It’s Your Problem Now

Stephen DUBNER: So let me ask you this. A lot of C.E.O.s seem to want to be as uninteresting and unrevealing, almost as inhuman as possible, at least when they’re speaking in any public or quasi-public setting. Is that strategy? Is it personality, or something else?

Carol BARTZ: I think they have bad advisers. They have, I think, one of three things. They either have that as a basic personality. Two, they’ve been advised to be as benign as possible. Or three, they are scared.

That’s the former tech C.E.O. Carol Bartz.

BARTZ: And I don’t buy any of that. People want to know where you stand, where your company stands, what you stand for, where you’re going.

Jeff SONNENFELD: Not all of them, but many of them, have become over-managed.

And that’s Jeff Sonnenfeld of Yale, a longtime observer of C.E.O.s.

SONNENFELD: The bureaucracies have overtaken them. When I started, C.E.O.s weren’t treated as celebrities to the larger world. They didn’t have this global, rock-star aura about them. And that’s been to their detriment. Their schedules have them running from event to event, that — it’s hard for them to have time to listen, and they often learn things way too late.

What sort of things do C.E.O.’s learn too late? Let me put it bluntly: they hear about the stuff after it’s already hit the fan. Today on our show: the next installment in our “Secret Life of a C.E.O.” series. Interviews with the C.E.O.s, past and present, of Facebook, Microsoft, PepsiCo, Reddit, Virgin, G.E., and more, telling us about all the different kinds of stuff they’ve had to handle. From consumer rebellions …

Ellen PAO: There were maybe 15 or 20 thousand people who were very upset.

… to changes in the business climate:

Steve BALLMER: We tried too hard to keep our old model alive.

And huge changes in technology:

Richard BRANSON: And then iTunes came along.

But what if you’re the one with the new technology, and it becomes so powerful that the world starts to panic?

Mark ZUCKERBERG: That’s a thing that Facebook and the Internet I think have really worked to change over the last 15 years.

C.E.O.s also get to deal with plain old fraud …

Jack WELCH: “Jack, we’ve got $400 million missing.”

And the joy of managing troublesome employees:

PAO: How do you fire the one person who’s actually going to get you over the finish line?

Hey, the corner office is empty. You want it? You sure you want it?

Satya NADELLA: Taking those decisions is not easy. It weighs very heavily on me personally.

*      *      *

It’s not your fault. You weren’t the one who moved the goalposts. You weren’t the one who made all your customers suddenly want something new. You weren’t the one who started a worldwide financial meltdown. So no, it’s not your fault. But it is your problem. Because you’re the C.E.O. So what are you going to do about it? What are you going to do if, for instance, you’re running a global soda-and-chips company just as the world decides that soda and chips are borderline toxic?

NOOYI: We knew we had to retool our portfolio.

That’s Indra Nooyi, the C.E.O. of PepsiCo and the star of our previous episode.

NOOYI: That was just not even a question. We knew that if we didn’t do it, our future was in jeopardy.

PepsiCo did retool its portfolio, adding what it calls “good for you” and “better for you” products to its “fun for you” snacks. And it’s been working: PepsiCo’s stock is at an all-time high; and Nooyi still has her job.

Richard BRANSON: We’ve never had a company go bankrupt.

That’s Richard Branson, founder of the Virgin Group.

BRANSON: And that’s in 50 years, so we’re proud of that.

Virgin is a conglomerate — a small one, as conglomerates go — with businesses ranging from travel and leisure to financial services to music.

BRANSON: We started off with record shops. And we built maybe 300 record shops around the world, Virgin Megastores. And then iTunes came along, and the Internet came along, and people, sadly, didn’t see the need to go into record shops anymore. And so we either sold or closed down most of those 300 record shops. But that spurred us on to move into mobile phones and into new technology that was evolving.

Branson has been a serial reinventor. In contrast to a consumer-goods company like PepsiCo, Virgin is essentially a lifestyle brand. So it can apply that brand to just about any sort of business. Many of which, we should say, haven’t worked out so well. It may be true, as Branson says, that he’s never had a bankruptcy. But the list of Virgin failures is long, and diverse, including Virgin Cola, Virgin Vodka, Virgin Games, Virgin Cars, Virgin Clothing, Virgin Brides, and Virgin Student, a social network Branson launched in 2000.

BRANSON: But by and large, we’ve had many more successes than we’ve had things which we’ve had to say — say goodbye to.

Jack WELCH: We carried businesses that had lost money for 10 years, 15 years — but we could afford it.

That’s Jack Welch, former C.E.O. of General Electric, talking about the state of the company when he took over, in 1981.

WELCH: But we couldn’t afford it any longer. The Japanese were coming. They were eating our lunch.

Welch responded in force. He slashed underperforming business units, along with tens of thousands of employees. And he pushed G.E., an old-school manufacturing and technology firm, into a variety of unrelated territories. Financial services and insurance. Even broadcasting. In 1986, G.E. bought the N.B.C. television network. Jeff Sonnenfeld again:

SONNENFELD: What was G.E. ever doing with N.B.C. anyway? It was good for cash flow, but did it, strategically reinforce the rest of what G.E. was doing? Did G.E. have a good logic for being in what turned out to be quite perilous insurance businesses? Some of these other businesses were very far afield from G.E.’s expertise, and became burdens later. And there were things that were done, in terms of accounting practices. Kidder, Peabody — that was, again, on that list of mistakes.

G.E. bought the brokerage firm Kidder, Peabody, for $600 million in cash, over the objection of some board members.

WELCH: That’s true. I supported my team that wanted to do it. And a couple of board members were smart enough to challenge it, and they were right.

Shortly after the acquisition, a former Kidder executive was found to have been selling insider information. Several years later, another scandal: a Kidder bond trader had faked hundreds of millions of dollars in profits to hide trading losses.

WELCH: I was going out the door on a Friday night, and the guy running Kidder, Peabody, he called me and said, “Jack, we’ve got $400 million missing.” I got sick to my stomach. I was torn up, and then I went down to Kidder, Peabody for the weekend to find out where the $400 million went. And we had to come out with a press release on Monday that our earnings were $400 million short. I went to the urinal that night. I was standing next to a guy, and he turned to me — the Kidder, Peabody guy, turned to me and said, “Jack, this won’t affect our bonus, will it?”

G.E.’s acquisition of Kidder, Peabody, one analyst later said, “was like G.E. lost its mind.” What was a company known for jet engines and turbines doing in financial services and insurance? But from a stock-market perspective, G.E. under Jack Welch exploded. When he took over, its stock price was just over a dollar. When he left, 20 years later, it was over $40. But today? G.E.’s stock is at around $16, and the future is not bright. Its current C.E.O. recently announced that G.E. still owes billions of dollars to cover liabilities from an insurance business it bought decades earlier.

Jack Welch, Richard Branson, and Indra Nooyi have all tried to master, with varying degrees of success, that buzziest of business-school buzzwords: the pivot. In fact, just about every C.E.O. we spoke with, when we asked about the biggest problems they faced, said it involved some kind of a pivot.

PAO: My name is Ellen Pao.

Ellen Pao grew up in Maplewood, New Jersey …

PAO: … which was a great town to grow up in.

Her parents weren’t from New Jersey.

PAO: I’m the daughter of immigrants from China.

Her parents were both engineers who came here to work on the American space program.

PAO: I ended up going to college at Princeton, and law school and then business school at Harvard. I practiced as a lawyer, and then went into the tech industry, where I’ve been for the past almost 20 years.

Part of those 20 years were spent at Reddit, the self-proclaimed “front page of the Internet.”

PAO: It was exciting to be someplace where there was all this positive energy and people were really tied to technology.

Pao became interim C.E.O. of Reddit in 2014. The site had roughly 135 million active users at the time. Fans of Reddit consider it the most genuine, engaging, and free-wheeling replication of humanity to be found online. Critics say it also reflects the worst of humanity, with pockets of misogyny, racism, and more. Pao wanted to change that culture.

PAO: One of the things that we did that was really quite daring at the time was to remove revenge porn and unauthorized nude photos from Reddit. And shortly thereafter, all of the big companies followed — Google, Facebook, Twitter. They were waiting to see what would happen if somebody did it, and it actually turned out to be pretty unremarkable. And we got rid of that content, and the site didn’t go up in flames, people didn’t leave in droves, there wasn’t a huge uprising.

There wasn’t an uprising then. But later:

PAO: The uprising came when we removed five of the most harassing sub-reddits from the site.

A sub-reddit is a user-run community within the larger Reddit site.

PAO: And those were sub-reddits where people would gang up on other individuals. In one case, they were trying to get somebody to kill themselves. In another case, they’re sharing private information. They were shaming individuals in a way that was very targeted and specific. We got rid of those five sub-reddits and people went up in arms. Yeah, there were maybe 15 or 20 thousand people who were very upset. One of the sub-reddits we got rid of was one that was fairly popular. They ended up putting up a lot of terrible pictures, terrible memes, terrible content about myself, about —

DUBNER: About you.

PAO: Yeah, about me. And photoshopped images, and just saying terrible things about myself and my family.

DUBNER: And what was that like for you?

PAO: Most of the site was either neutral or positive. I did get a lot of messages from people who were just relieved. One of the people said that they had had a friend of their sister’s had been, had experienced revenge porn, and had ended up committing suicide. The importance of taking down that content, and taking these steps to clean up the site and make it a place where everybody could feel comfortable was something that was really meaningful. I knew that it was the right thing to do. And that was what helped me get through that time.

The pivot that Ellen Pao attempted at Reddit was about a change of content, a change of tone. Richard Branson’s multiple pivots at Virgin were in response to changes in technology, and taste. Indra Nooyi’s big pivot at PepsiCo was about a shift in consumer preferences. The next pivot we’ll hear about covered all those bases, at a company far more valuable than PepsiCo, Virgin, and Reddit combined. We’ll start here:

Steve BALLMER: Five hundred dollars? That is the most expensive phone in the world and it doesn’t appeal to business customers because it doesn’t have a keyboard, which makes it not a very good e-mail machine.

That’s Steve Ballmer, the then-C.E.O. of Microsoft, when asked in 2007 about the impending launch of Apple’s iPhone. Microsoft had been in the phone business for years.

BALLMER: Right now, we’re selling millions and millions and millions of phones a year. Apple is selling zero phones a year. In six months, they’ll have the most expensive phone by far, ever in the marketplace, and let’s see. What’s the expression? Let’s see how the competition goes.

That competition, as you likely know, went very badly. Apple built a phone that a lot of people loved. And they got wireless carriers like AT&T to subsidize its high cost. Steve Ballmer, when we spoke with him, acknowledged his miscalculation.

BALLMER: So Apple did business-model innovation. We tried too hard to keep our old model alive as opposed to moving to a new model. We did not take the right path.

As Apple was climbing, Microsoft was falling hard. But it wasn’t just a phone failure; it was about computing generally. Back in 1999, Microsoft was the most valuable company in the world, with a market cap of around $620 billion. Adjusted for inflation, that would be about $930 billion today, which would still be the most valuable company — just ahead of Apple. At one point, Microsoft’s Windows software ran on more than 90 percent of the world’s computing devices. Today, if you count smartphones and other smart gadgets as “computing devices,” which they are, Windows is at around 11 percent. The mobile revolution was on, and Microsoft was very late to the barricades. Steve Ballmer’s solution, finally, was to buy the phone business of the Finnish firm Nokia. It wound up costing roughly $9 billion.

BALLMER: Yeah, the Nokia purchase was sort of not mine by the time it got done. I recommended it to our board in June of 2013. Our board turned me down. I got riled up about that topic and I didn’t think it was handled well. It was one thing to turn me down, it was another thing the way it happened. So anyway, I was a little upset about it. But then, after I made clear that it was time to go find a new leader, the board came back and said, “Let’s go ahead and buy it.” And by then they knew I was going to leave and they still wanted Nokia purchased.

Okay, so this is strange. Steve Ballmer wants to buy Nokia to help revive Microsoft. The board says no. Ballmer feels mistreated; not long after, he says he’s stepping down. Then the board says yes. Which puts the Nokia deal on the desk of the new C.E.O. Which was who?

NADELLA: Satya Nadella, C.E.O. of Microsoft.

In February of 2014, Satya Nadella became just the third C.E.O. in Microsoft history, after Steve Ballmer and co-founder Bill Gates. Nadella was a relatively low-key insider, not the shake-things-up outsider some people thought Microsoft needed. He was an engineer who’d been with the firm since 1992 and had ultimately been brought onto the senior management team. His mission then: to help Microsoft adapt to a world driven by mobile and cloud computing.

DUBNER: Steve Ballmer, your predecessor, famously pushed to purchase Nokia, the fading mobile-phone company, toward the end of his tenure. You voted against it, but the deal wound up going through a few months after you were appointed C.E.O. So, I’m just curious how this works. First of all, why did the board select a C.E.O. who voted against this gigantic recent acquisition?

NADELLA: I mean, just to put the facts and make them clear — I was not a board of director at Microsoft. I was part of the management team of Steve. So, it’s not like I had a vote. Steve just went around the room and wanted to get the pulse of his leadership team, and we had a good debate. And I felt that it is important for us to do things — given where we were in the mobile space at that point, which was the number-three slot with a huge gap between one, two, and three — to do something that was more unique and different and differentiated.

DUBNER: So, shortly after you were installed as C.E.O. you shut down Nokia, which resulted in a total write-off of the purchase and about 18,000 jobs lost. What was that — I mean, that’s a pretty big deal to be handling both — the mechanics of it and the emotion of it, shortly after you come in as a not-obvious choice as C.E.O. Just walk me through what that felt like on the way to accomplishing that.

NADELLA: First of all, these hard decisions around what to pick and focus on is something that I believe a C.E.O. uniquely has to do. That’s not something that you can delegate. I mean, ultimately, that’s your core responsibility, and especially taking those decisions that impact people’s lives and livelihood is not easy. It weighs very heavily on me personally. Therefore, I had to think it through. And then having thought it through and made the decision, we had to execute on it. To your point, where what was of paramount importance was to make sure that the employees being impacted were treated with dignity and were given all opportunities to find their next play, whether inside of Microsoft or outside it. That was my real concern and that’s where I poured my energy. But I knew that I had to define the core value propositions that we were going to create.

SONNENFELD: The company was getting quite mature in an industry that was facing really existential disruption.

Jeff Sonnenfeld again, the C.E.O. expert at the Yale School of Management.

SONNENFELD: He did come into a culture that had become very bureaucratic. It had become possessed by very strong turf issues.

Sonnenfeld thinks Microsoft has benefited from Nadella’s management style; it’s what you might call a mellow intensity. Sonnenfeld points to how Nadella handled the Nokia mess.

SONNENFELD He had voted against this merger with Nokia. He thought it was a mistake. But he doesn’t run around trying to shame Ballmer and Gates over it. He said, “Look, as we look at it now, in a clear-eyed way, that wasn’t the path to take.” Doesn’t mean that a mistake was due to idiocy or corruption. It’s that in the light of day, with somewhat different conditions, we see we need to take a different path now.

The path Nadella has chosen for Microsoft isn’t radically different. They’re still making big acquisitions: LinkedIn, for $26 billion, and another two-and-a-half billion for Mojang, the Swedish maker of Minecraft. Microsoft is still a software company that sometimes tries to sell hardware — which historically hasn’t worked so well. But lately, they’ve gotten better, especially with their Surface tablets and laptops. But Nadella is pushing Microsoft particularly hard in three new, or new-ish, directions. The first is cloud computing, one of Nadella’s specialties before he became C.E.O. Here’s what he writes in his recent book, which is called Hit Refresh: “When I took over our fledgling cloud business in 2011, analysts estimated cloud revenues were already multi-billions of dollars with Amazon in the lead and Microsoft nowhere to be seen. … Amazon was leading a revolution and we had not even mustered our troops.” Under Nadella, Amazon is still way out front of everybody, but Microsoft has seen huge gains, especially in the past couple years.

The second big shift under Nadella: partnerships. In its early days, Microsoft eagerly partnered with software firms like Adobe and hardware makers like Dell and H.P. Here’s what Nadella writes in his book: “I don’t think Google would have existed but for the PC browser. Microsoft enabled Google to build a toolbar for our Internet Explorer.” But as the tech ecosystem evolved and firms like Google, Apple, and others became a threat, Microsoft began treating partners and would-be partners as rivals. Nadella is trying to reverse that. One example: he’s allowed Microsoft apps like Office and Skype to run on mobile devices that use rival operating systems, including the open-source Linux, which Steve Ballmer once called a “cancer.” Nadella’s leadership thus far has been one pivot after the next.

NADELLA: Nothing can be taken for granted and there’s no such thing as a perpetual-motion machine. What you have to do is be good at being able to refresh yourself at the crucial times.

The third big push under Nadella? This one, you sense, is his favorite. Microsoft is trying to build what it calls “the ultimate computing experience” — essentially a layer of technology that would lie beneath pretty much everything we do. It’d be a blend of augmented reality, artificial intelligence, and quantum computing.

NADELLA: This is the real fun part, right? Just imagine if your hologram was right here interviewing me as opposed to just on the phone.

Nadella recently demoed a new augmented-reality model to Ford.

NADELLA: In the past, Ford would create these clay models, which weighed 5,000 pounds, that needed to be moved so that people can critique the new car. Whereas now, they have essentially these sessions where people in manufacturing, design, sales can all look at the model simultaneously, annotate it, leave voice comments — I mean, it’s just a complete new way to collaborate. In A.I., that the ability to reason over data and create intelligence is another amazing, amazing breakthrough.

I give you, again, a very tangible example. A group of people came together at Microsoft and created this new app called Seeing A.I. that anyone can download — from the Apple app store, in fact. It uses all of the cutting-edge, machine-learning A.I. techniques around computer vision from our cloud and brings about the capability for someone with visual impairment to be able to see. In fact, one of my colleagues Angela Mills, was telling me about how she has a visual impairment, and she uses that app now to confidently go into the cafeteria, order food. She walks in — and I had not even realized this could be such a challenge — which is she said, “I can now walk into a conference room at work knowing that that’s the conference room that I’m expected to be in, instead of barging into the wrong meeting for the first time.” And to know that A.I. can actually help someone fully participate in her job — it’s remarkable.

DUBNER: Satya, you’ve got over 120,000 employees around the globe. If we put them all together in one room, how many do you think you’d know by name?

NADELLA: Let’s say 5,000.

DUBNER: Wow, that’s impressive, really?

NADELLA: Yeah, 25 years, 5,000, 5 percent. Yeah, for sure.

Perhaps the biggest challenge for Microsoft, and Nadella: a quantum leap in computing itself.

NADELLA: Ultimately, I believe in order to bring about some of these magical experiences in A.I. capability, we will have to break free of some of the limits we’re hitting of physics, really. I mean, Moore’s Law — even when we grew the transistors exponentially, computing power was only growing linearly. But in order to reason over larger and larger amounts of data — I mean, think about the unsolved problems. We talk about global warming. What if there was a catalyst that could absorb carbon? We can’t. That problem cannot be solved — the organic chemistry problem there cannot be solved. It’d take a classical computer the amount of time that has transpired between Big Bang to now. But a quantum computer can solve that. We need to go after this bold new departure of building out a computer that’s very different — all the way from the math to the physics to the computer science of it.

Listening to Satya Nadella speak, you probably won’t be surprised to learn that he has proven to be a popular C.E.O. and — whether this is related, it’s hard to say — a very successful one. Microsoft has gained more than $400 billion in value since he took over.

*      *      *

Most problems a C.E.O. has to deal with — most of the pivots they’re forced into — are related to some sort of a failure. A failure to execute or maybe anticipate a shift in technology; a failure to understand your customers. But what if the problems you’re dealing with are driven not by failure but by success? What if your firm has become so rich, and pervasive, and powerful that just about everything you do is controversial? That’s pretty much what’s been happening the last several years with Facebook, and its founder and C.E.O. Mark Zuckerberg.

ZUCKERBERG: Good to meet you.

We sat down last summer on a very warm day in Chicago — in a trailer, outside an event space where he’d just given a keynote talk.

ZUCKERBERG: It’s just very warm in here. And now that you’ve turned the air conditioning off…

DUBNER: Well yeah, that’s our fault…

ZUCKERBERG: Let’s go with it.

Zuckerberg had just spoken to a very friendly audience:

ZUCKERBERG: Thank you all so much for coming out to our first ever Facebook community summit.

The audience was made of people who act as group administrators for all sorts of Facebook user groups: new moms and military moms; locksmiths and birdwatchers.

ZUCKERBERG: I’m Mark and I’m a member of the Zuckerberg family group. I’m also a member of about five different groups for people who like the same kind of dog as I do.

The purpose of the talk was to introduce new software tools that Facebook was making available to group admins.

ZUCKERBERG: So these new tools are going to help you sort and filter all of your requests by location and gender and more so you can group them together —

AUDIENCE: Thank you!

ZUCKERBERG: And we’re also going to help you remove bad actors and all their content from your groups.

Zuckerberg combines the enthusiasm of Oprah; the salesmanship of a young Bill Clinton; the moral uplift of Joel Osteen.

ZUCKERBERG: Laurie McGonigle is here. Laurie, where are you? Laurie runs a group for disabled veterans and tries to fill in some of the gaps that our government misses. And when one of your members hit a hard financial time and was close to getting their electricity turned off, your group rallied together to help pay off their bills. It’s amazing!

By pumping up groups like military moms and locksmiths — and, recently, by tweaking the Facebook news feed to show people more content about themselves and their friends — Zuckerberg has been working hard to reshape Facebook.

ZUCKERBERG: Now we’re going to change Facebook’s whole mission, as a company, in order to focus on this.

“This” being more community and connection. Less fake news, and racism, and … maybe less unhappiness. That’s where our conversation started, in the trailer, after his keynote.

DUBNER: I’d love you to talk for just a minute about the net effect of Facebook, or social media, or social networking on, let’s say, happiness.

ZUCKERBERG: So the way that I think about this is that technology amplifies human capacity. So people — there are good parts of people, and there are bad parts of people. I believe that on balance people are good, and that therefore amplifying that has positive effects. But I think that being open is also very important for society. But it can be challenging. Confronting truths or perspectives that don’t fit with ours, don’t necessarily make our lives easier in the near term.

DUBNER: So when it comes to political or social or gender or other affiliations, how do you think about weakening the silos?

ZUCKERBERG: So if you want to have a debate where people engage productively, the first and most important thing is to first connect with that person over something that you have in common. If you just go into an internet comment thread, and you start debating gun control —

DUBNER: That’s not going to end well…

ZUCKERBERG: — that’s probably not going to be super productive. It’s easy to dehumanize the other people; think about them as not human, not empathize with them. So a lot of what social networks can do well, and these communities, are first you connect over something that you have in common. So you recognize that the other person is a person.

DUBNER: Do you try to orchestrate that?

ZUCKERBERG: Yeah, building communities is one of the ways that you can. A group might come together because they like fishing. But then they go connect over other things, and they debate other things, and they find that, “Hey, we agree on other things; we disagree on them; but now we can have productive and empathetic discussions, because we’re all people, and we recognize our common humanity.”

DUBNER: So Facebook is obviously not a government. You don’t have an army, as far as I know. Do you?

ZUCKERBERG: No, we do not.

DUBNER: But in some ways, it’s become a nation-state in the way that we used to think about nation-states.

ZUCKERBERG: It’s a community.

DUBNER: Okay. But it’s a global community, organized by interests, activities, and it’s voluntary. Nation-state is an exaggeration. But what I’m getting at is this: governments throughout history and especially now, try their best, I would argue, to help their people. And they often don’t do a very good job, because the structure of government turns out to be pretty sub-optimal and the incentives kind of weird. In a way, Facebook, it strikes me, has more leverage over how people actually organize and live their lives. The choices they’re able to make, the information they’re able to get hold of —

ZUCKERBERG: We look out at the world and we say, “Okay, we’ve been focused on making the world more open and connected.” And I always thought that that would be enough to solve a lot of problems by itself. And for some it has. But the world is today more divided than I would have expected for the level of openness and connection that we have today. So now I just believe that we have a responsibility to also work on that. There are lots of different issues and things that help bind people together and make us stronger as a whole than the sum of our parts. A huge part of that is the economy — and our jobs and all that. And Facebook is a big player there, but we’re a relatively small part of the overall world economy.

But when it comes to helping people build communities, I’m actually not sure that there are many other institutions in the world that stand for building communities and have the tools to be able to empower people at as large scale to do that. Give people the freedom to share as much as you can. Give people the ability to get access to as much opportunity as possible. We work on basic things like improving the business model of telecom operators. Or we’re designing solar-powered planes to beam down access to the Internet, because that’s a basic thing. You can ask me all the questions you want about what we’re going to do, but it’s actually going to be other people doing this, and we succeed when we empower other people.

DUBNER: Now a cynic would say, “Well, sure, it’s in Facebook’s interest. The bigger they build a global community, the bigger and better the company is.” So let’s say that someone puts on their “I doubt the do-gooder part of you” hat. How do you respond to that?

ZUCKERBERG: A lot of people just can’t get out of their own way. For a lot of companies and governments, they would do better by giving people more freedom, and they don’t for whatever reason. You may be right that it is strategically the right thing to do, but that doesn’t stop everyone from — that doesn’t mean that everyone is doing what they should do. I really want to train our organization always to think about what is the impact that we can have by giving these people more power and freedom to go do what they want.

DUBNER: How many versions — or whatever the proper noun would be — of Facebook are running at any given time? And just explain that to people who use it. And what that idea represents.

ZUCKERBERG: One of the basic strategies of our company is to learn as quickly as we can. The best way to learn is to basically try things out and get feedback. So we built this whole framework that allows people within the company, any engineer, to change some code, create a new branch of what Facebook is, and ship that to some number of people. Maybe 10,000, whatever, some small portion of the community, in order to get good feedback from that. A huge part of how Facebook works is giving a large amount of freedom to our engineers at the company, and to people who use the product to make with it what they will. And trusting people to do that, rather than —

DUBNER: Was that hard for you to get to, or … ?

ZUCKERBERG: It’s hard every day. When you’re running something, you of course have the ability to make as many of the decisions as you would like. The real art is deciding to let people do things that you disagree with, because you know it’s just going to free up more creativity and people will feel like there’s more potential to try different things in the future that may be better. You let them go do those things, even if you disagree with them. I believe a lot in giving people freedom.

It’s one of the biggest challenges for any C.E.O.: deciding just how aggressively you want to manage your employees. There is, of course, a spectrum of behavior across the C.E.O. class. Jeff Sonnenfeld told us how Indra Nooyi of PepsiCo thinks about it:

SONNENFELD: She wants to delegate authority for others to grow and allow them to make some mistakes that she knows she could have prevented, but it’s part of their growing process.

There’s also the question of just how forthright you want to be when employees aren’t performing well. Jack Welch again:

WELCH: First of all, a C.E.O. must set a mission, a direction, work with a strong team to put it together — and candor carries the day. It’s a sin that people come to work not knowing where they stand. Everybody who works for you must know where they stand, what their boss thinks about them, what the company thinks about them. Understand, “radical candor” is not cruel. It’s the kindest thing you can do to somebody. Tell them where they stand early in their careers, so they know they can adjust. And they can change, or they can move on. They can be somewhere where they fit.

And then there are employees who are highly competent — but are still big trouble.

PAO: You’ll have the star engineer on a team, who’s driving most of the progress, who might be kind of a jerk.

That’s Ellen Pao again, the former interim C.E.O. of Reddit. She signaled a culture change with some of her early H.R. moves.

PAO: I was able to hire a lot of awesome people. We ended up bringing on board a team that was incredibly diverse. There were a lot of women on the executive team, there were under-represented women of color and people of color.

But Reddit, as you’ll recall, proved to be a contentious environment for Pao.

PAO: The uprising came when we removed five of the most harassing sub-reddits from the site.

The user uprising was just one of many decisions Reddit faced under Pao.

PAO: And if you look at who can make those decisions, at the end it’s the C.E.O., because a lot of these decisions are actually quite hard.

Decisions like … what to do with that “star engineer who might be kind of a jerk.”

PAO: And maybe that jerkiness manifests in inappropriate comics, or inappropriate touching. And the head of engineering is often unlikely to do anything about it, because they’ve got deadlines they have to meet. Maybe the product is supposed to ship in a couple of weeks, and how do you fire the one person who’s actually going to get you over the finish line? It’s up to the C.E.O. to say, “No, we need to get rid of this person. We’re going to move the ship date out. We’re going to disappoint some customers. But it is important enough to the company, and to me, that I need to make this call.” And if the C.E.O. is not making that call, if he or she is allowing this behavior, then it’s going to proliferate within your company.

Pao did make the calls at Reddit. Some of them were unpopular.

PAO: At the end of the day the board didn’t have the stomach for it. And they told me that they wanted me to reach either 350 million or 500 million users by the end of the year. And I just said, “No company has done that. I can’t commit to doing that.” And that ended up being the reason why they asked me to leave.

When Ellen Pao was pushed out at Reddit, she went from being a C.E.O. to … an example. An example of a phenomenon in business leadership that hasn’t gotten much attention. The glass ceiling you’ve heard of …

Michelle RYAN: The idea of the glass ceiling is that women can’t get beyond a certain point within the leadership hierarchy.

We’re talking about something even more dramatic. It’s called “the glass cliff.”

Olga SHURCHKOV: So the glass cliff is just one of these categories of explanations for why women may be under-represented in leadership.

On the next Freakonomics Radio: only about 6 percent of the Fortune 500 companies have a female C.E.O. Why so few?

BARTZ: I don’t believe a female is ever hired as a C.E.O., especially from the outside, for the reasons that she was the absolute, number-one pick.

When a company does install a woman as C.E.O., there’s a good chance that company was already headed for trouble.

Christy GLASS: We found that white women and people of color are significantly more likely than white men to be promoted C.E.O. to weakly performing firms.

And there’s a good chance that C.E.O. winds up getting pushed off the glass cliff.

BARTZ: Well, of course it’s depressing!

*      *      *

Freakonomics Radio is produced by WNYC Studios and Dubner Productions. This episode was produced by Max Miller. Our staff also includes Alison Hockenberry, Merritt Jacob, Greg Rosalsky, Stephanie Tam, Vera Carothers, Harry Huggins and Brian Gutierrez; for this series, the sound design is by David Herman, with help from Dan Dzula. The music throughout the episode was composed by Luis Guerra. You can subscribe to Freakonomics Radio on Apple Podcasts, or wherever you get your podcasts. You can also find us on Twitter, Facebook, or via email at radio@freakonomics.com.