is getting pummeled by AT&T. It’s been 2 years since the iPhone launch. AT&T isn’t just stealing a lot of subscribers from Verizon because of the iPhone. They’re stealing the best subscribers—the people who are willing to pay the most money for the biggest data plans for smartphones.

Verizon finally decides, “We’ve got to change the game here. We’ve got to be able to compete with the iPhone. We’re going to go all in on Android. We are going to buy a device and make this our flagship smartphone, position it against the iPhone, and invest hugely behind this thing.”

The device itself—the actual Droid, made by Motorola—was a great device. It had a big screen, big for the time, and a slide-out keyboard. It had a 5-megapixel camera, a removable battery, and all of these things the iPhone didn’t have. Probably the most important feature it had, though—the killer, killer app—was on the software side. It was the first Android device launched with Google Maps turn-by-turn navigation.

I didn’t realize that.

So, before the Droid, there was this whole consumer electronics product category of dedicated GPS devices. People old enough to remember may remember this: TomToms, Navman. People would buy these devices.

Garmin. Yep.

They would put them in their cars, and you also paid a monthly subscription fee for the turn-by-turn navigation service. Overnight, this entire product category gets obsoleted, Sherlocked, gone, because Google Maps is a better product with better navigation, and it’s free. No more monthly fees. It’s baked into your phone, the device you already have with you. Why on earth would anybody buy, let alone pay monthly for, a standalone GPS product again?

Yep.

And you know what doesn’t have it? The iPhone. The iPhone version of Google Maps—you had to manually advance the steps. It would pull up the route, and then you could tap the button to be like, “I’ve made this turn. Now show me the next step.”

That’s right.

Part of it.

That’s such a funny—you’re exactly right. I remember that, too.

Not really what you want to do while you’re driving, man.

Yeah. It’s crazy how not that long ago this was.

Totally.

That was the killer feature.

But even more important than all the features was the marketing and the muscle that Verizon put behind this. They licensed the Droid name from Lucasfilm.

That’s right. I think Lucasfilm was mentioned at the bottom in the credits of every commercial.

Yes, every commercial. They did this series of commercials that we’ve been referencing. Man, if you lived in the U.S. and were older than 12 at this time, this is burned into your memory. It was so great.

The first 80%—90%—of the ad was an Apple-style ad knockoff, with bright, happy, upbeat music and a white background. It had the fading Apple-style text, and it said, “I don’t multitask. I don’t have a removable battery,” et cetera. Then, in the very last 5 seconds of the ad, there was a hard cut with static noise. It was black and edgy, and then it said, “Droid does.”

The CMO of Verizon—Verizon did all of this—said that the campaign was designed to “wake up the market.” Boy, did it ever. That original Droid, I think, sold 250,000 units the first weekend it was on sale, and then it sold 1 million units faster than the original iPhone had. There was just so much pent-up demand for a real smartphone on the Verizon network.

Plus all the—yeah, this has turn-by-turn navigation, but even if you put aside whether it was better or not, it just was a real smartphone on Verizon.

Time magazine named the Droid its product of the year for 2009.

Wow.

The bigger thing, though, is that Verizon went all in behind it, even though they would add the iPhone later. It creates sort of this seed of what the Android user base would become today, at least in America, because Verizon went all in on Android, all in on Droid. Over the next couple of years, they followed the original Droid up with, let’s see, there was the HTC Droid Incredible, the Droid X, the Droid 2, the Droid Bionic, and the Droid Maxx. All of these had major marketing campaigns behind them.

It was game over for the segment of the market that was not Apple—the different OEM-from-operating-system model. Google just ran away with it. And before this, Microsoft had a shot.

They really did.

They were at a systemic disadvantage because they were going to carriers and saying, “Why don’t you pay us $5, $10?” And Google was going to them and saying, “Here you go. This is free. You can have the source code, and you can modify it as you see fit.”

Even today, I think Samsung has their own OS, Samsung One UI or something like that, that looks different. I mean, it’s Android, but it’s the open-source version of Android that they’ve customized. That’s the thing that’s on, I don’t know, 1 billion phones. And 3, we aren’t Microsoft.

Yeah, you guys don’t want to be Compaq, right?

Microsoft managed to suck up all the profit in that entire value chain. And handset makers, you currently make money. So why would you go work with Microsoft, who did that to the PC makers?

And as a little sweetener on top of all this, you know how I mentioned it was free and open source?

It’s actually less than free.

We’re actually going to pay you.

Yeah.

For searches that originate on your phone, we will give a little revenue share to both the carrier and the OEM, the handset maker.

Yep. So this was not widely publicized at the time, as you can imagine, but Bill Gurley wrote a blog post where he had heard from friends that Google was paying carriers and OEMs to use Android, even though Android was free. And he wrote this incredible blog post about it called “The Less Than Free Business Model.”

He basically predicted that Android was going to run away with this. If you’re a carrier or an OEM, sure, there’s a segment of the market that’s going to demand Apple. That’s fine. But Microsoft is dead. Palm is dead. BlackBerry is dead. There’s no way you can compete with free, let alone less than free, where they are paying you to take something of value for free.

Yep.

And from Google’s side, it’s the exact same thing as that thing we talked about with open-sourcing Chromium. They’re happy to give a few percentage points in traffic acquisition costs of their search revenue to people who are ensuring that the platform underneath them doesn’t belong to someone else.

There were some risks that it was all Apple, and then that creates 2 problems for Google. 1, they pay Apple a lot more money than they pay the combination of the carrier and the OEM maker. Those get a much smaller spiff. 2, this means that Google controls more of the underlying environment that they operate in.

Imagine how terrible it would be for them if mobile Safari was the new Internet Explorer and their entire franchise was at risk of Apple saying, “And we’re going to point traffic over here.” Google is happy to toss a couple of points over to these guys. I can’t think of another example of a dominant technology business and business model that has successfully survived and transitioned a major platform shift.

Yes.

And thrived in that next platform as well.

Mobile was a platform shift. A huge one. I mean, going from PC to the web was a platform shift. Going from PC and the web to mobile was an even bigger platform shift.

Play it out even further back in history than this. IBM was dominant in mainframes and then lost their dominance in the PC era. Microsoft was dominant in PCs and then lost their dominance in the web era. Google was dominant on the web and stayed dominant in the mobile era.

I mean, they didn’t derive giant profits from mobile, directly off of selling phones or selling the OS. They make some money on the Play Store, but not giant amounts relative to the rest of their money and what other players like Apple make.

But they kept search going.

But they managed to stay relevant to consumers with these hundreds of millions, billions of devices that they shipped, and their business was doing better than ever. I mean, all of these Android phones that are shipping, especially in the earlier years, what is the most prominent part of the UI on the touchscreen? A giant freaking Google search bar right there at the top.

Right?

The state of play of being a big tech company—and this dates back 80 years—is that technology moves fast, and the new paradigms disrupt everyone that came before you. So you get 1 era, and you have to make the absolute most of the 1 era that you grew up in. After that, you’re probably going to lose relevance. You might keep your money machine going for a long time. Famously, IBM made more revenue than Microsoft for a lot longer than people think.

Or even take Microsoft and Windows. Windows is still big today.

Yep. But the importance of that platform is going to fade and fade and fade.

But yeah, you won’t be able to transition your business model into the next era. Google did it.

And occasionally someone misses the second era but comes back for the third, like Apple figured out mobile. They never won a previous era. They were a player in PCs, but they didn’t win. Almost no one gets 2, and almost no one gets 2 successive ones.

Yeah. I mean, guys, like we said, this episode is the hit parade. Android basically, from the end of 2009 onward, just washes over the world like a tidal wave. In holiday 2009, when the Droid comes out, total Android market share of the smartphone market is still in the G1 range, like 5% or 6% global market share. 1 year later, 30%.

Wow.

They go from 5% to 30% in 1 year. They announced that over 200,000 Android devices are shipping every day around the world. The next year, in 2011, Android’s market share is 50%. And 2 years after that, by the end of 2013, it is 80% market share.

In many ways, it’s the Visa network-of-networks thing, where they don’t have to make every phone. They don’t have just 1 horse in the race. They’re getting leverage by having 2, 3, 4, 5 major manufacturers of these devices that are all independently doing their own marketing.

And there’s a very clever arrangement where you can just have the Android Open Source Project, and you can build your own mobile phone and launch it, and you don’t have our app store, and you don’t have to default to Google Search, and you don’t get Google Maps. You just have the operating system, and it’s great. Anyone can do that.

But why wouldn’t you want to have our app store? It’s where all the apps are. And if you do that, then you get all the great Google services, all the apps. You get the native Gmail and the native Maps and all this great stuff we’ve written. And if you do that, then Google’s the default search, and we’ll pay you for that—

And then you make money.

Yeah.

But by the way, if you want all this stuff that your consumers are going to demand, you are going to default to Google Search.

That’s the payment. Yeah, that’s the offer you can’t refuse.

Yes.

Now, here’s the actual crazy thing. As I said, by 2013, Android’s global market share is 80%. That’s actually higher than it is today. Today, it’s down to, I think, 72%, and Apple is 27.99%. Apple’s share has really grown. No question, Android pushed the iPhone to be better on many dimensions.

Things like cheaper iPhones, bigger screens, better cameras—I mean, on and on and on and on—things that I don’t think Apple would have done if Android hadn’t been pushing them.

Probably not big, cheap screens, but some of the cameras, I think.

Maybe. I don’t know. For years, the iPhones did not have good cameras. A big part of that Droid marketing push was the 5-megapixel camera. The original few iPhones had a 2-megapixel camera, I think. Like, it was crappy.

That’s right.

They’ve definitely pushed each other.

Yep. So then the other quick thing to mention on Android: there was 1 interesting moment of tension with Samsung in the early to mid-2010s.

Samsung basically said, “Oh, okay. The iPhone is the premium device. Android is this incredibly flexible platform. What if we just take Android and copy the iPhone with Android?” And they got really good at it. The Galaxy devices were just shipping in huge, huge numbers. And then Samsung started stripping out Google services and putting their own Samsung services in on some of their devices.

That was a bridge too far for Google.

So this is when Google started the Pixel program. Google had done the Nexus program, making their own hardware before. The Pixel, though, was and is a sort of reference device that consumers could buy, but more so to show the rest of the OEM market—the non-Samsung market—“Hey, here are reference designs, essentially, for great premium devices, great cameras, all the features you want. Here, copy these.”

It’s the same thing as the Microsoft Surface strategy, why Ballmer was so adamant: “We’ve got to make a Surface. We’ve got to show the OEMs how to do this.”

Right?

It’s funny. I’ve been trying to think about what the business of Android is—Google having Android versus Google not having Android. And I tried to pull up the most credible numbers I possibly could.

There's basically 2 things that you just have to add together to create the value. One is how much money they make from the Play Store, which has become significant. It didn't used to be, but it is now. And then the second is how much money they're saving by not having the searches originate from a platform that they don't own.

I used to think, because it's Android, they don't have to pay money. They have to pay $20 billion to Apple. It's not zero. They do actually have to pay. Like we talked about, Dave and I figured out as we were going through the financial disclosures and stuff, they do pay the OEMs and they do pay the carriers. The question is how much? Because once you can figure out how much, then you can do a little bit of napkin math to figure out how much they're still saving by it not being Apple.

Google paid out $55 billion in total traffic acquisition costs last year. I'm just using the current numbers to try to figure out what the splits have always been. Traffic acquisition costs are actually the sum of 2 different numbers from 2 different businesses, because they love to obfuscate things. One is what we're actually looking for: the acquisition of traffic to Google Search. The other component is money that we paid to publishers where our ads show up, in the sort of DoubleClick-AdSense world.

Yeah.

Now, we know that averages about a 70/30 split, and we know that they made $30 billion last year, gross, in the Google Network. So you could say, okay, they probably paid out about $21 billion of that $55 billion in the AdSense, DoubleClick, Google Network world. So that backs our $55 billion down to $34 billion. That's $34 billion in actual traffic acquisition for Google Search.

And we know $20 billion was iPhone.

Right, for Safari searches. So that means there's $14 billion that gets distributed to non-Apple traffic acquisition distribution partners, which in their annual report they define as browser providers, mobile carriers, original equipment manufacturers, and software developers. It's basically $14 billion to the Android mobile carriers and OEMs, plus Firefox.

Yep. What am I missing? I'm going to guess Firefox is less than $1 billion. Call it somewhere around half a billion-ish.

Yeah, there's probably some version of the old portal deals that still exist.

Properties on the web that have Google Search baked into them.

Okay. So let's cut $4 billion off for Firefox, the other web properties, and other things.

Yep. Okay, so $10 billion going to the carriers and OEMs. It's actually pretty significant that $10 billion is going to carriers and OEMs. It's half of what they're paying Apple.

Half of what they're paying Apple, but for many, many, many more devices.

Right. And so clearly, the revenue share to the carriers and OEMs is a much smaller percentage than what they have to pay Apple. I'd guess a quarter.

Either way, I actually think, after walking all the way through it, the bigger component of this is just derisking their future. It's not how many billions. They don't care about giving $10 billion up for this.

Yeah. As we've been saying all episode, Google is more than happy to pay traffic acquisition costs to any and everyone.

Yes. And then the direct value that they make from the Play Store actually came out in a lawsuit. In 2019, Play Store revenue was $11.2 billion. Gross profit was $8.5 billion, and operating income was $7 billion.

Now, $7 billion is not nothing, but it's still a far cry from Google's core business of ads from Search, Gmail, and Maps. That same year, the core business did almost $100 billion in revenue. Something like $85 billion in gross profit is my best estimate, and around $30 billion in operating income.

So even though the Play Store made $7 billion in 2019, the important thing is that Android is still primarily protecting the core Search ads business and making sure that traffic doesn't go elsewhere. This levered Google's web business into the mobile era. How amazing is that?

Yeah, that's true. It probably generated several hundred billion profit dollars that they may not have had those years otherwise.

Yeah. So I guess what I'm saying is, obviously, Android was a giant success. The biggest reason, even though they save—I don't know—$10 billion to $15 billion a year by not having to pay it to Apple, and even though they generate $8 billion—I'm sure at this point it's bigger, I don't know, $10 billion to $15 billion a year—it's really about just protecting the core, not about saving costs.

Yeah. And in this one, they almost missed it. If they hadn't bought Android, that window was closing fast.

And Microsoft did miss it.

Fast, fast, fast. Yep.

And so at some point, Andy Rubin leaves, and Sundar Pichai actually takes over the combined teams. Our hero here is starting to gather more responsibilities. It was just the application clients, and then it was Chrome, and in 2013 it becomes Chrome and Android. Whenever you see Sundar on stage, he is very proud of Google's 2 open platforms.

Yep.

So today there are more than 3 billion active Android devices. I think it's even higher than that now.

This is just silly. There's like 7 billion people in the world. They're over 3 billion active Android phones.

Yes. So you're probably thinking, coming into this 2010–2011 era, they're really feeling themselves over there at Google. We've jumped over some failures, but it's been hit after hit after hit in a lot of these areas that really matter.

Just like we talked about on the Microsoft episodes, it really doesn't matter when you fail, how many times you fail, or even the size of your failures, if your hits are these giant, world-changing, platform-type tech businesses that endure for decades. And that's what they had on their hands.

Yep. And it sure looked at this time like there was another big technology category out there, of a similar size, that Google should be playing in: social.

Yes. And this is the Google+ story. I'd say rest in peace, but I don't think anybody misses it.

Yeah.

All right. Well, I want to start this story the way that people expect us to start the story. And I have a little bit of a different take on it as we get partway in.

Google had been interested in social for a long time. They weren't blind to it. In 2007, they tried to do OpenSocial, and they basically failed at that because Facebook didn't participate, and Facebook was social. So everything else combined didn't really matter.

Oh, you didn't start where I thought you were going to start. The craziest thing is that Google had Facebook before Facebook: Orkut.

Yeah, that's true, which I think was like a 20% time project that then blew up in Brazil.

Totally. No, no. Yeah. Okay. So there was a Turkish engineer who worked for Google named Orkut Büyükkökten, and his passion was social networking. Friendster was a thing at the time.

In January, before Gmail, before the Google IPO, and before Facebook launched on the Harvard campus, he launched a social network within Google in his 20% time called Orkut. It didn't become that big in America, but at its peak, I think it got 300 million users. It was the biggest social network in Brazil and the biggest social network in India.

Wow.

And Google was like, I don't know, it doesn't seem that important.

All right. So, OpenSocial in 2007, Google Wave in 2009. By the way, can we just pause and say, in 2009, this is right after Chrome, right after Android? Google is a big place, and Google is a siloed place at this point.

It's kind of crazy that Android is happening over in this other building, and there's this fight with Apple, and at the same time they're doing Google Wave. It's weird that this is all sort of concurrent. The company was focused in a lot of different directions.

But it was so decentralized that it actually worked.

Well, it worked early.

Yeah, it worked really well to get all this stuff off the ground.

It was so interesting doing the research for this episode because so many of the people we talked to—even people who were leaders of a lot of these products—because Google was so decentralized and so siloed, they were focused on their thing, on Android or Chrome or whatever.

We'd ask, what was the overall strategy? What was the through line to all of this? We kept getting answers like, well, it was just Googliness. People worked on what they thought was cool, and it was good for the web.

That is absolutely true, but there was this overlay of a very, very thin layer of strategy that held the whole web together.

I think the strategy was pretty tight at the top level, and they just didn't actually need to communicate it down very far. Most people that I talked to said, I don't know, I was just trying to build great products that people love.

I think it was a feature, not a bug, that it didn't communicate down, because it let the teams below build really, really great products.

Yep. And they never really had to think about how this was going to help the ads business, and that was okay.

Yep.

So Wave failed because really nobody knew what to use it for, despite a dazzling and wonderful introductory video. Buzz, then, in 2010, created this big privacy debacle right at launch. It was super short-lived, and then it shut down.

So then in 2010, Urs Hölzle—very senior at Google at this point, probably a Distinguished Engineer and Senior Vice President—

Yeah, the guy who created the distributed infrastructure.

Right. After the Buzz failure, he was inspired to write this memo, kind of like Bill Gates's 1995 Internet Tidal Wave memo. A sea change was going on: the internet was becoming more people-oriented, and social media could be a problem for them.

The social media challenge requires a decisive and substantial response involving a significant deployment of personnel right away.

Essentially, the internet was now starting to organize around people in this Web 2.0 era, not just pages and applications—the things that were sort of the domain of Google. And so, here’s where I want to pause, David, and take it in a little bit different direction than I think you’re probably expecting: therefore, they went after Facebook. I think it’s a little bit more related to the palace intrigue at Google and a little bit less on-the-nose strategically.

And a little bit less on-the-nose strategically.

If you zoom out and look at the company right now, it’s pretty fragmented. It’s got different fiefdoms, with big personalities at the top of each of these fiefdoms: Android, Chrome, Search, YouTube, developer relations, all trying to will a Google platform into existence. There are different products with competing goals. Ultimately, they all help Google’s overarching mission, but there are a lot of elbows starting to come out.

Android was its own fiefdom, totally off on its own island, fighting an existential battle. Chrome is starting to do the same stuff as Android; they’re building their own operating system. It’s not clear what belongs in an Android camp versus a Chrome camp, and Sundar hasn’t unified them yet. Search is very protected—a separate team, especially the core people doing search ranking and monetization. No one touches them.

YouTube is totally separate. Gmail is massive, and it really is the only one in 2009–10 at the company that owns identity, since it’s the only Google property that you actually have to log into. YouTube has its own entirely different username-and-password system. It’s a mess, right? It’s a complete mess.

So Larry’s sensing this. He’s not CEO at the time, but he’s realizing the company is all over the place. He decides he’s just going to come back and get the company on track. I think Plus is kind of just the thing he picked as the single thing to try to galvanize and unify the company around. No matter what they picked and how they executed it, it was going to create a lot of carnage.

I can buy that.

There was a big shift that needed to happen in one way, shape, or form. Google+ ended up being kind of the ugly thing they did.

A recentralizing of authority, so to speak, within the company.

Right. So, in May 2010, they get the top 50 people at Google’s leadership assembled to discuss what to do.

The argument for this is that this is more of a convenient crisis.

Yeah.

And it might be a real crisis also, but it’s exactly what you’re saying, David.

So officially, in January 2011, Google announces that Larry Page will return as CEO in a few months, in April. Right away, Larry moves his office into what would become the Plus building. Wow. Yep.

So, they had just come out of this chapter. They’ve got this amazing business. The whole Chrome and Bing thing was defense against Microsoft. Android was defense against Apple and Microsoft. Google+ is now defense against Facebook.

Yep.

And legitimately, you could imagine a world where social ends up becoming way more important—the only places to put ads and the places where people are asking for information. There were rumors for a long time that Facebook was going to build a search engine. You have the attention; you can hijack it and do other stuff with it.

These were walled gardens. Facebook was a walled garden. Google Search couldn’t index what happened inside of Facebook, right?

And so, yeah, you could see how this is an existential threat. The traffic is growing. Oh my gosh, what if this becomes AOL all over again?

Right. And that’s the main thing. One tier down from that is that Facebook doesn’t even allow other ad servers.

At least with AOL, we could do a deal with them and power their monetization. Facebook just hired Sheryl Sandberg. They’re doing this themselves.

They’re doing it all in-house, a closed-loop system.

Yep.

So, Google+, what was Plus, and how did it get built?

It was a one-year sprint following this point—the 50 getting together—and it was built in a very, very un-Googly way. It was not organic, David, like these passion projects you’re talking about.

It was instilled from on high down upon all of the products.

It was not based on a core technical insight. It was not consensus-driven.

It was top-down, command-and-control style, led by the person that you mentioned earlier, Vic Gundotra. Now, who was Vic Gundotra?

Vic was this interesting character. Like we said earlier, he had been leading Google’s developer efforts in the pre-Android days.

And he was sort of the frontman. He was the MC at Google I/O.

Right? If you were looking for somebody to communicate and push down this new, top-down vision across the company, he would be a logical choice.

Yes. I don’t know if he raised his hand. I don’t know if Larry said, “Hey, I really think you should do this on our behalf.” But what is definitely true is that it became Vic’s thing, and Eric, Larry, and Sergey stepped back and let Vic run with it. He was given an enormous amount of institutional authority.

And we should say, too, you alluded to this earlier: What was Plus? It wasn’t just a social product in and of itself. It was baked into all of Google. It was inserted into every other product that Google had.

There’s a quote from Vic to the press at the time about what Google+ is. He says, “This is the next generation of Google. It is Google plus one.” Oh boy. Oh boy.

There’s a lot of hubris in saying that.

Oh, it gives me the heebie-jeebies.

So, yeah, it was a Facebook-style thing, but its goal, in addition to being a Facebook-style thing, was to leverage all of Google’s assets and make all Google things Google+ things. So they moved big headcounts out of each team and onto the Google+ team. They reached deep to integrate with these other products, and it’s very clear who the boss was in all these negotiations. You had a clear mandate: Your job this half, this year, is to do these Plus integrations.

Yeah, your OKR. Google famously ran on OKRs; it was now all about Plus-pluses.

And Danny Crichton, who would go on to become the managing editor at TechCrunch, at this point in time was a Google intern. He wrote about it later, and he said, “Due to this integration, much of it was forced. The culture around the company at Google had become deeply poisonous by the time I started. I still remember talking to one member of the Picasa team, which was Google’s photo repository that they bought, who told me to fuck off when I asked about integrating Plus into the product. He was hardly the only one.”

Companywide bonuses were based on the success of Google+. They even went so far as to put little Plus-one buttons on mobile advertisements, like those little banner ads at the bottom.

Yes. This is the best. Google had bought AdMob.

Yes.

And the mobile display ad units.

You could Plus-one it. Who the hell wants to Plus-one an ad? I mean, this is like Facebook’s Like button, but Plus’s version. And they’re like, “Any Google thing should be Plus-oneable.” So they even reached into YouTube comments, and YouTube comments became Google+ posts.

I mean, they almost killed the golden geese.

Right? They almost killed all of these golden geese that they had.

Yes. And so, Plus, from a product perspective, wasn’t just Facebook. They brought a lot of really interesting ideas. Google Hangouts came out of this. Google Photos came out of this. There were these things called Sparks. They really rethought a lot of social networking.

The issue is that nobody really wanted to rethink social networking. That was a Google priority to get people to use this, not a user-driven one. They tried to essentially put rocket fuel onto something that really didn’t have product-market fit.

Well, I really think the key huge mistake with Google Plus—one of the huge mistakes with Google+—was—

You don’t need a Facebook when there’s already Facebook.

Not even that. Facebook was already dying. Mark Zuckerberg had already realized that the future of social was not what it looked like at this point in time.

Meanwhile, as Google is launching Google+—

In June 2011—

2011, 2012, 2013—these were the big years for Google+. What is Mark Zuckerberg doing? He’s buying—

Instagram.

He’s buying WhatsApp, and he’s remaking, essentially, Facebook into what Meta would become. What we used to think of as social networking had bifurcated into 2 things: public media—that is, YouTube, Instagram, UGC—and private messaging.

And here’s Google launching—I kid you not, this is the craziest thing—desktop-first—

With a desktop-only UI to arrange your friends into Circles.

Circles. That’s right. Circles.

Which is, on its own, such a computer-science way of thinking about it. “Oh, my friends are in sometimes overlapping, sometimes non-overlapping groups that I want to carefully label so that I can identify deterministically who I want to share what with.”

Right? Nobody wants to do that.

Yeah. Here’s the thing that just leapt out to me about Plus: This was Google’s Windows Longhorn/Windows Vista. In our Microsoft saga, we talked about how Vista and Longhorn were the most damaging things to the company because of the distraction and the siphoning of resources and the best talent away from working on what really mattered.

Now, the question I was asking myself and others in researching this was, okay, what were the negative consequences of that?

With Microsoft, it was clear. What was the negative consequence for Google? The whole reason Microsoft let Google fester, from their perspective, for all these years and didn’t kneecap them was that they were tied up with all the distraction from Vista—

And losing relevance with developers because they kept selling them a platform that kept not shipping, and then when it eventually did ship, it wasn’t good.

Right. So then I was sort of trying to figure out, okay, what were the similar consequences for Google of the Plus era?

At first, I couldn't really think of any. I was like, “Oh, well, Android's pretty good. YouTube's pretty good. Chrome's pretty good. Search is still pretty good. Gemini AI comes out later. It's all pretty good.” But there were 2 things.

Messaging, probably. I bet in a non-Google+ world, WhatsApp, something like that, could be owned by Google.

Yeah. 2 things. One is messaging. I totally missed messaging. When I was a business school student at Stanford, Eric Schmidt was executive chairman, and he started co-teaching a class at the GSB. I took his class. I was one of his students during those years. It was awesome—one of the best classes I ever took.

The quarter when I was taking the class was when Facebook bought WhatsApp. I remember Eric Schmidt coming into class right after it happened and just saying, “Goddamn it, we missed it. We totally missed it.”

It's because Google was distracted. So that was one thing. Then I realized the other, bigger thing is cloud. Google should have been massively investing in cloud. There were all sorts of reasons that they didn't. We're going to save this for the next episode, but especially think about where the impetus for this came from: the memo known as the Zuckquake memo. Google should have been focused on cloud. It should not have been focused on social, and Google had the wrong strategy in cloud for many years. As a result, its cloud business is way behind Amazon and Microsoft.

Maybe they turned off some talent. Maybe there were some good people who got burned by the culture souring. You could argue this destroyed product velocity.

People complain today that Google is always working on really interesting technology, and they just never get cool products out the door. That is by far the biggest complaint you hear about Google.

Slow, big, and bureaucratic.

Yep. Folks on the inside and outside these days say it's just too slow. That's probably the biggest negative consequence.

Maybe you could trace that here.

I bet you can, because before this, we just spent this whole episode talking about all these amazing things they were building, shipping, acquiring, and transforming. Until Gemini—and arguably, that's a question mark—I actually don't know where Gemini stacks up. Is it a third-place product? Still a question mark. Until Gemini, what great breakthrough consumer service did they launch after Google+?

I got nothing.

Oh.

Yeah, that's pretty wild.

After having this incredible 10-year run—

And there was a lot of stuff they tried. I think there were some things for Android users. Think about Google Now, which predated Google Assistant, and maybe Google Home.

These are not world-changing products.

Right. It's funny. The thing that I keep thinking about from the Google+ failure is this big existential Facebook threat they were worried about. There was a strategy memo in 2013 where Neal Mohan said there was a risk that Facebook would become the starting point of the internet. Google knew social was the future and tried to win it. But, interestingly, they didn't, and they've been fine.

Right. Right. Right. It was all totally fine.

Facebook was really freaked out, too, that Google was going to come in and win it. Google was this giant, and Facebook was recently public and going through its own problems. Even though it was kind of a nothing burger and Google+ was a footnote in history, both companies were completely all-in on this being the big battle. Ultimately, Google wasn't a credible threat to Facebook.

Facebook went in a different direction anyway.

Facebook went in a different direction. Yeah. So it's almost like the end of Burn After Reading. Have you ever seen that movie?

No.

I won't spoil anything, but the feeling you have at the end is that you just watched all this crazy stuff happen and you're like, “Whoa, wait, did any of that matter?” That's how Google+ feels to me.

Yeah.

Funny. Plus did have 2 great surviving products: Hangouts, which became Meet, and Photos.

Yep. Photos is a billion-user product today.

Wow. Huge. The biggest thing—and I think this is getting back to my original postulate of never waste a crisis—is that you know what we have today? Google accounts.

Yep.

You know what Google is today? It's one company. It's not these little fiefdoms here and there, with different people amassing power and building things in different ways. I'm sure there's still plenty of that, but everything about Google got more unified from this era. They have a failed product and a smoking crater to show for it, but they have a unified look across all their products and a unified login that I think would be pretty important for them going forward. Anyway, my snarky finish on all this is that it's tempting to say Google lost in social because of this giant smoking crater, but actually, all of social ended up pivoting to either look like messaging or look like YouTube. YouTube is kind of the winning paradigm in quote-unquote social media, in user-generated-content media.

Yep.

So they should have just done nothing and watched the money printer go brrr.

Yep. To put a bow on it, Vic Gundotra ended up leaving the company in 2014. In 2019, they finally shut Google+ down. There's a blog post about it, and they cite a big security breach as the reason: “Oh, no. We've discovered this huge security vulnerability. Thus, we need to shut down all of Google+.”

Dude, it's so bad. There have been 50 Google products that all sound kind of the same. They launched one called Currents at one point, and when they shut Google+ down, this was horrible. Many people had written articles as posts on Google+, and they're just gone.

Yeah, that's right. It was sort of an impediment to doing some of the research for this episode because those posts are gone.

If you go to plus.google.com, google.com, anything, it just redirects you to Google Currents. However, Google Currents has now been shut down. There's a Google Workspace blog post announcing the Currents shutdown. Every time you click any Google+ link anywhere on the web, you go to a blog post that tells you about the shutdown of Currents. That's the most Googley thing.

Ah.

They've got to do a better job with those.

Well, by 2015—

It's our last section: the bridge to Alphabet.

Yeah, it's clear it's time for a new era at Google. The company announces that it is reinventing itself and becoming an entirely different company. Google is becoming Alphabet in August of 2015, and Larry Page will be the CEO of this new Alphabet holding company. Sundar Pichai will be the CEO of Google, which will be by far the largest and really primary operating company within Alphabet. Interestingly, they didn't at all decide to split up YouTube or any of the products. They just spent all these years unifying it all. That's all Google.

They broke out Google X.

Yes, Google X. They broke that out. Waymo was still part of X at this point in time; it would later spin out and is now part of Alphabet on its own. The Other Bets—really quite clever nomenclature here—were Nest, which they had just acquired; Google Fiber; Calico; and Verily, their 2 health companies; as well as Google X Lab. Then there were Google Ventures and CapitalG, the 2 investing entities that they had. So the question is, why did they do this? Why did Larry become CEO of Alphabet? Why did Sundar become CEO of Google?

I think this kind of had to happen as a healing after Google+. Sundar was a leader who had real credibility going back to the early days, with Chrome and with Android—2 of these core, great products and platforms that we've talked about the whole episode. Those platforms really drove the Google flywheel all along. Interestingly, he had never worked in search or ads.

Right.

But these were the platforms that had shoehorned Google into the mobile era and protected it from its greatest existential threat. Also, Sundar's personality, I think, was a way to reunify the company and bring everybody back together.

Definitely strikes me as a peacemaker among big egos. Yes. And that is where we're going to leave Alphabet and Google for the moment. Ben, give us a sense of how big this company had gotten.

At the end of 2015, it's gotten huge. It's $75 billion in revenue. $52 billion of that is first-party sites: Google websites, AdWords, Gmail, and Maps. $15 billion, the smaller part, is over in DoubleClick and AdSense land. Actually, that's pretty low-margin revenue, so again, the lion's share is in Google websites.

YouTube is profitable at this point, and Google's bottom-line operating income was about $23 billion. Its Other Bets lost about $3.5 billion at this point. The Other Bets are extremely interesting and will be the focus of our next episode. But the big takeaway here is that the business was still, in 2015, and essentially is still today, search ads.

Yep. What strikes me listening to you say those numbers in 2015 is that they're huge, but Google is so much bigger today on these same businesses with this same business model.

Right. There was another 5x scaling to go over the next 10 years.

Yes, it's crazy.

Google back then was about 20% the size of Google now, and basically nothing has changed when it comes to the business model and products. Nothing has changed since 2002.

Right. Well, I think this era, what we talked about all episode, was about stewarding that business through these sea changes. But nothing has changed about what the core business is.

It just turned out that that seed of an idea—search ads—actually scaled to the biggest market in the world.

Yep. All right. Just like the last episode with Gmail at the end, I've got one little coda, one little teaser for next time.

Great, Ben. What if I told you that between 2015 and 2016—so, this next year, this next 12 months after the Alphabet transition—all of the following people were Google employees: Alex Krizhevsky of AlexNet, the dawn of machine learning and AI; his PhD adviser, Geoffrey Hinton, the godfather of AI; his collaborator on the AlexNet paper, Ilya Sutskever, founding scientist of OpenAI; Dario Amodei, co-founder with his sister of Anthropic; Andrej Karpathy, until recently chief AI scientist at Tesla; Chris Olah; Noam Shazeer; Ian Goodfellow—hello—and, of course, the co-founders of DeepMind, which Google acquired in 2014: Demis Hassabis, Shane Legg, and Mustafa Suleyman. Mustafa runs AI at Microsoft today. Andrew Ng from Stanford, Quoc Le, Oriol Vinyals—and, oh yeah, in addition to all of those people, the authors of the Transformer paper, because Google invented the Transformer and published the paper in June of 2017.

Right? Which is the novel mechanism that all LLMs today, from every big foundational-model research lab, are based on. When I was talking to folks in the research for this episode and AI came up, one of them said, “You know, I have to remind people when I'm talking to partners out in the ecosystem that the T in ChatGPT stands for Transformer, and that we invented that.”

Because it is also during this time, while Ilya is working at Google, that he poses the question to his research colleagues and the Google Brain team: “Gosh, what do you guys think if we just built one really, really, really big neural network and we set it loose with training data on the entire internet?” Which, by the way, of course we can do here at Google because, thanks to the combination of the search index—we index the entire internet—and all the products that we just talked about on this whole episode, we have all this data.

And all this content out there. If we did that, do you think it would learn everything?

Well, David, that feels like quite the groundwork for the next episode.

That feels like a story for next time. Yeah. But through that lens, there is another way to view everything that happens at Google during this 10-year period we just discussed, which is that they're just collecting all the assets, all the information, and all the talent for AI.

It's nuts. There's this whole other world of research. Who would be the people that would drive the next decade or 5 decades of change? And they basically had them all in one place at one time. They were all employees of Google.

So I want to end with one more quote. This time from Larry Page, all the way back in the year 2000. This is Larry talking in the year 2000:

“Artificial intelligence would be the ultimate version of Google. So if we had the ultimate search engine, it would understand everything on the web. It would understand exactly what you wanted, and it would give you the right thing. That's obviously artificial intelligence, to be able to answer any question basically, because almost everything is on the web, right? We're nowhere near doing that now. However, we can get incrementally closer to that, and that is basically what we work on, and that's tremendously interesting from an intellectual standpoint.

“We have all this data. If you printed out our index, it would be 70 miles high now. We have all this computation. We have about 6,000 computers. This is 25 years ago. We have enough disk space to store like 100 copies of the whole web. So, you have a really interesting confluence of a lot of different things: a lot of computation, a lot of data that didn't used to be available. And from an engineering and scientific standpoint, building things that make use of this is a really interesting intellectual exercise. So I expect we'll be working on that for a while.”

Incredible. This is 25 years ago that he said this.

Amazing.

All right. Should we do some analysis?

Let's do some analysis.

All right.

Let's do power. And for those who are new listeners, power is the section where we analyze which of the 7 powers Google has from Hamilton Helmer's framework that enables a business to achieve persistent differential returns, or be more profitable than their nearest competitor, and do so sustainably. Google is very, very weird to analyze for this because most of the way you think about Google is actually not where the economic transaction is.

If you want to analyze the business, it is: Why are advertisers spending a marginal dollar with Google versus spending it elsewhere? And Google has the 7 powers that show up in numerous instances all over their business. But I think the interesting way for us to do this analysis, David, is let's look at each one. Just assume Google has them all and say where is the biggest, or a very large, example in our mind of where each of them show up.

Great. I like that. So counterpositioning typically doesn't show up for incumbents, for large companies.

This is the exception, though, with Google.

Right? Where you just look at their new businesses—for example, in this episode, talking about Android. They massively counterpositioned against Microsoft: the less-than-free business.

Less-than-free business model. I mean, this is the clearest example of counterpositioning I think that has ever existed.

Oh, hey, my competitors require you to pay them. How about I pay you instead?

Right? And my competitors can't do that because they don't have the business model of advertising based on search such that they can justify doing this.

Right. Scale economies, especially as they're adding all these apps, all these users, across all this surface area.

Now, if you're an advertiser and you want to reach users across search, display, or video, Google is a one-stop shop.

Right? You don't have to independently spend operational time and headcount on all these different platforms. You sort of get the one.

And that's not even to mention the scale economies on the infrastructure side we talked about last time. I mean, they show up in every business here, but that's just one example. And the fact that the more advertisers there are and the more users there are, the more profit Google makes because each little individual auction on every individual search finds a maximal price.

Yes. Network economies.

YouTube. Hello. More creators, more viewers. Creators make money from having views of their videos.

Application developers on Android and users on Android—the two-sided network economies there. Yes, everywhere.

Not to mention, in the core business too, in search, more users searching is more valuable to me as an advertiser because I have a deeper pool of people I can advertise to. So I can just deploy more dollars on your channel if it's working.

Yep. Switching costs. How about Gmail? Oh, I've got my last 20 years of email history in Gmail, all stored for free. Yeah, I'm not switching.

In the core business, there's not as much switching cost. I suppose there's a little bit of, “Oh, because I've spent a lot of money, the targeting is very good at allocating my spend,” but the switching costs in the core business for an advertiser are not as prominent as other powers. I don't think I continue to spend on Google because it's hard to switch. I continue to spend on Google because they have all the high-intent users for products other than, you know, Amazon. That's one of the 2 big search boxes in the world where people type in when they want to buy a product. So I'm going to advertise there. It has little to do with switching costs, I think.

Yeah. But for users of Gmail—

Oh, and several of the other products have enormous switching costs.

For users all across the board. Yeah, I won't leave YouTube at this point. The algorithm's dialed to my interests.

Oh, yeah. That's a great point of switching costs: the algorithm on YouTube.

Yeah.

Branding. I think in the heyday of Google that we're talking about in this episode, when they're launching all these incredible products, yes, these products—one, because they're incredible and because they were free—but also there was such a halo around the company. If there was a new Google product, I would be chomping at the bit to go try it.

Yeah, that's super true. I remember I was desperate for Google Wave invites. The product completely failed, but I was completely dazzled by it and desperate to get an invite and access. The Google name meant something.

And still does, by the way, which I think held them back in AI for a while. They know the Google name means something, so they are reticent to throw their name on it until they got kind of shoved off the cliff.

Yep.

Cornered resource. Yes. Well, certainly heading into the AI era now: YouTube, the YouTube catalog you can train on.

Yeah.

All the data they have.

Yeah. It's funny. I was about to say their infrastructure, but I think that's actually a scale economy: they've built out the infrastructure they have so they can run all their products as cheaply as they can.

Yep. I think the infrastructure is also a process power.

Yeah.

In the era we've been talking about, they could launch all these products on their infrastructure just way cheaper than anyone else.

You know what's a cornered resource? They have built internal software and systems that are better than what is available outside of Google.

Great point. A lot of the time, they even create open-source projects that are similar to their internal stuff, but they don't actually give away the internal stuff.

Inside Google, they still run Borg.

They run far less Kubernetes than they run Borg.

Yeah. When you talk to engineers who’ve left Google, they miss the infrastructure.

So Google has it all.

And we can name a lot more examples, but we’ve got to go.

All right. Playbook.

All right. I tried to get most of them in as we were going through the story. The first is that Google really wanted to become a platform company, and I was noodling on whether they ever did this successfully. David, we sort of touched on this idea that they are advancing the platform of the web without owning the platform of the web. So if they didn’t have Android, how would you answer the question: Is Google a platform company?

Yeah. I would say it’s like a shadow platform company. It’s like an ecosystem company.

Right. And even with Android, okay, great: They own the target development platform. Their money is still made elsewhere. It’s not a platform business. They may have a platform orientation as a company. They build a bunch of stuff for developers to build their applications on top of, but where their bread is buttered is really as an advertising company. It’s important when push really comes to shove on big strategic decisions the company has to make.

Yeah.

Like Apple, a pure-play platform company. Microsoft, a pure-play platform company. They either sell software or hardware, and then they need the platform around it to bolster their sales. Google’s very indirect.

Yep.

All right. So that was one. The other one is they made tons of small acquisitions, famously. I mean, that run in the 2010s. Aside from the big ones—YouTube, Android, DoubleClick, and AdMob—there was also what became Google Groups, Spreadsheets, Docs, and Blogger.

They bought Applied Semantics with the patents and some of the technology for AdSense. They bought the technology for Google Maps. They bought Urchin for Google Analytics, Dodgeball, FeedBurner, reCAPTCHA, Slide, Jambool, Like.com, Widevine, AdMeld, Punchd, Zagat, Sparrow, and Wavii. I mean, I could just keep going. There are hundreds of companies they bought.

In talking to folks for the research, there was this amazing part of Google culture that also fit the strategy perfectly: Help the web and the rich web and web apps bloom. Come work at Google with these incredible people, meet your co-founders, go start a startup, leave Google—

Right?

We will then reacquire you back into Google in a couple of years. It happened dozens or hundreds of times. I remember seeing this happen from the outside and thinking, “Google is nuts to let this happen.” But I realize now, no, this was all part of the strategy.

Yeah.

It’s all good for the web.

Yep. You can run very indirect, generous, long-term strategies like that with a money printer like AdWords.

Yes. I know I keep coming back to that, but that is at the core of what drives everything.

This one’s a little bit less playbook, but just an observation. I watched the Google I/O keynote with Glass, and I watched a bunch of Glass content. I even launched a Google Glass app at a Startup Weekend back in the day.

Oh, nice.

So, after watching all this Glass content—and it’s the butt of every joke now—Meta Ray-Bans and Google Glass are the same thing feature-wise. The gestures on the side, the fact that it could take a photo. I mean, Google Glass was a little more advanced. It could run very basic text-based apps, but I’m sure when Meta launches its little hologram version of the glasses, that’s going to be eerily similar.

Of course, you could say, “Oh, it’s just timing.” But here’s the thing: Google’s made you look like a cyborg. Meta’s are for normal people. And there is no better metaphor for the cultural difference between Facebook and Google than this. Google’s a bunch of wacky academics who did not really understand why this would make the product fail. Facebook is founded on the idea that you’re trying to be cool.

Yeah. Meta went and did a partnership with Luxottica to get the tech—

Into glasses that normal people wear.

Yes. It was crazy watching these demos because I’m like, these are the Meta AR demos. It just happens to have a cool factor versus not.

Yeah. And then my last one is this idea that they did figure out a way culturally to get people amped about: Just build great products. Figure out how to do something really hard from an engineering perspective that ends up being really useful, and ship things that people love. It’s not that you didn’t have to think about a business model, but a lot of the time, for many years after launching a product, you really didn’t.

Yep. It’s like we talked about earlier: There was this thin layer of really, really, really tight, really great strategy that was just a few people at the top of the company, but below that was just, “Make a great product.”

Yes. All right. I’ve got 2 for playbook. One that I’m going to make my quintessence: I just want to underscore again—we said this in the Android chapter—but Android was the mother of all wins. It was so big to win with Android. Nobody stretches a business model across technology eras. Nobody. And Google did it—

In a dominant way, where they are the dominant company in the next era as well.

Yeah. It is the Google version of Azure from our Microsoft series. It absolves any and all sins—not that there were many at Google. The only one was Google+. The only way it could have gone better is if, instead of launching Android, they launched the iPhone and also got the iPhone profits, rather than just some small dollars that protected their core business.

Yep. That was my playbook. And then my quintessence is: It is wild that this one company has 8 products with over 1 billion users and started this era with just 1 search that didn’t even have 1 billion users yet. Search, Android, Chrome, YouTube, Gmail, Maps, Drive, and Photos. And then, if you count the Play Store as separate from Android—which Google does—I think that’s a bit of a stretch, but if you do, then they have 9 products with over 1 billion users.

Just for context, Meta is the next-highest count of products in 1 company with over 1 billion users. They have 4: the blue app, WhatsApp, Instagram, and Messenger. Meta likes to claim they have 5. They like to say that Meta AI, in aggregate, has over 1 billion users embedded across all their products.

Yeah. But this whole superintelligence thing is an admission that the active users of Meta AI is a little stretchy.

If the Play Store doesn’t really count on its own, Meta AI for sure doesn’t really count on its own. So Meta has 4.

Apple, I think, only has 3, maybe 4. So the 3 Apple has for sure are iPhone, iMessage, and Safari. iPad, maybe? I don’t think so. Mac, definitely not.

I basically don’t count any iPhone app because they all come for free when you get the phone.

Okay. So, by your definition, Apple has 1 with iPhone.

I think Apple has 1.

Okay. All right.

Let’s take that same definition. How many of these came for free at Google? Google Search and Android—those are 2 completely different distribution channels. Chrome—

Yep. I don’t think any of these came for free.

I mean, Google helped Chrome.

No. Yeah. These are all independently achieved billion-plus users.

Gmail and Google Drive sort of advantage each other. So I think you can subtract 1 of those out.

Yeah. But it’s not to the extent that iMessage is default with an iPhone, right?

Maps is advantaged by Android. They ship a whole lot of Maps. But probably whatever the phone was would have a great Google Maps app.

Yep. Okay. All right. I buy it. Apple has 1. Microsoft has 2: Windows and LinkedIn. Amazon doesn’t have any billion-user products. Google’s got 8. That’s incredible.

Call it 7 or 6. I think it’s reasonable to subtract.

Okay, fine.

But still, that’s exactly right.

Whatever. That’s my point. This period at Google is a run like nobody’s ever had.

Yeah, absolutely right.

All right, what you got?

So quintessence for me is the thing that I can’t stop thinking about from the episode, and I decided this time—I knew what it was going in—I decided to hide it all the way until the end. So we haven’t talked about this thing yet.

Almost all of Google’s successful products are based on a core technology insight that is underneath the whole thing.

The type of insight that could be in an academic journal.

Yep.

Someone told me this, and I’ve been using it as a little litmus test for whether a product will work or not. As you look through—I mean, look at the original search. That is, by definition, the PageRank algorithm. It’s a core technology insight.

I mean, they published it as an academic paper.

The way that the ad-based auction works is a core technology insight. It’s almost mechanical in its elegance and its brilliance and its simplicity. It is a technology insight, and so was everything we talked about in Google Part I, our first episode.

Gmail—the way that they were able to offer 1 GB of storage—AJAX, fast, responsive web applications. You look at Maps and Docs with real-time collaboration: breakthrough core technology insights.

Yeah.

YouTube.

Yeah, totally. Serving video on demand to the entire world.

Absolutely.

Being able to scale that and make it a real going concern.

Yep.

Android. I can't name one magical core insight. This one may be the exception because it's technically hard and all that, but there's not an elegant thing that's the reason Android succeeded. It was perfect execution in a lot of ways: strategically, distribution, marketing, partnerships.

Okay. Wait, no, no. I got what it is. It's the same thing as the iPhone. It was an incredible achievement to wrestle OS X into iOS and to get it to run on a battery-powered mobile device that fit in your pocket. And Android did the same thing with Linux. They wrestled Linux into a battery-powered mobile device that fits in your pocket.

Yep. Less of an elegant, satisfying core insight, I think, and not the reason that it worked. I mean, it's not the reason why Android succeeded. Unlike these other ones, there's a clear line between them. It's almost like the Google products that succeed wildly organically, except for Android, are ones where there's almost no product. The technology solution is just so incredible that it is directly the user experience, and you get the technology breakthrough as the experience.

Yeah. Yeah. I see where you're going.

But then look at the other ones. Google+, Google Wave—these are products. These are user experiences that people come up with that don't necessarily have a breakthrough technology underneath them. Google Photos is actually quite the opposite. All of the AI stuff that's been happening on Google Photos for a very long time, that's why it worked. People wanted all these incredible magic features that come with Google Photos. It's funny: someone told me this in the research, and it's been batting around in my head. Then I'm reading Eric Schmidt's book, and Eric Schmidt said he would ask PMs, “What is your core technical insight that makes it all work?” And if there wasn't a good answer, he wouldn't fund the project. They figured this out at Google, too.

It's a googly thing that this genius technology is the product itself. And if you try to craft some cool idea that you have that is not just directly translating a tech breakthrough, it's not going to be the type of product that succeeds at Google. They don't know how. Some people can make an Instagram, and those people are not Google.

Ironic that Kevin was a former Google employee who left to start a startup.

Yes.

Anyway, I think that has made it extremely clear to me when Google products succeed and when they fail.

Love it. Spot on.

All right, carve-outs.

Carve-outs. I've got one, and then I've got my long-awaited follow-up.

Oh my God. We've been awaiting with bated breath. Listeners, what game console did David buy?

I'm going to make everybody wait for one more minute. My actual carve-out for the episode is when we were in New York for Radio City. My whole family came, the girls came, and we stayed for the rest of the week after the show. We took the girls to the Bluey experience at the CAMP store in New York City, and it was awesome. It lived up to expectations and lived up to the hype.

They basically have recreated the Bluey house in this physical space in New York City. The house is almost a character in the show, and they have recreated it. They just let you and your kids in to roam free in the house. Then you have a magical moment at the end of the experience.

Ah.

It was cool. Highly, highly recommend it if you are in the Bluey demographic and happen to be in New York.

Okay. What game console did you pick?

I bought the Steam Deck.

The Steam Deck?

I bought the Steam Deck, and it's great. Although, truth be told, I haven't had much time to play it this past month with everything we've had going on at Radio City and then preparing this episode. But it's great.

How'd you pick? What was the ultimate deciding factor?

It ultimately came down to this: as much as I desperately wanted my older daughter to be ready to play Mario Kart with me, she's just not. And so, if you're buying a console for just you to enjoy, I went with the Steam Deck. I was like, I would probably enjoy the Steam Deck more.

Do you endorse it? Do you recommend it?

Yeah. What Valve has done with the Steam Deck, I didn't realize until buying it and using it, is incredible. They have abstracted a PC gaming machine into a console experience. I've always liked PC-type games, but I haven't been a PC gamer in many, many years because I'm not going to build a gaming machine. You could just buy one, but I don't need another PC. Where am I going to put it? What am I going to do with it? I want the console simplicity of just buying the damn thing, turning it on, buying the games, and playing them.

Right?

Valve has created that in handheld form. It's awesome. You don't have to worry about any of the drivers or specs. It's really, really impressive.

All right, good to know.

So, I will buy a Switch 2 at some point, probably in the next year or so, but for now, Steam Deck. All right, what are your carve-outs?

I've got 3.

Oh, great.

My first one—and I swear to God, this is unrelated to their sponsorship—is Claude.

Amazing. It's so great.

It's so good. Using AI has completely changed the way that I prepare for these episodes now, and I cannot imagine going back.

I hear AI is a thing. I hear AI is a thing.

So that's the first one. I just find myself in it all day. Now, 2 is the Sony RX100 VII, or 7. So, I recently bought a different camera, the Fujifilm X100VI, or the 6.

Yeah, that's what you had in New York.

Yeah. And it's great. It's like the internet's favorite camera. It has these amazing film-simulation color profiles. It's a camera, though. I carry it around my neck because it's a camera that you hold and use, and it's very fun shooting 35-millimeter equivalent. It feels like I'm taking pictures the way that pictures were meant to be taken.

It's not a full DSLR, but it is a big thing.

Exactly. It's handheld, but I wouldn't call it a pocket camera. Now, the funny thing is, the thing that I'm actually talking about as my carve-out is the Sony camera. Sheel Mohnot tweeted actually this morning. I should have been preparing for this episode, and I was replying to him on Twitter instead, that he's been considering getting this Sony or another point-and-shoot camera, and I just remembered how much I love this camera.

The Sony RX100 VII fits in my pocket. It's very small. It has a giant zoom lens for its size, and I was just looking at some of the pictures that I've taken with it. It's the perfect thing to bring with a phone. Whenever I am space-constrained—which is usually; I just don't really want a camera around my neck—the perfect combo is to bring a phone and bring the Sony.

I am aware that it's not a full-frame camera. I'm aware that it's not as photographic as my Fuji, but it is the most practical one for most things that I want to do. And for many, many shots, it is far superior to shooting on a camera phone. So, I don't know. I just love it. It's a 2019 camera, and they really need to come out with one that has USB-C because it's annoying to charge, but other than that, it's just awesome. So, I highly recommend it.

There you go. You're bringing it full circle on this episode. A point-and-shoot camera.

Point-and-shoot camera, especially paired with Lightroom, has this great AI feature called Denoise that they launched, and it's now rolled out in production. It is incredible.

Love it.

Then I have one more. A listener recently sent me—he started a clothing company called Kerosimi—and it is an incredible garment. It is just this really, really nice cashmere shirt. I've been wearing it all recording. It's great on a cool day. It's great on a warm day. It's my current favorite shirt. And so I wanted to thank the listener who sent it to me and say: you have built a very nice clothing company. The prices are high, but the products are just excellent.

Well, I've been staring at you for the past 7 and a half hours here, and I've been thinking the whole time, God, Ben is looking good.

My normal thing, if I could just wear it every day, is a long-sleeve dark crew neck. I don't have to think about it. You can look nice in it. It just goes with everything. It's the capsule wardrobe idea. And this is the finest version of that that I've worn. It's really great. Kerosimi.