Remembering Susan Wojcicki

I was deeply saddened to hear the news of Susan Wojcicki’s recent passing. She’d been battling cancer for the past two years. Susan was the former C.E.O. of YouTube and the fifth guest I ever had on the show back in 2020. It was right in the middle of Covid lockdowns and it was the only time we ever spoke, but I’ve had a soft spot in my heart for Susan ever since. Two things stand out for me in this interview. The first is that she actually gave meaningful answers to tough questions. I found that most C.E.O.s are so concerned about how they and their companies are perceived that they give completely nondescript responses when talking in public. Susan, in contrast, seemed eager to have a real conversation. The second thing that caught me off guard about the interview is that she was actually nice. It’s really hard to rise that high in the business world and stay nice. She will be deeply missed.

*      *      *

I’m so excited to talk today with Susan Wojcicki. She’s the C.E.O. of YouTube. She’s had an amazing career in tech, and she’s done it without really attracting all the hatred that seems to circle so many of the powerful people in tech. She doesn’t really have any enemies.

Susan WOJCICKI: I’ll go to places really remote and then I’ll see someone. They’ll be like, “Hey, I learned Korean on YouTube.” And so the opportunities for technology have just changed our lives in so many ways. And I recognize that there are all these questions about responsibility. But I’m overall really optimistic about our opportunities to help people all over the world have access to better information and learning and ultimately live better lives as a result.

Welcome to People I (Mostly) Admire, with Steve Levitt.

She’s managed to get herself into great situations and then to make incredible decisions. Everything she touches seems to turn to gold. Is it her intellect? Is it her instincts? I’ve never spoken with her before. I’m really curious to hear what makes Susan tick. Now, the wild card today is that given Covid lockdown, we’re both doing these interviews from home. I’ve got my six kids at home and she’s got her five kids and her husband at home. And we will see what kind of chaos might ensue. 

*      *      *

Steven LEVITT: So you grew up living on the Stanford campus. You were the daughter of a physics professor at Stanford and a high school teacher. You went to Harvard. You majored in History and Lit. And then, as I understand it, you decided you wanted to get a Ph.D. in economics, of all things. Is that really true?

WOJCICKI: Yes. Yes, it’s really true.

LEVITT: Where did that desire come from?

WOJCICKI: Oh, I had taken the introductory economics class as an undergraduate. Once I started it, I realized how interesting it was as a field and just loved it, so I wound up getting a master’s in economics and I had the idea of going on and getting a Ph.D. and then continuing to work in economics. And that’s when I discovered technology along the way. And so I wound up having to detour. And there was definitely a day where I had to make that decision: was I going to apply or was I going to go back to technology? And it was a rough day. I had tears in my eyes because I wasn’t sure if I was going to the right place. I could have continued and really enjoyed being an economist, but technology just seemed like it was taking off. It was in my backyard and I looked down both paths and I chose the technology one.

LEVITT: And I’m just guessing you don’t look back very often and say, “Damn, I wish I had gone for that econ Ph.D. My life would’ve been so much more awesome and enjoyable if I’d done the Ph.D.” I think you definitely made the right choice. So you live in — lived, at least — in Menlo Park. Is it common for people in Menlo Park to rent out their garages?

WOJCICKI: No, it’s definitely not common. I was just looking to rent it so I had rent income and could pay the mortgage as a just-graduated-from-business-school student with a lot of student loans. And, it just so happened to be that it was a startup that found me and it was Sergey and Larry, the founders of Google, who wound up renting our house.

LEVITT: So you just put an ad in the local newspaper or something? “We have a garage” — or, how does that work?

WOJCICKI: I did put an ad at the Stanford Daily or the equivalent at the time, but it turned out I had a friend who was good friends with Sergey and they were getting started at the same time and they were looking for office space. And this was during the first boom of technology and so they really struggled to find any space. And I said, “Well, we have three rooms here in the garage. They’re really tiny rooms.” And they thought it was great because they were living in their dorm and they were super excited — there was a washer and dryer and there was a hot tub in the backyard. They really thought this was the best thing ever. And at the time it seemed totally normal. All my friends would come over and say, “Who are those guys?” I’d be like, “Oh, they’re just some Internet guys at my house.” And then afterwards they found out they were Google. So it created a lot of fun stories for us. And I enjoyed it. I enjoyed having them around and talking to them. And late at night they’d be there. I’d just go hang out with them and talk about what they were building. So, yeah, it’s hard to imagine in today’s world that happening. But that’s the way it was 20 years ago.

LEVITT: So, Susan, you eventually get so enamored with the young Google company that’s in your garage, so you join them as their 16th employee. Were you actually working in your own garage, as an employee?

WOJCICKI: I wish. I should have. But no, they had actually moved out already. Initially, when they were in my garage, they got up to seven people and they had just hired engineers. But then they realized they needed to hire someone who did stuff other than engineering, and they weren’t really sure what that was. But they thought whatever it was, I could probably figure it out since I had an M.B.A., which was more than they had at that point. I really didn’t know anything myself, but I knew a little more than they did. And they hired me and my title was marketing manager, and they weren’t really sure what that meant, but they said, “We’d like to create a global brand, and we’d like you to figure out how to make that happen. And you have no budget to do that. But why don’t you just sit down here and get started?” And I was probably young and naive enough to just say, “Sure.” And really quickly moved from marketing to product development and creating products. And that’s where I’ve been ever since I joined Google.

LEVITT: Were you the only woman when you joined Google?

WOJCICKI: There were one or two others. Head of H.R.. was there. Oh, yeah. And there was another woman who had joined, she was doing deals. And she joined the day before me. So technically I was the third woman. But it was small.

LEVITT: And still, tech these days, women are highly under represented. Why do you think that’s so? Do you have theories on that?

WOJCICKI: Yeah, I definitely have theories, having been in it for so long, and then also having watched my own kids grow up with this. I have to say, I think it starts early on with kids in high school or in middle school. For some reason, there may be more boys that gravitate towards doing work with a computer. So by the time you get to college, some people actually have quite a lot of knowledge. And then getting started is actually really hard. So what I did with my own kids, I really encouraged my girls when they were young to make sure they had a strong C.S. background so they never felt overpowered by others in the class. And I think the way to fix this ultimately is to have computer science in say, middle school, and make that available to everybody. If everybody graduated from middle school knowing how to code a basic language, everybody would have that ability and that base to continue to go on, and I think it would make a huge difference in having more diverse groups in computer science.

LEVITT: Well, that certainly fits into the bandwagon that I think we teach totally the wrong things in school, and the earlier the better to get people involved in data science and programming and otherwise. Although I have found with my own children’s experience with coding, is really— I feel like we’ve gotten really good at teaching older people how to code. And I’m not sure we’ve really cracked the nut of how to teach younger kids to code. Do you have any feelings on that?

WOJCICKI: I think the challenge — what I’ve really seen — has just been the availability. The number of people who can teach it is limited and people need to have a computer to be able to learn it. And those are scarce resources and a lot of schools don’t have it. So I think if we were able to solve that resourcing issue. I do think the tools are out there. Like Scratch, where they learn basic concepts, is a perfectly fine way for kids to start learning, and then to start learning a language whether it’s Java or JavaScript or something that can start showing them. And I think as soon as kids make that connection, “Wow, I can actually build something,” that’s useful. As soon as they just make that connection, then you’re home free. I mean, you think about math or reading. Learning math or reading when you’re first out, you’re like, “Why am I learning letters and numbers? Makes no sense.” But as soon as you can read, you’re like, “Oh yeah, of course, it’s really important. I need to be able to do this.”

LEVITT: Let’s talk about the Google ad environment, because everyone knows how brilliant the Google algorithm is and how impactful that is. But I have this sense that without the cleverly designed ad environment, Google wouldn’t be at all what it is today. So were you involved at the beginning? It sounds like you were with you know the building of how the ads are done and the second-price auction and all that? People probably don’t really know that much about it. Could you describe some?

WOJCICKI: Sure. When Google first started, there was only banner advertising, and people used to say there’s no money in search. What Google said was people come and they do a search on something and the ad should be related to what they searched on. So if you searched on the latest N.F.L. game, there should be tickets or paraphernalia from that. And we went to all the ad networks and we said, “Can you do this? And can you enable us to serve ads on a million different terms?” And nobody could do that. So even though we had 30 people and no one in the company knew anything about advertising — we had like not a single advertising expert in the company — we decided to build an ad system, which in retrospect is just crazy. Looking back on it. And we built these really basic text ads. And we ran those ads. And I just want to tell you that nobody clicked on them. It was a complete failure. The really big change, and the huge breakthrough that we made was when we started looking at the clickthrough — how the ads performed — and having it be a C.P.C.-based solution, which is a cost per click. So we basically said how much money are we making and then how many people are clicking on it? So I would compare that to inventory in a store. If you have a store, you don’t just put your expensive items in the front. You have the expensive items that a lot of people are buying. And that was basically the click through. And so we started looking at both the cost per click as well as the click through. Just that breakthrough, which was really simple, made a huge difference. It was better for our users. And we actually generated more revenue when we started factoring in whether people clicked on those ads or not. And I remember there was a period of time at Google where nobody got this concept. I remember we all went around and we all tried to say something about C.P.C., clickthrough, and everybody failed. I mean, today we just — it just rolls off of our tongue in terms of how we think about our ad and our ad development. But we didn’t really have any expertise in this area.

LEVITT: Well, things that are new are hard. I think people dramatically underestimate how hard it is to do anything new. And once you figure it out or your teams figure it out, then the next generation, you just explain it. And it makes sense. But to actually figure stuff out, is hard.

WOJCICKI: It was.

LEVITT: Yeah. And the other thing that I think is under appreciated about how brilliant the Google system is that the typical company has a huge sales force that goes out and negotiates prices with their clients. But can you imagine if that’s what Google tried to do? And the thing that’s so remarkable about your system is, number one, that it’s totally self-serve, right? It’s almost completely self-serve, maybe for the big customers you do some kind of sales force. And the other one is that you do this very economic-y thing, which is you have your companies that want to place ads, they tell you essentially how much they’re willing to pay for different spots. And they even do it in this very clever and complicated thing called the Vickrey second price auction, which to an economist makes our heart beat a little faster because it’s supposed to make people reveal their true preferences. So when someone’s searching for a new car, that’s worth a whole lot of money compared to when someone’s searching for, I don’t know, what the weather is going to be today probably, and the system just takes care of that without having to be hand-held by a sales force.

WOJCICKI: Yeah. We used to have a sales force that sold all the ads that are on the top of the Google search page. And it used to be that we automated everything that was on the right hand side of the page. That’s the way it used to be. And we realized quickly that we were generating more revenue with this automated system. And so there was actually a big transition where we had to go to our sales team and say, “Hey, you’re going to have to use this second-price auction solution that we have.” And so you can imagine that did not go down well initially with all of our sales teams. But that was actually a huge improvement for us because suddenly we were dynamically adjusting our page and prices based on supply and demand. So, factoring in the C.P.C. that people are willing to pay, the clickthrough, and then using the Vickrey second-price auction, and then doing that across all of the inventory, all of the ad slots on the page, made a huge difference in terms of just having the most efficient way of figuring out who wants to be shown on this page and what users were interested on that page.

LEVITT: Yeah, really, it is an economist’s dream. I really can’t think of another company that has so embraced the basic principles of economics, and you certainly can’t argue with all that’s gone down.

You’re listening to People I (Mostly) Admire with Steve Levitt, and his conversation with the C.E.O. of YouTube, Susan Wojcicki. They’ll return after this short break.

*      *      *

LEVITT: So skipping forward, by 2005 or 2006, you are now in charge of Google’s video service. And usually in firms, there’s such an unwillingness to accept that anyone else could do it better than your firm. And you had the backbone to say, “Look, YouTube is doing something special. Let’s go bring them inside, not fight them.”

WOJCICKI: Yeah, it was a really hard time for us, because we had actually created Google Video before YouTube came out. So we actually came out first. So we had this moment where we saw it and were like, “Wow, this is a huge opportunity for us. We’re doing so great.” And then YouTube came out. And this all happened within a year, and we saw it slipping away to YouTube. And our teams, of course, were very motivated to compete with YouTube. So going and saying, “Hey, let’s go buy YouTube,” was a really big decision. I mean, it was hard because you had to say, “Hey, we weren’t as good as we thought we were.” But it was clear to me that that was the right decision for us.

LEVITT: Do you think that it helped that you had been so successful at Google up to that time, to be able to say, “Look, we just can’t handle that”? I just think, in all the corporate environments I’ve ever been around, you’re just not allowed to say, “We’re not good enough. Let’s spend 1.65 billion dollars to buy this little upstart, because we can’t beat them ourselves.”

WOJCICKI: It was hard. But I’ve attended Google’s management meetings pretty much since the beginning of Google. And so it was easy for me to take myself out of the equation and just look at what was good for the larger company, irrespective of what was good for me. And, I wound up doing the model that showed that it was worth paying 1.65 billion. And also just to let you know, we only had a day that we were able to go through the whole process. And so I probably did the model in an hour at the most. And afterwards, I wound up getting asked a lot of questions about it. And why did you assume all these things? And I always have to remind people: I only had an hour. I really couldn’t do it with that much detail. But I wound up not working on it initially, even after I had done all this work to encourage us to buy it, because we had a little reorg and I wound up taking over and managing all of the ads. So it took eight years until I could work on YouTube again.

LEVITT: Well, I just want to go on record as saying that I remember vividly a lunch conversation I had in the days after Google acquired YouTube, and I was sitting in the Faculty Club — the Quad Club — at the University of Chicago with a bunch of other economists, including some Nobel Prize winners. And honestly, we almost never agreed on anything. But all of us collectively agreed 100 percent around the room that that was the stupidest business move that had ever been made in the history of mankind. And history has proven 15 years later that actually it was probably the opposite. It might have been the single best acquisition of all time. But that’s probably why we’re sitting in the Faculty Club and why you’re running YouTube now, which you took over as the C.E.O. in 2014. And let me just start by saying that YouTube is amazing. People watch a billion hours of content on the site a day, and that’s a level of success that I think even the model you were building back in an hour when you did the acquisition probably never imagined that kind of success. But of course, with that success comes all sorts of responsibilities and challenges you probably never imagined either, like hate speech and terrorism, misinformation, foreign governments trying to influence elections. I mean, basically, now YouTube and Facebook and Twitter are essentially blamed for every problem that exists in the world. And the critics are screaming “Why don’t you just fix these things?” So I’m guessing it’s not that simple. These are hard problems. And the reality is that as hard as you try, it’s just not realistic to think that you’re just going to fix these problems. Is that a fair assessment?

WOJCICKI: Well, I think we have been focused on this area. We call it responsibility, or trust and safety. And we have been really focused on it — I’d say we started around 2016. That’s maybe really where the light bulbs start to go off, that we needed to make a lot of changes — and have implemented a lot of changes across our entire system that I think have made a huge difference. And we’ll continue to improve them, we’ll continue to staff them. But I think if you look at the systems that we had then versus what they are now, it shows that we have been able to respond really quickly and our systems are built in a way to make sure that we can deliver trusted results for our users. And during Covid, I was so grateful for those years and years of work, because we were able to so quickly have accurate information for our users. And that was all due to the work that we had done over the last couple of years. So, of course, there will always be some new loophole that somebody will figure out, and we’ll close that loophole.

LEVITT: Yeah, I mean, that’s the nature of adversarial interactions, is you fix one thing and the bad guys go and try and break a new thing. But somehow, even more fundamentally, I think as I watched the debate over the tech space, you’re being asked to make decisions that even in the best of cases governments have difficult times making. And I just find it odd that these are companies that were founded in basements and in garages 25 years ago. And now the executives are being asked to make fundamental decisions about life and society, which I think is just — you can’t possibly be equipped to do. So, in general, companies hate the idea of being regulated, but would you actually welcome more regulation? Because that would put the onus for these, I would say, essentially impossible decisions on government instead of you as the executive?

WOJCICKI: Gosh, I have a lot of different thoughts in this area. I mean, first of all, it would be much easier if there was some regulation. And we just said, “Hey, we’re just complying with the current rules.” That would simplify everything for us. On the other hand, as a private company, we can actually make in many ways much more detailed, fine-grained and really thoughtful decisions in this area. And you look at something like Covid. So there were new conspiracies that were coming up almost every week: 5.G. causes the symptoms of Covid. It’s not a virus. And so, would the government be able to respond quickly enough to be able to come up with the right set of requirements? I’m not really sure. And I also think there’s something good about having multiple companies come up with different perspectives on this, and that we’re trying to — as much as possible — to balance all the different needs of the constituents around these really important and tough topics.

LEVITT: There’s so much public hatred of Facebook and Twitter and personal enmity, really, towards Mark Zuckerberg, Sheryl Sandberg, Jack Dorsey. Why do you think you basically are completely off the hook on that? How come you don’t get lumped in with other folks?

WOJCICKI: Well, I mean, YouTube in many ways serves a few really important functions. During Covid, we served over 400 billion impressions related to information that came from public health officials, local public health officials around the globe. And so a lot of people during Covid — again, just to use that as an example — were looking up how to fix something in their house, how to give yourself a haircut, how to exercise at home, how to do yoga. So as an information source, I think we’re a really important part of that equation. And we are also able to connect a lot of people. I saw a lot of people who otherwise wouldn’t have been able to meet with whether it was their local school, or their religious organization, come together. And I also just think people come to us looking for a laugh. And we’ve always had as a top search request, “funny video.” That’s always been a top thing that we’ve seen our users come for. And so I think it’s a combination of those.

LEVITT: Okay, I’m sure I’m aggravating many listeners and they’re angry that I’m not challenging you and saying you’re wrong about everything.

WOJCICKI: You’re welcome to. Go ahead.

LEVITT: Well, no, the problem is, I actually agree with what you’re saying. So I don’t want to argue. But there’s actually one area where I do think you could do better, and just to simplify things: assume that everyone could agree on what was hate speech and what wasn’t hate speech. And my impression is that the goal of YouTube is to minimize the amount of hate speech content on the platform today. You set things up because you want to make sure that today you take as much down as possible, and you spend enormous resources identifying and immediately removing all sorts of offensive content. Okay. But that seems to me to be an enormously shortsighted approach, in the sense that it ignores the fact that, in the jargon of game theory, the battle between you and the hate speech people is a repeated game. And shouldn’t YouTube’s strategy reflect the fact that you care not about how much hate speech is on the platform today, but you care about how much hate speech is on the platform a month from now, or a year from now, or five years from now? And it seems to me that if you took a different longer-term approach — so not the goal of taking down everything in the day, but actually tricking the people who are doing hate speech and leaving their stuff up, but making it so that very few people see it, and tracking them, and maybe working with authorities to actually get them arrested and stuff. There’s a whole menu of strategies that it seems to me that you and the other tech companies have punted on to achieve a short term goal. So how do you respond to that?

WOJCICKI: Well, I guess I would say that I don’t think we have necessarily gone after a short term goal. What we have focused on is, first of all, defining hate speech in a way that we think is appropriate, is enforceable, and is fair. And we’ve made a significant number of increases in terms of the removals that we have there. So last quarter we removed 11 million videos. And you can look and see what percentage have come from hate.

LEVITT: But that’s my point, which is, a really short term way to deal with hate speech is to take down people’s videos and then let them start a new account and put them up again.

WOJCICKI: So, first of all, we’ve taken an approach of not enabling there to be a business created on hate content. And I do think that that has long term implications. So as soon as you have revenue and a revenue model then you can start to grow it. And you can hire people. And by removing the videos we remove the economic incentive to do it. But we also have an issue where a lot of times there’s content that is not necessarily violative but really brushes up against our guidelines. There’s always going to be content that’s up against that line, wherever you have drawn it. And so what we’ve done with that is we have reduced the viewership to that by not having it be recommended to our users. If you look in the U.S., we’ve actually had an over 75 percent reduction of the content that we classify as borderline content. And that has also had an impact. When people create content on YouTube, they have two goals. One is that lots of people see it. And the second is that they generate revenue. So we don’t want to be too overarching in our approach. And so, there’ll be some content that we’ll allow on the platform. But they can’t monetize it, and we won’t recommend it. So, basically there’s a whole grayscale in terms of how we approach these different types of content. It’s not just black or white, on or off the platform. There are a lot of different dials that we use. And at the end if you produce content and you can’t generate any revenue and no one’s going to watch it, you’ve taken away a lot of the incentives for people to create that content.

LEVITT: Okay, I’d love to shift away from businessy stuff into a more personal realm.

WOJCICKI: Sure.

LEVITT: Let me ask you, what kind of advice would you give to a 15-year-old who’s hoping to build a tech career?

WOJCICKI: I would say, first of all, I think tech’s a wonderful career and it’s a huge area of opportunity. And I’ve always found it to be easier to be successful if you’re a company that’s growing. And what I’ve always looked for are the areas that are high growth but are not necessarily universally recognized as high growth yet. I’ve also looked for companies that have some kind of mission and that are doing something meaningful. And so, Google could have failed, but at least it would have failed with a big mission, which was to organize the world’s information to make it universally accessible and useful.

LEVITT: Yeah, I love that advice. How about let’s say, now you’re a 30-year-old. You’re not in tech, but you’re unhappy with the path that you’ve been on, and you’re looking to get into tech. Any advice? Would you say, “Forget it; it’s too late”? Or do you think there are entrees for people like that to actually have a shot?

WOJCICKI: There are definitely entrees. I took my first computer science class when I was a senior in college, and I almost didn’t take it because I almost thought it was too late for me. But of course, I think it’s really never too late. And not everyone is going to be an engineer. So if they did finance for another company, they can do finance at a tech company. Having demonstrated interest in technology is always a benefit. But I think there’s lots of opportunities, if you’re in your 30s, find how you can leverage your current skill set and make a transition.

LEVITT: So we both have a bunch of kids. I’ve got six and you have five, and we also both have a big age spread in our kids. And so in principle, we both could have had a lot of learning along the way. And I’m often asked by people, “Hey, what did you learn in the first round of kids that you raised that has been beneficial in the second round?” And honestly, I’m stymied by that. And the only explanation I have is, I was so sleep deprived that I can barely remember what it was like to have the young kids the first time around. I, literally, find myself tongue-tied when people ask me advice about raising kids. Are there things you’ve learned that could be helpful to other people about raising kids?

WOJCICKI: If I look back at my young self and me as a young mother, I probably worried a lot about following what were the established rules. Like your kids shouldn’t sleep in your bed, and your kids shouldn’t do this, and your kids should be toilet trained by this age. And there are all these things about what your kid is supposed to do by these certain times. And of course, there’s some milestones where if they don’t reach it, you worry because there’s maybe some developmental delay that you need to help them reach. But I’d say by the time I got to my fifth, I just didn’t really worry about that anymore. And I wish with my older ones I hadn’t been so worried about all these different milestones and just enjoyed the moment. I’d say the second thing I’ve learned as my kids have gotten older is, I think it’s really good for kids to have chores in the house. And with my younger kids, I’ve done it earlier. It’s very hard to tell a 12- or a 14-year-old, “Hey, now all of a sudden we expect you to do stuff in the house.” So I’m sure we all could do more and get our kids to be more productive in helping the house. But I wish I had done that earlier. I learned that as my kids got older. Four-year-olds don’t get angry about it because they’re excited because they’re contributing to the house. It’s easier if they’ve always just been expected to do a lot of things in the house.

LEVITT: Finally, last question. Do you have advice for people about living a good life, a life worth living, a life they could look back on and be proud of?

WOJCICKI: One of the things that I have thought a lot about is just making sure that I live a life without regrets. So I’ll just say, there’ve been some decisions that we’ve had to make at YouTube that have been really, really hard. Hard for me in the sense that maybe they’ve introduced a lot of risk, but yet I think they’re the right thing for us to do. And so I want to do what I think is the right thing over the long term. And I would also say I think it’s important to have a multi-dimensional life, not focus too much on just one area. And as a working mom, to be fair, you’re always making tradeoffs that are not optimal. There’s always something that you missed for your kids. So I’m not trying to say that that’s not going to happen. But I just want to make sure that when I look back on my life, that I feel that I did the right thing for my friends, my family, and in my job, and for society.

Susan Wojcicki died recently at the age of 56. 

And this Friday we’re back with a brand new episode, featuring Chris Anderson, the mastermind behind the TED conference and the wildly popular TED Talks. As always, thanks for listening and we’ll see you back soon.

*      *      *

People I (Mostly) Admire is part of the Freakonomics Radio Network, which also includes Freakonomics Radio, No Stupid Questions, and The Economics of Everyday Things. All our shows are produced by Stitcher and Renbud Radio. This episode was produced by Matt Hickey and engineered by David Herman with help from James Foster. Our theme music was composed by Luis Guerra. We can be reached at PIMA@freakonomics.com. Thanks for listening.