The Facebook Dilemma | Interview Of Mike Hoefflinger: Former Facebook Director of Global Business Marketing

The Facebook Dilemma | Interview Of Mike Hoefflinger: Former Facebook Director of Global Business Marketing
The Facebook Dilemma | Interview Of Mike Hoefflinger: Former Facebook Director of Global Business Marketing

Mike Hoefflinger was the director of global business marketing at Facebook from 2009-2015. He is the author of Becoming Facebook and currently works as an executive-in-residence at XSeed Capital.

This is the transcript of an interview with FRONTLINE’s James Jacoby conducted on February 27, 2018. It has been edited in parts for clarity and length.

First of all, what was your role at Facebook? What did you come aboard to do and about when was it?

My job was to do the marketing for the advertising side of the business. I had run the Intel Inside program for Intel which, of course, is on the buy side of advertising all around the world and Facebook was one of our important partners. They had begun to grow and to become a digital property that we wanted to communicate through. I started to realize as a marketer that there was something special happening at Facebook and so I joined essentially the sell side of advertising. And my job was to market the value of Facebook advertising to advertisers big and small, and that started in January of 2009.

Facebook’s Beginnings

Describe the atmosphere that you walked into in 2009 at Facebook.

Well, on the one hand, Facebook was beginning to build some momentum. But they were not yet the number one social network in the United States. That was still MySpace. So there was certainly a feeling of enthusiasm but it was very, very early. As a business, they had only done $270 million in advertising revenue the year before and there was a feeling that as an advertiser, if you are on Facebook it was still very experimental, as opposed to much more established advertising platforms like television and Google. And so there was a feeling that there was a lot of potential but that it was very, very early in us becoming an established place for people and companies to advertise.

What was the business model at that point in Facebook’s history?

Well, Sheryl Sandberg had arrived the year before I did and had done a very good job of coalescing across Facebook’s different teams that advertising would be the primary business model of Facebook. But at that point, it was still primarily a desktop website. It’s hard to remember now but that’s, of course, how it started because [founder and CEO] Mark [Zuckerberg] didn’t have smartphones in 2004 when he was building Facebook. And so we would do little ads on the side of the Facebook experience and the right-hand column, and some of those would actually be sold by Microsoft’s advertising sales force. And we were just trying to kind of figure out what was the ad product that we could build. We knew that there was a lot of potential, but we also felt, and more importantly, advertisers felt that we hadn’t quite figured it all out yet.

What did you see the potential being at that point time? What was it that you wanted to harness about Facebook’s power?

The fundamental thing that we saw very early and that has proven to be completely true, even though it was a lot of iteration to get there, is Facebook’s understanding of people and the amount of time that people would spend in a network like Facebook where they were fundamentally connected to other people. And so we simply over the years got better and better and better at helping advertisers target their messages effectively to the right people.

And when that happens, you create value for the advertiser and you create something that also creates value for people because they see commercial messages and communications. They are simply more appropriate and more relevant. And that, fundamentally, is what has made Facebook as big as they are. And back in 2009, we were still fumbling with that capability or that superpower and over time Facebook simply got very good at actually harnessing it for the sake of both parties in that advertising relationship: people and the business.

And what does that mean – understanding people? What’s the basis of that?

People were their authentic self on Facebook because they were connected to other people that they knew in real life. And so the things that they would do on Facebook, the things that they were communicating on Facebook, the things that they had told Facebook, everything from their name to their age to where they lived, the things that they liked on Facebook pages all were an authentic, a fairly authentic version of that person, and had provided more information, really, than any other website had been able to collect prior to that.

And so because people were being their authentic selves, because that’s what maximized the experience that they had getting connected to other people in their life, Facebook understood them in a way that allowed them to give advertisers a chance to better communicate with them. You know the story we always tell each other: [Of] the 120 million people in the Super Bowl, only 7 million of them are interested in buying a truck, but all 120 million of us have to watch the F-150 commercial. On Facebook, you can find those 7 million people very exactly. So everybody is better off. You see things that are more relevant and the advertisers being more effective with what eventually would become hundreds and eventually billions of dollars of investment.

Ads Targeting

So really, targeting is key. Tell me about that. How was targeting key at that point?

It’s two things that have to come together to make any advertising actually work. You know, advertisers, especially advertisers that are extremely good at what they do, actually genuinely care about the people that [they] are trying to communicate to because they’re trying to build a very, very long-term relationship with those people, as opposed to something that is a very quick, rapid transaction tomorrow and then we’ll move on and we’ll never meet again. And so they care about who they are talking to and what they’re communicating.

And on their side, we’re trying to craft something: a message, a story, a piece of video that is meaningful to that person, that they’ve thought about very much. They’ve thought about that person since they were developing the product. And so they have thought a lot about who they are trying to talk to and what they want to say. And so they’re desperate to find a medium that can find exactly those people. That’s the only people they want to talk to because they know that to other people, their product or service is not as relevant. And from a selfish perspective, their advertising spend will not be as effective.

In 2009, you’re coming in and there must have been some excitement about the fact that they have this trove of data there. And then the question is: What do you do with it?

The biggest question that we have to ask ourselves is not what advertising products could we build, but what advertising product would advertisers actually understand and want to use and use successfully. And so over the course of time, if anything, we had to simplify the advertising product that we’re building rather than making it more complicated. And that was probably the central journey between 2009 and the latter parts of 2011, and eventually that very pivotal year of 2012, where the IPO had arrived and revenue pressure was as high as it had ever been.

The whole story was not about building more, but in some ways building a product that was simpler, more powerful, easier to buy, had a stronger degree of communication to people, and was more effective. So it was actually a story of stripping complexity away rather than building more and more complicated things.

You’ve been in the tech world a long time, but was there some sense of astonishment at how much data there was to be able to understand people, their motivations, what makes them tick, who they are? Was that something that astonished you early on?

If anything, there is such a thing as having too much data and really struggling with the ability to make sense of it for the sake of both the advertiser and people. And in a way, we had to grow through that and actually understand how to harness, if you will, the superpower for everybody’s kind of best outcome. And that’s how eventually things came about, like lookalike audiences where Facebook was actually able, through complex [artificial intelligence] algorithms, to find people who were like customers that businesses already had. And it’s proven to be one of the most powerful tools that Facebook has and is based on a maturation of Facebook’s understanding of not only the information that it had but what to do with it to be effective for both people and business.

Creating The “Like” Button

Can you give me a sense of how important the Like button is and was as an invention?

The Like button is probably the second most important thing Facebook has ever built after the News Feeds for two reasons. One is it allowed people a lightweight way to react to what a friend of theirs had posted on Facebook, which creates a degree of human connection. It’s very light, it’s very fast. But it’s something that just shows: I see you, I see what you just said, I see what you posted. I liked it. I’m a part of your story, not just on Facebook but also in the real world. So it was phenomenally important for people [in] their ability to react inside of News Feed.

And secondarily, it was incredibly important because it allowed us to understand who are the people you care more about that cause you to react and who are the businesses, the pages, the other interests on Facebook that are important to you. And that gave us a degree of constantly-increasing understanding about people, beyond  just their name and whether they live in Michigan.

And it also adds, I’d imagine, exponentially to the amount of data of understanding those people. Right? Tell me a bit about that as a kind of revelation there.

Totally fundamental because it allows us to understand the bonds between people to a certain extent. It’s one thing that you’re connected to 150 people or 500 people. It’s another to understand which of those connections are more important to you than others. What topics are more important to you than others? Do you have a tendency to Like things from a particular person? You have a tendency to Like particular conversational topics? And as far as advertisers are concerned, who are the entertainers that you Like? What are the cars that you Like? What are the bands that you Like? What are the football teams that you Like? What are the businesses that you Like? Who are the influencers that you Like? Right?

…I like to go to the Caribbean. I love this place. Every place you checked into is yet another incremental piece of understanding. All right? So it’s not just that we understood a person in a particular way and then 2.5 billion  other people showed up. It’s that for every person over time Facebook is getting consistently more intelligent about who they are and what they would care about.

Algorithms And The Facebook Mission

And at the heart of that intelligence is a computer, something that’s processing all this. Like how does that actually work?

Well, there has to be a computer and the all-powerful algorithm, because to the human mind this is all too complicated. Once you’ve got two and a quarter billion people on Facebook or Instagram an hour, a day, maybe every waking hour they’re checking it. They have maybe 1,500 things going on in their Facebook universe every day. If we were to communicate all of that to them all the time it would be utterly overwhelming. And you would run to the hills and make a point of not using Facebook or Instagram rather than making a point of using it more or feeling like it’s time well spent. And the only way to sort through that is to try to write down the rules that we think [define] things that are important to you and things that are not important to you. And in computer science we call that an algorithm.

Explain that. So you’re basically training a computer to learn about someone. Right? [Tell me as] a layperson what that means to process all this information [that] gives you what you want.

It comes down to the same things that we care about in the real world: Who are we connected to and what are we saying to each other? That’s how we’ve operated as humans for tens of thousands of years. You would argue maybe the feeling of being connected is the fundamental human condition. And this is just technology improving the scale and speed with which we do those things. That’s what the computer is trying to do. [It’s] trying to understand who are the important people in your life and what are the important things in your life.

So of course, if I know that you’re posting a life event on Facebook, the algorithm is going to fundamentally prefer that because you’ve said: This is important to me. We don’t have to calculate anything. You’ve said this is a big deal. So that’s always going to come very high. And then we literally understand, we believe we understand in rank order the importance of all of your connections on Facebook from one to 150, from one to 500, from one to 2,000; and the more important we believe the person is because you’ve connected with them more often, you’ve responded faster, you’ve reacted more, you’ve message with them when they message you, you’ve returned that message faster – any indication that we can get that they’re important to you and, of course, if you said they’re a family member.

We try to understand who is talking. Then we try to understand what are they talking about. Over time, the smarter and smarter artificial intelligence can get, the more we understand the meaning of something that you post on Facebook. And maybe there are certain topics [like] family that you care more about than other topics. And so if people are talking about those topics, we’re going to try to prioritize that.

And by prioritize that, you mean feed them more of that?

Well, the universe of Facebook is made up of events and things that are happening with your friends or the pages or businesses that you’ve connected to, and as I said, on average maybe 1,500 things in that list every day. And our job is to find the 150 to 300 that matter the most and then order them as perfectly as we possibly can [according to] how they would interest you or the role they would play in you feeling connected to those people. And so we’re always trying to figure out: What do we throw out? What do we keep? The stuff we keep, what order do we show it to you in?

Maybe you only have 15 seconds right now, so the thing I show you first is really important. It needs to be something that is about a person that’s important to you or a topic that you have shown us that you prefer or something that all the other people around you are also really engaged with – something that has caused other people to Like something, to Like it quickly after it was posted. That’s a signal that it’s pretty likely that if the people around you engage with this that you would want to do that, too.

So the priority for the computer and for the algorithm is to engage you?

It’s probably one of the most fascinating pieces of Facebook’s evolution and maturation of the last 14 years. They’ve been profoundly good at what they call their North Star Metric – having clarity across the company about the thing that you’re optimizing for. They’ve always said that is engagement. It started initially with how many monthly active users do we have. Then it quickly became how many daily active users do we have. Then it quickly became how many people have used us six of the last seven days. It became how much time people are spending with us. Then it became are people clicking more on things [in] the News Feeds, spending more time reading it, stopping it on their smartphone.

They have always been very dedicated to optimizing for that variable and it’s what has caused them to grow in a way that’s very unaccidental. We all think, it just got big because … It got big because that’s what they were trying to do. What has happened now, 14 years later, is we’re discovering that engagement alone may not be the future of Facebook’s North Star variable. We begin to use phrases like “time well spent,” which is about not just engagement but a feeling, a sentiment that you have about News Feeds; the feeling that you get from reading News Feed; the trust that you have in News Feed. So it’s a story about Facebook being so good at what they were fundamentally trying to do that is created through their success a new problem.

Facebook’s Mission

Bring me in [to] how it felt that this North Star, this mission, is what you’re there to do, which is to actively engage users. How pervasive a mission was that inside of the company? Give me a sense of that.

Well, the one thing that Mark Zuckerberg has been so good at is being incredibly clear and compelling about the mission that Facebook has always had, which is the desire to make the world more open and connected and to build community.

And it is the one immutable thing about Facebook that will still be here 10 or 20 years from now. Everything else might actually change. And so people have a tendency to be at Facebook, to work at Facebook because that mission resonated for them and it was something that, you know, Mark doesn’t just say when we do ordered calisthenics in the morning and we yell the mission to each other. Right? We would actually say it to each other, you know, when Mark wasn’t around, when Sheryl wasn’t around, when nobody was around to test us on the mission. We would ask ourselves if this thing that we do would make the world more open and connected.

So there was always great clarity on the mission and there was great clarity in how we would measure whether we were building something that people wanted to use. And it allows you to communicate clearly as opposed to subjectively. Where you wouldn’t yell at each other in a meeting [and] say, “I think X” and the other person would say, “I think Y.” You would simply ask yourself whether it would make the product one turn of the wheel better or one turn of the wheel worse.

Is there not some sort of disconnect between a mission of making the world open and connected and just simply engaging people? I mean, isn’t there a disconnect between those two things?

If people were using the product, a lot of people were using the product, they were using the product often, then there is a strong correlation [with the fact that] you are doing a great job against the mission. So there was a strong feeling that if we were to measure engagement, if we were to pay attention to engagement, if we were to try to grow engagement very intentionally, then we would be doing well against the mission.

This is helpful …just to understand more globally what’s happened.

You know, that’s how we arrived at this moment, because they have done such a good job of being intentional about this, which I think is one of the biggest things I’ve tried to communicate in the book [Becoming Facebook: The 10 Challenges That Defined the Company That’s Disrupting the World]: This was not an accident. This is exactly what they were trying to do. And it’s also why we’re here now.

Tell me as simply as possible, what is the North Star Metric?

The North Star Metric of Facebook is engagement – the notion of whether or not people are using the product and using the product frequently. Over time, we would go from simpler metrics, like the monthly active users, to much harder metrics, like the daily active users, and even harder metrics, like the number of people that had used Facebook six out of the last seven days. Metrics like whether any particular change in the product would cause people to use it less or more. It’s simply people telling us by voting with the thing they have the least of – their time – whether this is a product that was important to them.

And that was of the utmost importance to everyone inside the company?

Everybody was clear on the fact that that’s what we were doing, and it allowed us to use common language. Were we getting better or were we getting worse? Were we building something that more people wanted to use more often or not?

…Was that a concern inside of the company, that in some way that there was a fine line between engagement and addiction?

Any goal you pick, any focus you have, especially the more successful you’re being, might over time have a consequence that was unintended or less well understood. There was really very little way for Facebook to understand the scale of impact until they had reached that scale of impact. And I think we find them today simply doing what they always have, which is to understand what the data from people is telling them and what they can do about it. Today that’s become a profound and in many ways emotional issue precisely because Facebook has reached the scale that they have.

Sheryl Sandberg

Who is Sheryl Sandberg and what was her role at the company? You came in just after her and I’m just curious who Sheryl was in the mix.

Sheryl is probably one of the all-time business athletes in Silicon Valley and had that reputation even before Mark and Sheryl began talking with each other about her potentially leaving Google, where she was running the high scale of their advertising business – very successful, it was a $10 billion a year business when she left – and coming over and being the chief operating officer and arguably most important partner to Mark Zuckerberg.

She had a reputation of being someone that people wanted to work for. Thousands of people were working for Sheryl at Google at the time that she left there.

…Sheryl had a reputation as being very detail-oriented, [having] the ability to really get things done even in complicated situations. And that is probably in no small part due to the fact that she was in earlier parts of her career in the U.S. government the chief of staff to the [Bill] Clinton-era Treasury Secretary Larry Summers. And she brought that to Silicon Valley, first joining Google, and then she began to talk to Mark about taking the next step in her career. And the two of them were trying to figure out if they would be a great match for each other.

Sheryl [was] trying to evaluate whether she would leave a giant business that was clearly doing well to take a big risk in her career, but to have greater influence over a company. Mark [was] trying to figure out whether this could be a great partner for him so that he could concentrate on vision and product while Sheryl was taking care of the business. They discovered over a number of dinners that they had [during] that holiday period in 2008, that they thought they could be very good partners, even though on paper they were maybe more different than alike. Mark had dropped out of Harvard. Sheryl was head of her class.

Facebook’s Business Model

Was Mark at that point in time uncomfortable in any way with the advertising business, the profit motive? Was there any sense of concern that the pure social network, the pure mission of connecting people, might in some way get corrupted by the profit motive?

I don’t know that it was a concern about the profit motive corrupting the business, but maybe a feeling that – great as he was at vision and product, which is proven at that point, and the ability to lead a team that would build it – he had very little experience and perhaps little confidence in building a big advertising business. They had a feeling even at that point, even prior to having built the ad product that is so powerful and effective now, that it would be the fundamental business of Facebook, and without having a business there would be no pursuit over the coming decades of his mission. And so he always knew that he had to build something that was great for people first. But he also knew that he had to build a business that could continue and thrive. He knew enough to reach out, to try to understand who could be a partner who could do this for [him]. And that’s how Sheryl arose as someone who was expert in precisely that kind of challenge.

Was there any point on the inside in your experiences where those two things seemed in conflict with one another? The advertising business model, engine of the company, and the social mission of the company were at odds with one another?

I think it’s a conversation that we would have inside of Facebook on an almost daily basis, and one that Zuckerberg stewarded very carefully and very consistently and in a very balanced way. We would always build something that was good for people first, and we would work and iterate and change until we had something that people were using, back to the North Star Metric. And we would always be honest with ourselves whether that was actually happening and when that was happening. Then we would build a product that would allow businesses to participate in that goodness. And so he was able always to be very effective in these meetings. Priority one is building for people. But we will have a business that’s going to attach to this. And that’s where the partnership [with] Sheryl came in.

Maybe there’s the benefit of hindsight right now, but in terms of this idea of building something that’s good for people and then building something that engages people, was there not a conversation about whether those two missions are aligned with one another?

I think it’s difficult to figure out the relationship between those two things until you’re well and good in the middle of building it all, which I think is the general approach Facebook has a tendency to take when it comes to building product. They know that they will not be able to predict every outcome of what they are building. But they have, either through their own vision of the future or the data that people are providing in the way that they are using Facebook, a distinct idea of a thing we should build. They then go and build that thing and then they try to make that thing great and understand whether it continues to be great.

And we’re now finding ourselves in this space where News Feed, for example, has been wildly successful, probably in some ways more so even than its parents, Mark Zuckerberg or Chief Product Officer Chris Cox, could have ever hoped for. And in ways that are so perfect for our modern smartphone life that even Zuck, who we should give a lot of credit for product design, could not have built it more perfectly: a vertical stream that never ends of bite-size information about the people who matter to you.

So we find ourselves now in a moment where we have to understand we have built something, it has been successful, more so than even we dreamt. Now what do we need to understand about it to understand whether it is as they would say “time well spent,” which is kind of an important game-changer in the evolution of Facebook? But just another point of having to improve what they have constructed because of the success they’ve had to just even get to this point.

The News Feed

So let’s talk about News Feed for a second. How important is News Feed?

News Feed is comfortably  the most important thing Facebook has ever built because it perfectly coalesced the value that people get from Facebook, which is this notion of feeling connected to the people who are important in their lives. Nothing makes that more forward in their experience than News Feed. It really [enables] navigating a social sphere, the things that are happening with the people and things that you care about inside of Facebook. Navigating without News Feed would be unthinkable. And at the same time, I think we have to say even when compared to unbelievably important media like television, print, radio, that in the 11 years or so  since News Feed’s birth, it has become one of the most important media ever invented. And I think it is profoundly impactful, as we are clearly finding now.

In your book, you describe it as a lens on the world. Explain that for me. What does that mean?

Well, there’s two reasons that it’s a lens on your world. One is that you have decided what the people and things are that are important to you. You’ve made it very clear to Facebook that you are connected to these people and not those people, to these things and businesses and not those things and businesses. So that is you saying, “This is what matters to me, and I am going to use Facebook to understand the world through the eyes of these people and things.”

And then Facebook has laid on top of that increasingly sophisticated technology that says, “We understand what you told us. These are the people and things that matter to you. We’ve also paid attention to how you use Facebook, how you react to all of these things, and we have tried to put those two things together to show you more relevant things today than we showed you yesterday.” And so it is, like any other medium, a way to see the world, but is perhaps one of the greatest ways to eliminate noise in our lives, especially our modern lives, and to view the world through the people and things that we care about.

Facebook And Filter Bubbles

…Was there concern inside the company about this idea of a filter bubble that you are only seeing through a lens of yourself and that this whole idea of personalization was in some way not connecting people but actually separating them to some degree?

Well, certainly in hindsight, it is a filter bubble by definition. You have chosen the people that you care about who are likely to be people you know. And you know those people most likely because they think, act and live like you. It is more part of this human condition of confirmation bias. We try to find our tribe, when we’ve been finding our tribes for tens of thousands of years. This just happens to be a digital version of our tribe.

So it is by definition set up from the very moment that you make connections on Facebook to people and pages a filter bubble. And so that, of course, can be incredibly engaging. We do love to see things, to hear things from people that we care about, things that are aligned with the way that we think about the world. You see something that aligns with how you think [and] you’re like, “Yes, this is great. I understand the world. The world is the way I think it is. I feel better today than I did yesterday.”

But a filter bubble may not inherently be good for the world. So now again, you’ve actually been successful, [but] if you hadn’t been successful you wouldn’t even have the privilege of having to deal with this problem. Now you have to deal with this problem. Is it worsening, the notion of filter bubble? Is it worsening the notion of tribalism? Is it worsening the notion of Balkanization? Is this digital Balkanization on top of geographic Balkanization on top of political Balkanization? And what can we do to work at the edges, to maybe expose you not to opposing points of view – because people have a tendency to reject those outright – but to a more nuanced and expansive version of the world, the ability, as I say in the book, to maybe restitch the Pangaea of the mind, to draw connections and learnings across political boundaries, geographic boundaries, thought boundaries? That potential is there and somewhere between Mark Zuckerberg’s genius in building a product and our desire as people to not narrow our view but widen it, lies the future.

There was a notion of a filter bubble that was kind of talked about as a critique of Facebook back in 2011 or so. How did those critiques resonate inside at the time of the idea of a filter bubble and that’s what you were building?

The risk of filter bubbles was just one of the many items that we had to deal with on a daily basis inside of Facebook to make sure that we were building a product that was better today than it was yesterday. And so of course, you would try to be as intellectually honest about this as you could, and filter bubbles are a real thing. And so you begin to think about what can you do in the product; what other people can surface to Facebook users that might widen their lens. What can we surface to you if you liked an article that somebody had shared on Facebook? What other things can we surface about that topic that are related that might broaden your view on that topic?

And then of course, especially in the last few years, how can we work as hard as possible to help you trust News Feed, to help you trust what you’re reading in News Feed, what you’re sharing in News Feed, without becoming an overbearing editorial presence? If we allowed a News Feed only for New York Times stories to be shared, it might be a more accurate News Feed, but it would be a profoundly less diverse experience. And so it was all part of constantly everyday wrestling with how to build a better product. But it was just item number 18 on a list of many things that we were constantly wrestling with to make a better product.

The Downside Of The Algorithm

…The North Star is engagement. So quality is not the North Star.

And over time we understood that you had to be subtle about even that. We understood that we had to go from looking at the billions of data points, literally billions of data points, that were coming into Facebook every day to a greater degree of subtlety about understanding the sentiment that people had about Facebook. So every day we would poll tens of thousands of people on Facebook about how they felt about their News Feed. Are you seeing what you want to see? And then eventually go as far as having thousands of people that were contracted  by Facebook read their Facebook feed and react to it by saying, “This is what I had wanted to see. This is what I actually saw.” Here is what we could do about Facebook in ways that go far beyond whether you liked something or laughed about it or were angry about it.

But use the subtlety of literally an army of thousands of people that were telling Facebook, “Here is the deeper subtlety of how I feel about my News Feed” beyond the raw data that we could collect. It’s all a part of the constant maturation that Facebook had for understanding its product. Initially, it was looking at the data. We have a ton of data. Let’s look at the data. This is what we like to do. And then we began to understand that was not enough. And now we understand all of those things are not enough still, and that we may need to evolve how you think about the very notion of engagement.

Facebook’s Business Model

…Sheryl’s [job was] mission impossible at the time. Right? So we’re still early on, but why was it mission impossible? What made Sheryl’s job, when she comes in, so difficult?

She had come from Google where advertising, for a number of reasons, was mission very, very, very possible as a matter of fact and still, maybe in some ways to this day, the greatest internet business because it was so obvious that you would apply the ability to advertise against the words and things people were searching for. That made perfect sense to people. If I typed ‘red crayon’ into the Google search box today, of course, I will get some what we call organic results. But I would also expect people that make red crayons to tell me, “We make great red crayons. Why don’t you come visit us?”

So it all made perfect sense. And when you arrive at Facebook, connecting to other people and hearing what happens in their lives makes perfect sense to us. Putting advertising in the middle of that is much less obvious and potentially dangerous, in that we go from seeing a story about a friend of ours who just had their second child to another story about a friend of ours that has a new job, and in between to see an ad from Ford might have been, if handled indelicately, impossible. And that’s the job Zuckerberg gave Sandberg the day she arrived at Facebook.

…I read your book and there’s the sense that there was this kind of Manhattan Project going on inside of Facebook in building the ad business. What was the holy grail of building that ad business?

The most important business decision Facebook ever made was to allow advertising in the middle of News Feed, first on desktop and then much more importantly, we know now, inside the mobile experience. And it would take years for us as a company to be good enough and to have confidence that we could build this in a way that wouldn’t ruin Facebook’s most important asset: News Feed. To be able to accomplish that, we worked for years to get to the point in Q4 of 2011 when we finally said we have enough advertisers, enough good advertisers, enough understanding about people, a complex and sophisticated enough auction that would connect advertisers to people that we can allow this to occur in the middle of News Feed.

It was absolutely an existential bet for Zuckerberg to take his most important asset, this thing that they have slaved to make successful, News Feed, and to allow advertising in it. And it would be the future of not only the total Facebook experience, but also Facebook’s business because it crucially reaccelerated revenue at a time when people were slowly beginning to lose confidence that Facebook could be the meteoric business that a gold standard like Google had been before.

…As you’re building this platform, this targeting tool that you’re bringing out to advertisers, was there ever a concern that targeting could be used by bad actors?

I think over time the more you build and grow a system, naturally you become more and more aware of what it’s actually capable of. And your dreams are, of course, in the beginning that your system is extremely capable. What you need to learn and what requires maturity in building products is to understand whether it has become too powerful or whether it has become powerful in a way that you never intended for it to be used; to have advertising that is discriminatory by virtue of its targeting, or to have entities that are using it to increase their ability to communicate certain messages to certain people.

And that’s, of course, what we’ve discovered now – [that] entities like the Internet Research Agency and other Russian meddling were doing. They were using the system and becoming extremely good at the system to further their messages and to push those messages into particularly the United States. And so in response to all of that, just another iteration of building a product, Zuckerberg and the team now have beyond doubled down on keeping the News Feed trustworthy, and as I say in the book, clean and well-lit. It may seem like a ridiculous analogy, but this is just the 11th year of keeping your News Feed clean and well-lit. For those of us who were on early, this is a more dramatically profound version of making sure you weren’t seeing too many FarmVille invites.

Ads Targeting

Were there actually concerns at the time about building the most powerful targeting tool ever in human history?

Well… that was the goal. There would be a lot of goodness if Facebook were to accomplish that. That would really enhance Facebook’s ability to be a long-term business and to pursue over decades the mission that Zuckerberg has set out. It would prove to be an important product for the advertisers. It would, hopefully, drive improved, higher-quality advertising for people. And so you’re working at your very limits as a company to even build that product. And then you discover that the engines [of] that product can be manipulated in its pedestrian form that might look like, in the early days, advertising from third-level diet plans or 18th-level internet dating sites that would use questionable visuals creatively to get you to join them.

Then over time you begin to realize we have to police the system. We have to get bad actors out of the system. And initially, those bad actors are incredibly poorly-done dieting ads. And over time those bad actors become more sophisticated. The very power that your system has draws intelligent operators from the shadows. And that is exactly what Facebook calls these people: information operations. And they have people who formerly were in foreign service and the FBI who’ve probably been spies for America who are working on the very team to prevent those information operators from being successful.

They have, at the end of the day, the perfect control to determine what you can share into News Feed and what, conversely, your friends will see in News Feed. We have to remember that structurally Mark Zuckerberg has set up the company to be in control, to essentially make the News Feed go dark tomorrow if he wants. And what he is wrestling with now is having set up the ability to control Facebook. He now has the responsibility to control Facebook.

The Facebook IPO

…Bring me into the story of what the IPO kind of sets into effect in terms of chasing revenue and really the pressure that’s put on the company to perform at that point.

We worked in an unlabeled building a few blocks away from Facebook to write the S-1 [securities registration document]. So that was quite the process. But I don’t know to what extent you care about that, but I understand what you’re saying. Well, being a part of the lead-up to the IPO on the inside was a matter of already being in the process of trying to optimize the business. We knew that we needed to complete what we called a pivot to being a mobile-first company because it was very clear that’s what people were becoming. It was mobile-first they were rushing headlong into, really not just prioritizing smartphones, but doing it to the extent of many, if not all, of the ways that they were communicating.

So we needed to rapidly finish that pivot. We needed to rapidly figure out how to build an even better advertising product. If anything, we had perhaps overcomplicated the advertising products of those years 2009, ’10, maybe even the early part of ’11. We knew that more could be done. We knew we could build a better product and we were in the midst of building a better product. The world didn’t know that. The world just was wondering when this $100 billion IPO was going to happen. And it was, of course, an enormous amount of pressure because Zuckerberg had held off on doing this IPO because he was in those days concerned about the undue pressure that it might put on the way that Facebook made its decisions.

And then of course, we announced the IPO early in 2012. And the scrutiny became absolutely huge. And especially the final days and weeks leading into the IPO there were some not great events. Facebook had to amend its own S-1 and say that if we couldn’t make the pivot to mobile, there might be significant negative impact to revenue. There were very big name advertisers including GM that had said they were beginning to pull advertising from the platform because they weren’t 100 percent sure that it was effective.

And that was in the very days leading up to the IPO. So while internally we are trying, as Mark Zuckerberg would say, to stay focused and keep shipping the very products that would ensure business success, the pressure externally was mounting to almost inconceivable levels. It then crescendoed in ringing the bell and having the IPO, but having a very mixed at best first day, including gigantic chaos in the Nasdaq’s ability to actually clear the first few hours of transactions, which creates an enormous amount of uncertainty in the market and therefore bad news. [That in turn] winds up being the first of over 100 days of pretty [relentless]  declines in Facebook’s stock price and valuation.

Internally, what was that like? What was it like for you in those days? I mean, the pressure and the incentives to push through.

It was a very mixed bag because on the one hand we had been working since maybe the middle of 2011 to make those crucial evolutions we knew would need to happen: the pivot to mobile and of course, the simplification, the improvement of the advertising product and especially allowing the advertising product in News Feed, including mobile News Feed. So there is part of you inside that knows that we are doing the right things that will produce the outcome that will rebuild the confidence in Facebook because it is going to have great results. We knew if we could scale this, that would happen.

But of course, you are seeing the external perception of Facebook suffer. When you go down from a $100 billion in valuation to $50 billion in valuation, it allows people to conclude that maybe Facebook isn’t going to be a thing; that we all knew that it wasn’t going to be this big; that it’s not a great business; that worst case, Zuckerberg’s desire to build this mission over the next few decades might not come together. And so it was a perfect example of Zuckerberg being able to keep everyone internally a little bit calm, level-headed.

Another version of one thing he would say to us almost every Friday in the company Q&As is: “It’s never as good as they say and it’s never as bad as they say.” He is a very even-keeled CEO and we saw that from him every Friday in the Q&As in ways that were honest, and he would never shirk a hard question. We knew that the game plan that we had is one that we could have confidence in even though from the outside it looked really, really not good.

Was there at that point when you start to become accountable to shareholders, when you start to have to issue quarterly reports, quarterly earnings – does the company change fundamentally at that point?

What’s fascinating is that for the longest time Zuckerberg especially, but also all of us because we cared a lot what Mark thought [and] felt about the future, felt that the IPO could be very damaging to the way that you make product decisions, business decisions; that it might hurt our ability or even desire to build things for people first and then to build the related business product. What we had to go through actually had again been in play for several years before that. Between the way that Sheryl was running the business [and] that our CFO at the time, David Ebersman, was running the financials of Facebook, we already were in command of how we planned, how we grew the business, the discipline that we had for it. The seriousness with which Sheryl approached the growth of the business. It was almost as though we were a little public without knowing it.

And then what we discovered, and I think especially Mark in the years since the IPO, is that the discipline to stand up in front of the world every quarter and tell them what we’re up to – what we’re trying to accomplish, how that’s going, how things are improving, what products we’re building, whether those are causing success for advertisers and “sentiment positivity”  for our users – gave us additional discipline and the ability and desire to make decisions even cleaner even faster. And in some ways Facebook, and I think Mark at the very top, felt that we were a better company after the IPO even than we were before.

Was there concern that it was growth at all costs? Was there concern that growth and this kind of missionary zeal could blind you to the darker sides of this medium, of this technology?

What was fascinating from the early going – I still remember these meetings with Mark back in 2009 when we were still stumbling our way toward ecstasy on the advertising product – is that Mark was able to exquisitely balance the fact that, first and foremost, we needed to build something that was great for people across the board, something that makes the world more open and connected and builds community. And then we would build the business for that.

And we would be in meetings where we would communicate to him about the desires that advertisers had for new advertising products, products that were more expressive, more visual, more dynamic, and eventually even in the feed, years before they were allowed to be in feed. Mark would always take a balanced perspective which is, “We’ve got to build this great for people first. If I agree to this advertising product now and it doesn’t work with people, then we have to pull it back. And all we’re going to do is confuse these advertisers.”

So… it was always clear that we would build for people first and then we would build the business and that we had proven to ourselves that we could do both of those things and do them well.

So it was never growth at all costs. And even in the very algorithm that drives Facebook’s advertising auction, we don’t choose the ad that is willing to pay the highest price. We choose the ad that is most likely to create the best outcome for businesses and people. So internally, I’ve seen Mark [and] Sheryl, over and over again, pass up or delay immediate-term revenue-goosing for the much longer success of the business but also mission.

There’s a notion that really the user is the product and the customer is the advertiser. Do you agree with that?

Well, Facebook is a complicated marketplace. In some ways it’s three sides. We’ve got people that use Facebook with other people that use Facebook, even a marketplace between people. Then of course, there are businesses, advertisers’ pages, whether they are investing in advertising or simply communicating for free. And then there is Facebook itself, and that is an incredibly complex dynamic of making sure that you’re taking care of all three of those players. And you know, as we’ve said before, if we don’t take care of people – and we would say to advertisers directly like this – if we don’t take care of people first, the party ends for absolutely everybody. And so you know, taking that care, that stewardship was always central to any kind of additional success we wanted to build.

Ads In The News Feed

…So situate us back, for instance, [to] when advertising starts to get into the News Feed. Did Facebook research whether its users could tell the difference between an ad and content or news or something that someone had posted or something that a news organization had put out there?

The amount of intensity that we put around measuring the advertising product was huge because that is probably the only road to success with advertisers, [is] to go from being something that somebody would experiment with to something that an advertiser is committed to over the long term. … The only way to cross over from the bucket of experiment to the bucket of commitment is to prove to the advertiser that their communications are effective; that people [recognize] the difference between a commercial message and something from their friends; that people changed their opinion, their emotion about products and services when they saw things; that people would click on things; that people would visit stores; that the company’s business will be improved; that the brand metrics of a company that was advertising would grow.

So we had to measure the effectiveness and the effect of advertising to within an inch of our lives to prove that to advertisers and to communicate to them what was happening here. We were discovering things like in advertising – which we used to think that if we can’t show somebody a 30-second TV ad that we can’t really influence them to those degrees – all of a sudden, we begin to realize that even three seconds of seeing something visual in your News Feed on your smartphone could be incredibly effective, could make a difference.

…So you think most Facebook users can actually tell the difference between something that’s paid for and something that’s not paid for?

Certainly, when it comes to the way the News Feed is designed, that’s made very clear. It’s labeled very specifically as a sponsored [ad]. But beyond that you are trying to find a communication between that business and that person that is authentic and that fits with the style of how News Feed works, the way I experience News Feed. You don’t take a straight-up television ad and dump it into the middle of News Feed. It simply won’t work the way that it does on television. People won’t perceive it the way that they do on television. They won’t be influenced by it the way they would be on television. And so you’re constantly trying to educate advertisers about what effectively communicating in Facebook’s News Feed would look like. And it [means] actually changing the approach that you’re taking, making it feel casual, organic, making it short, pithy, making [it] communicate something about that advertiser, as opposed to just driving the sale that you’re having on the car lot that weekend. That dynamic of what is effective inside of Facebook’s News Feed is to communicate organically in a way that feels familiar, while still, as you say, actually distinguishing that paid message from an organic message.

And that was something that internally you guys felt confident that your users were being served well in that they could tell the difference between something paid and something that wasn’t paid for?

We did. And as with anything else in the product, even that constantly evolves. And we’ve seen just in the last year or so, Facebook now getting more serious about being transparent about what ads, certain pages, certain entities inside of Facebook, are paying to put into the News Feed. As part of this investigation [of] manipulation, they are now making all ads that advertisers are putting into the News Feeds transparently findable  by users, so that you’re very clear this is [an ad], whether it’s a car company or somebody who’s trying to manipulate our elections. These are the messages that they are paying for. You can see them very transparently right here. Everything is clearly identified transparently. And what we’re discovering is that the way you do that, the way you keep this all clear, also has to evolve over time like the product itself. So I think Facebook realizes that the more successful you become, the more responsibility you have, and the more you need to think about whether your product continues to deliver on the trust and promise that you started with many years ago when life was simple.

The Arab Spring

…Arab Spring is 2011 or so. You’re at the company. Arab Spring comes along, it happens and it’s dubbed the social media revolution. What was that like inside?

Facebook employees have a tendency to be at Facebook because prior to their arrival they felt connected to the mission that Mark was trying to drive of making the world more open and connected. And we would see that happen, of course, in some very small ways and some medium-sized ways. We would see it in our own lives: the way our News Feed behaved, the way that we felt more connected to people, the way we understood moments that were happening with people that otherwise we might not have been privy to – some of them maybe just pedestrian, some of them entertaining, some of them maybe profound, some of them sad.

And so we knew that this was working. Then slowly but surely, even before Arab Spring, you would hear stories about how these very large, very important communities would organize using these tools, often primarily Facebook but certainly also things like Twitter. I mean, we heard about a single person opposing FARC [Revolutionary Armed Forces of Colombia], the [rebel] movement in South America, by just using Facebook to organize these absolutely giant protests of hundreds of thousands of real people in the real world opposing this dangerous [rebel] movement that was taking hostages in the streets in public, and being able to use a tool like Facebook to organize that movement. Of course, it’s their actions that are causing the change. Facebook is just a tool that they were using. But all of a sudden, it was becoming clear that something much bigger was happening than just finding out important and sometimes profound things about our friends, and that the world was changing because of the tools Facebook was building.

Then along comes Arab Spring, a mega version of this ability for people to share their point of view to build common momentum and to organize themselves in the digital world so that they could protest in the physical world. And of course, to us it was extremely humbling because we weren’t the ones out on Tahrir Square risking our lives talking about what was important to us. But we had built a tool that helped these people do that. And maybe nothing was more important in understanding that this mission was so real and so valuable than seeing these people using these tools to do these things. It was incredibly humbling internally.

So it reifies this notion to you inside the company that this is a tool for good.

Which doubles your desire to not only continue to build a better product but to build a great business that can give that product oxygen for decades. If we can help these people do that, imagine if we could do it for a billion more people. Imagine what might happen, as Zuckerberg has [said] himself, if we could even help connect people to the internet who aren’t connected to it today. And so he would stand up and invest hundreds of millions of dollars in connecting people who are almost guaranteed not to make him profitable.

You begin to get into this virtuous cycle internally where you [feel] like: I felt pretty good about this mission. Now I feel even better about this mission. And I see Zuckerberg doing things and planning for the future in ways that I hadn’t even understood before. So now I’m even more committed to the mission. It gets a degree of internal commitment to the product and an effort that is beyond what you would see at an average company. Nothing wrong with average companies, but the effort that people put in, the willingness to stay with these impossible missions, seemingly impossible missions, is what allows you to eventually break through. At Facebook that came from the mission the same way that at Apple it comes from wanting to build something great for Steve Jobs.

Was there ever a point where there were questions internally about this mission being naive optimism?

I think the short answer is completely yes and I think that’s why we loved it. It became clear to us especially in a moment like when we crossed a billion monthly active users for the first time.

I remember so many of us were looking at each other and our primary emotion was: I can’t believe we did that. And the way I recall Mark at the time, I remember thinking: Mark always expected to get to a billion and now we’re going to go from there. Mark always expected to actually be much bigger than this, and when he says he wants to make the world more open and connected, he’s not just talking about 150 million people, like it was when I joined, or a billion on that ridiculous day where all of us [thought]: How did we get here? But the two and a quarter billion that it is now. I don’t think Mark is going to stop until he gets to everybody.

The Evolution Of Facebook

…I asked you about naive optimism though because naive optimism can blind you to the downsides. This ethos of “Move Fast and Break Things” and not really worry about the consequences, are we not dealing with that? One of the critiques of the growth of this company is that naive optimism blinded the company internally and this mission, this almost evangelical mission, blinded the company internally to very serious risks along the way.

The story of Facebook, more than anything else, is also the story of Mark Zuckerberg and the joint maturation of both of those things – Facebook as a company and entity in the world that has position and power and therefore responsibility, and Zuckerberg as perhaps an unequalled leader with unequalled power and the maturation that he himself is going through. The notion that maybe there was a point in time when you were locked in a hacker house in Palo Alto [and] were saying to each other “Move Fast and Break Things” is awesome and fun and motivating, and a time where “Move Fast and Break Things” is disorienting to people and all the things you’re trying to accomplish in your very mission. And you stop saying that, and you move on and you say other things.

You even evolve the mission to include the notion of community, rather than just openness and connectedness. You understand that having people use Facebook a lot may not be the best way to measure whether Facebook is good. You understand that it might be time for you to travel around America to understand what is going in Louisiana and Minnesota as opposed to just Palo Alto, California. You understand that you are helping people change the real world in real and dangerous ways. And in the last year and a half, you also understand that you’re giving some very bad people some very powerful tools.

How does that make you feel, that you helped build that powerful tool that bad people used?

I’m personally a techno optimist. What I mean by that is I think the world, net-net, is profoundly better off when we build new things faster that make our lives easier and richer. I felt that very strongly when I was at Intel working for Andy Grove and we were building the microprocessors that make computers possible and arguably the very foundations of the internet. I felt that very strongly when I was at Facebook, a tool that is for us entertaining, human, profound, and dangerous, which by the way, just describes people. Everything I just said describes Facebook the same way it describes people and the fundamental benefit and risk of Facebook is that it is the single greatest tool for people in the world.

It is in that sense better than anything Google’s ever built, better than anything Apple has built. And with that simply comes the notion of having to deal with what you have built, and trying to continue to build something that is undeniably positive. What I believe is that if we never try to build these things, we never get to have these moments of crisis. And having been in Silicon Valley now for 25 years, I believe in one thing above all, which is that we will keep building. We will never stop, and we’re always concerned whether what we have today is one day better than it was yesterday. The thing that’s so rare about leaders like Mark is to have total control over something, to have the willpower to execute that something, but the ability on any given day to listen to other voices or to even understand that what you used to believe or what you used to understand may no longer be true or may have been changed by the very thing that you built.

Political Advertising And Targeting

Facebook at a certain point makes a push into politics and to helping political people running for office to use these tools to communicate with potential voters. Were you a part of the expansion of Facebook into politics at all?

Not directly but it was advertising, like all other advertising, and that comes down to whether or not you have something relevant to communicate to people and finding those people so that you can communicate it to them. Whether you are American Express or a dentist in Idaho or someone running for a local government race in Arizona, it’s all a matter of finding the people who are important to you and no one is better than finding those people than Facebook. No one is better at giving you access to those people more often on more devices for more time than Facebook. And so it naturally extended to politicians trying to communicate the reason that they should stay in office, the reason that they should come into office.

Many politicians as well as small, medium and gigantic business have used Facebook for that very purpose and succeeded, and because of their success made the world better in some small ways. But of course, you’re also going to find people who might get profoundly good at using that system to spread questionable information, false information, incendiary information. And in this last election cycle, it was a shock to find out that the Republican team behind these efforts appears to have been, depending on how you measure, more competent and able and more committed than the Democratic team, especially coming off an election both four and eight years prior to that where the Democratic team essentially ran circles around the digital Republican team. So [it’s] kind of an amazing example of what happens when you become effective at using a tool, and [it’s a] question of whether that [tool] is good or bad or just is.

Was there a question within Facebook as Facebook made a push to help promote their tools to political leaders or people that are running for elected office globally, including places that did not have a tradition of democracy, was there a concern inside the company about giving that powerful tool to people running for office?

Initially, it was felt that it would be extremely positive to have powerful tools available for politicians to use effectively to find their constituencies or their desired constituencies and to effect the change in their political career that they wanted to effect, because we believed in our product. We knew the product was successful in reaching people and communicating with them. So there was this feeling that there was a strong degree of contribution to the democracy if we allow politicians of all sizes, by name – definitely it doesn’t have to be Trump or Clinton – to communicate to reach the constituencies that they had and to grow the success of their political career. And then you began to think about what happens if these tools are used for the darker parts of political campaigns. And we don’t just need Facebook to tell that story. We have decades and decades of attack television advertising that shows us just how dark these corners of the political campaigns can get.

They’ve now simply been translated over to this particular medium and then accelerated by certain entities’ ability to use that system very effectively. And all of a sudden internally you have this feeling that our job was to help make people successful communicating their message on Facebook. And it didn’t matter whether you were Democrat or Republican, Libertarian, some party we’d never heard of, whether you were an advertiser for consumer packaged goods or autos. Our job was to try to help you be successful in communicating. And then you began to realize that we had to begin to have a point of view about what you were communicating, and that is the dawn of an entirely new era of defending the trustworthiness of the Facebook News Feed.

When did that dawn on you?

I think we’re in the process still of understanding what all of this means and we’re seeing Facebook very publicly talk about: This is our game plan. Here are some things we’ve already done. Here’s how that went. We see them, as Mark has done since he was 12 years old, building things. So we see them last year moving towards marking articles that are being shared on Facebook as disputed if certain fact-checking organizations deem them to be disputed. And there was a feeling that this could be very, very useful – that if we informed people that the link they were about to share was disputed by a number of third-party fact-checking organizations that this would be tremendously useful to people and they would stop spreading bad information. Instead Facebook learned that marking something as disputed caused it to be shared more. This, of course, [is] a perfect example of confirmation bias in massive action, and so that’s not going to work. And now Facebook has told us: OK, we’re not going to do [that] anymore. That didn’t work the way we thought that it would.

Now we’re going to actually begin to get into the business of grading the reliability of certain news organizations. And we’re going to work with all of our users to actually understand who you deem to be reliable and who you don’t so that we can slowly throttle the News Feed and to eliminate fundamentally sourcing this information and spreading this information.

Assessing Facebook’s Response

There is a sense that it’s been years that people have been writing and shouting about the fact that News Feed has had hoaxes that propaganda spreads that internationally, elections have been meddled with by the Russians and others. This has been known. So why wasn’t the company more proactive? Why wasn’t it more responsive earlier on to a lot of what was already known?

I think since the dawn of News Feed, Facebook has understood that they need to keep it clean and well-lit. And that goes back to eradicating nudity, hate speech, violent content. They’ve had very, very large teams numbering in the thousands that have been doing that for the better part of a decade. And the evolution of what needs to be done to keep News Feed clean and well-lit has simply become increasingly sophisticated. I think Facebook over time has always wanted to have a relatively light touch in being editorial with what can be shared in News Feed, other than those things that are definitively bad and offensive; obviously among them, nudity, hate speech, violence, things that are against the law in the countries that they operate in. Those are relatively easy to target to be eliminated and taken out so that we can trust each other in this community. But then there are some areas that are much grayer. And the notion of shutting down certain sources, the notion of actually not allowing certain things to be shared, or the notion of actually suppressing entirely those things in the way that the algorithm selects for News Feed, puts Facebook in an editorial position that I think they have consistently been very nervous to take up.

I think that’s probably one of the most profound changes in the last 15 to 18 months since the 2016 election, which is Facebook’s understanding that they are going to put their finger on the scale of what can be shared into News Feed. [They are] going as far as saying in the last few weeks that clearly not all news sources are created equal, which is as much as they have ever said about the fact that they are about to really tighten down what can be shared in News Feed.

Bring me inside. In the past, were you part of discussions where you saw that reluctance to want to put their finger on the scales of becoming editor to some degree?

I mean it’s [brought up] at meetings that happen almost every day, because it is all about what can be shared into News Feed and what is selected to show you in News Feed. It’s arguably a 24/7 conversation inside of Facebook. Sometimes it’s simple acts like life events are really important, we should make those a bigger deal. Or hey, not everybody wants to see baby pictures all the time and we need to maybe hire 1,000 people to give us a better sense of how we can find good things for News Feed.

Then it becomes much more profound things like: Do we allow paid messaging, advertising in News Feed? You sit in these meetings and it’s every day, 24/7, a defense of the trustworthiness of News Feed. You’re seeing that generally for the world to be more open and connected, you want to take an attitude of allowing as many things in News Feed as possible. The dark version of News Feed is that the Egyptian News Feed prevents the points of view of civil opposition from surfacing the same way that we interdict Internet Research Agency content to happen in the U.S. That will be the darkest version of getting too controlling in News Feeds.

So you see in his conversations, his Q&As, meetings, Zuckerberg wrestled with: If we want to try to make the world more open and connected, we should allow for content to flow. But now that’s clearly no longer… That’s too simplistic a way to look at the world. And you see him…

So there weren’t conversations inside Facebook about the quality of the news and information on the platform? I mean this is becoming one of the largest if not the largest distributor of news and information in the world, essential to the functioning of democracy. Right? How is it that there weren’t conversations about the fact that the quality of the News Feed and the news and information on News Feed was not paramount?

If anything, it’s a case of the conversation about News Feed never stopping. It’s Facebook’s most important asset. So it’s simply a notion of: What does quality mean? Many years ago the problem was simple. Terrible ads by people that are trying to take advantage of consumers have to be eradicated. Nudity has to be eradicated. Those may be complex problems but they’re very easy to solve. But what quality means now is becoming much more sophisticated. And the notion that what is working so beautifully – which is news, information, links being shared in Facebook had to be something that we spent more time on – dawned slowly.

Facebook began to even build a part of this organization that interfaced with media companies, from The New York Times at the very top to much, much, much, much smaller, even local news at a lower scale. And so you saw this notion that if we want to defend the reliability of the trustworthiness, the sentiment that people have about News Feed, we need to work on that every day. And what quality means is changing.

So you see Mark Zuckerberg sitting in that [Techonomy] interview  [with David Kirkpatrick in] November 2016, being asked about whether it’s possible that Facebook influenced the U.S. elections, and answering with data, as he is prone to, because it is to a large extent truth, at least in the eyes of a product builder. Literally using the word that it might be “crazy,” quote, unquote, to think that Facebook affected the election if only this very small amount of things that were shared in the News Feed were hoaxes or what we call now fake news. Then you literally, in front of our very eyes, see Zuckerberg’s notion of what this really means evolve, as he tells us on Facebook what he’s been thinking about it, how he is changing his mind, how the game plan is changing, how many more people they’re hiring to protect the security and trust of News Feed. What’s amazing is it might be hard to see from the outside. But that is a very honest reflection of precisely the conversation that he is having in a secret meeting room inside of Facebook.

Was it crazy to suggest that this influenced the election?

I think that moment is actually a fascinating study in Mark Zuckerberg. The word crazy has a lot of connotations. And to someone who was beginning to feel maybe on the outside of Facebook that there might have been real manipulation going on here, Zuckerberg saying that it was crazy was a value judgment of how people were feeling about Facebook. What I believe Zuckerberg actually meant in that moment was a very rational analysis of how much content inside of Facebook actually came from these questionable sources.

That’s what he meant by the word crazy. But to me it’s one of the most pivotal words in the entire history of Facebook. And he himself has come to understand both how his use of that word was perceived as well as the much greater responsibility that he himself has and had as of that very moment in the role that Facebook plays in our lives. It was probably one of the most profound pivot points in his entire tenure running Facebook.

What did you think of that moment in time of [then-Facebook General Counsel] Colin Stretch and others appearing before Congress?

As I was watching that, I was going back to all the moments that I had spent with Sheryl Sandberg in particular during my career at Facebook, remembering now that Sheryl, of course, grew her ability to manage chaos and complexity by working in the very government that was now casting a very, very serious analytical eye on Facebook. I was imagining the seriousness with which she was treating this because it is absolutely vital to Facebook, to the health of the mission and also secondarily to the health of the company as a business, that they stay ahead of any regulation that might be passed down because we as a nation, a democracy or a government don’t feel that Facebook is doing enough by itself.

Precisely because Sandberg has experience in the government and understands how this works, she is going to put an extra degree of attention on Facebook regulating itself, understanding itself, and stepping [up] to its responsibility before the government does it in a way that might actually lack understanding of the technology, lack of understanding of what Facebook is capable of, and lack understanding of the second- and third-order effects that could occur. So as I was watching Stretch himself in those meetings, I was simply thinking back on what was Sandberg [was] doing right now to lead the company in this crucial moment. That was, again, existential for them, like so many moments before, because she has that additional benefit of having operated within the government and understanding how these two pieces have to come together.

Did you think that they were forthcoming? Did you think that Facebook was actually forthcoming in that hearing?

To me, Facebook had pivoted from perhaps a degree of rejection of how great their power could actually be to accepting that prior and significantly prior to those hearings. They began to move quickly in their security organization to identify the sources that were bringing misinformation [and] the scale of that misinformation.

…Knowing these players, having been in many rooms with Colin himself, certainly Sheryl, Mark, to me it felt as though they had sprung into action just in time and begun to be transparent before they stepped into those rooms. I think also at the time, [they] knew that no matter how effective they might have been in those hearings, that much more work was ahead and that they would have to do more than even what the government would make them do to get to the other side of this issue.

Your Data On Facebook

…How much does Facebook know about us?

Well, of course, [it] all starts with the data that you’ve explicitly given Facebook – your basic personal information, where you live. And then of course, what you do on Facebook every day – the things that you like, the people that you’re connected to, how you react to things. The things you read, the things in your News Feed that you stop with your thumb, you spend a few seconds on. The things in your News Feed that you skip. The things you’ve clicked on in the past. Stories you’ve Liked. Those are things that you give to Facebook directly.

In addition to that, Facebook has the ability to understand and pair with the data they already have about you how you traverse the internet. They have the ability to bring in information that businesses have about you. You’re an existing customer of a certain business [and] that business can upload that information to Facebook [and] combine it with the information that Facebook has about you. And then there are data brokers in the world that understand other things about you – the cars you may have in your garage, the banks that you have credit cards with.

That information can also be matched to the information that you already have on Facebook through your phone number, your address, your emails. That all builds a composite of things that Facebook understands about you and some of that information is extremely useful in understanding you well and targeting communications. Some of that information is not as useful as one might think but it’s really a compendium of many things that Facebook understands about you.

When you say “understands about us,” does having all that information make us more manipulable  because you know so much about us?

What it does is it allows someone who wants to communicate with that user a better or deeper picture of who that person is. I was making the analogy earlier that on the Super Bowl television broadcast, all 120 million of us have to suffer through that F-150 commercial whether we want to buy a truck or not. But there is data in the universe, other data sources, that can tell you when somebody bought a truck or whether they’re in the market to buy another truck, and that data can be matched to you on Facebook without giving away your individual information to anyone.

But that way, if Ford wants to advertise the F-150 on Facebook, they know that they’re only talking to the 7 million Americans that are in the market for a Ford truck. So it allows that communication to hopefully be of greater value between the business and the consumer. Does that in the end make it more manipulable? Not to a company like Ford.

Ford would never get caught in the act of attempting to grossly manipulate its customer because it would ruin that customer relationship for life. The hard part is always at the edges, in the shadows, where there are operators who care maybe about manipulating one particular thing in one particular moment of time, and don’t care about you as a person in the longer term. So it’s usually nonbusiness interests because a business itself is way too incented to take care of its customer relationships to manipulate them.

But I guess one question I have is by giving up so much data and by Facebook having so much data, does that not leave the users more vulnerable to manipulation?

The reason this is such a complex and, I think, emotional issue for all of us isn’t just that it’s touching the very notion of democracy. But I think it’s a question about where does responsibility lie. Does responsibility lie with the medium through which the message came to you? Does responsibility lie with you, the person who is evaluating that information? Who is responsible for how you think and what you believe?

…One real question is: Was this targeting mechanism, was this advertising platform, a manipulation machine to some degree that has been exploited by bad actors? That is a very basic question that everyone has to ask right now.

…When we’re communicating with our friends, we’re trying to influence each other or we’re trying to help each other. We’re trying to share things. And so of course, the very success of Facebook is its ability to communicate and influence each other. And so you have to say it definitively allows for influence. And that is part of its beauty, but also part of the danger, especially at this scale and speed [and] especially for those informational operators who have become very good at taking advantage of the platform.

…I mean internally when we were asking ourselves whether we could put ads in News Feed, we would be in meeting after meeting where elements within the product development team who were looking to constantly defend the user experience, exactly as they should, were fundamentally worried about commercial messages of any kind inside a stream that was so profoundly human.

Exactly as they should, they took the side of the user and the business side of Facebook. Sheryl’s team, my teams are teams that were working with our customers. Were looking for the opportunity for advertisers to get the same rights and privileges and opportunities to communicate with people the way that people were communicating with people. We knew that if advertisers were not communicating effectively with people, they would not be successful as a business.

So we trusted that our advertisers would be effective when they were communicating with people. And so we had these two positions, both equally valid, one coming from the perspective of always defending the user; the other coming from the perspective of trying to build tools that would be successful for advertisers. We would have very heated debates and it was eventually Mark who had to decide when the moment had come … that we could feel good enough about the ads that we were selecting to put in front of users that they would be perceived almost as useful, as entertaining, as the content that we were receiving from friends. And like everything else at Facebook, in those meetings it came down to Mark’s understanding of the situation and his willingness to be responsible for the decision that he made because he was the only one making that decision.

The Facebook IPO

[Tell me about the IPO.]

The pressure heading into the IPO, of course, was to prove that Facebook was a great business. Otherwise, we’d have no shareholders.

Mark probably up to that moment had never been under greater pressure to allow a business-oriented element in the product so that we could grow and re-accelerate the business. He had to blend in the middle of those meetings a desire to continue to build something great for people, a desire for his product organization to always know that he was looking out for people first.

But also the notion of making the business great and allowing it to accelerate and become increasingly successful with our customers. We were seeing that ads in News Feed in some cases were 10 times as engaging as the ads we had before. We knew we were sitting on a truly effective medium. But he had to ignore as much as possible the financial pressures, the public pressures of an IPO and make a decision that was right and pure for Facebook, independent of the expectations that come from a $100 billion IPO, the third biggest IPO in the United States at that point in time.

I thought there were stories, anecdotes that you were going to share.

I mean, that’s what the meetings were like. We’re sitting there talking about how good this product is from the Sheryl universe, roughly like the Jets and the Sharks fighting with each other. But the product guys were like: “We fundamentally can’t do this. We’ve never done this, we can’t do it, it can’t work. You know it’s going to be terrible. Right?”

That was what the meetings were like?

Yeah.

Really? Like the product guys are saying this is not possible?

Well, it’s not desirable, right, is what they were saying. And again, they were coming from a position of representing users precisely as they should. Right? Sheryl, myself, others were coming from the perspective of representing advertisers; we’re trying to grow a business here. The business’s job is to help the mission. Right? Can we find common ground? And it was Zuck as Supreme Court justice. Right? He has to arbitrate this thing.

And what decides it?

Well Zuck does. Because Zuck decides everything. …Zuckerberg is like all these mythologized leaders of consumer technology companies – an autocrat. He is both structurally and conceptually the only one at Facebook who makes a decision.

Facebook And The Distribution Of News

And one decision that Facebook made was to go whole hog into the news distribution business. And you were there during that period. I’m kind of curious what the motivation was to introduce much more news into the News Feed.

Well, if anything, it comes from the same place it always does, which is we were seeing that that’s what people wanted to do. They were sharing news with each other. It was beginning to get to the point where people felt that Facebook was the internet because it was their entry into the internet, the way that 10 years before that we thought Google was the internet because that was our entry into the internet. And so we recognized that this is what people wanted to do.

So we had to make it more possible. We had to make it easier. But more importantly, we had to have the media companies themselves feel that having this content be shared would be good for their businesses; that they would want to have more of it happen; that we would make it easier for them to do so. And so we had to go from essentially having no team that dealt with media companies to having a very large team that dealt with media companies; to make the flow of information easier but also the business of media and news easier to carry out over Facebook.

And was there a concern at that point in time that in some ways Facebook would then essentially own the news distribution business and the responsibility of what that meant to be the world’s largest distributor of news and information?

That is the effect that it had. And [that] was the reason that we needed to build bridges with media and news companies so that this could work effectively for all parties involved. Right? That media companies could share inside of Facebook; that people would see [their] content inside of Facebook; that Facebook would be more successful because of that. But also that it would not cause the media [and] news companies’ businesses to be negatively impacted – if anything, that it would be positively impacted.

And we see now with these recent changes that Zuckerberg has reduced the amount of content that can make its way to people from businesses or news media companies, and has amplified the amount of content that comes from my friends that I see in my News Feed. So he has had to make a decision that was all about people feeling better about their News Feed, all about more trustworthy content being in the News Feed, that has unequivocally damaged the distribution of these important partners that he has spent years building relationships with.

Was there a sense inside about the idea not just that your users wanted news in their News Feed and were engaged with news, but that, wow, we’re inside Facebook? We have all of this data on all of these people, and that, in fact, we know more about the audience of one publication or one television channel than it actually has on its own audience.

What was actually very important is that as we were building those bridges to the media companies they would have exactly the information that we had about the people that Liked their pages, that Liked their content. So we spent an enormous amount of time working and building tools that the average person really never sees – analytics tools that allow these media content companies and businesses to understand who are these people that are reading my content.

Not only did we want to inform the businesses, that was very, very valuable.

…In many ways the analytics that we were able to provide to businesses and the media companies was a greater understanding of the people who were reading their stories or consuming their products and services [than] even those companies had themselves. So it was very important for us that they would understand who these people are. That helps the business. But it also helps that business or that media information company craft things that are more thoughtful about the audience that they had on Facebook.

In one respect [it] also plays into this idea that it made it harder for media companies or content providers to be up the middle on things, but they needed to do things that maybe played to the biases that the algorithm helps to perpetuate.

Yeah, it’s a little bit of a chicken-and-egg problem. Like, every business or publication has an idea of the kinds of customers – readers – that they want to serve and they have an editorial point of view that’s aimed at those readers, some down the middle, some right, some left. And that’s just on media alone. And so it’s a little bit chicken-and-egg. Right? The goal, of course, was for these tools to make it more possible for these media companies or businesses to do exactly what they wanted to do and to attract the audiences that they wanted to attract rather than to communicate to anybody that would show up.

Was there a recognition of the danger that an algorithm gives a structural advantage to something that may be more sensational, that may be more partisan, that may be more incendiary? Was there a recognition of the danger of that, especially when it comes to the distribution of news and information? I mean, that’s a basic question.

There was a recognition of that and, of course, we now call that clickbait. In a way, entire sets of content companies grew up around the act of writing a clickbait-y headline. Right we now call them BuzzFeed headlines and Upworthy headlines for that very reason. Entire content companies grew up optimizing around getting people to click on things and then putting ads against that on their own landing pages.

So it became evident that we once again needed to improve the system that a click alone was not an indicator that this was interesting. We began to implement more sophisticated things like actually understanding if you clicked on something and then read it and came back to Facebook extremely quickly – what we know as a bounce in the digital business – was a sign that you had just been sucked into a clickbait, read about three sentences of it, and then regretted it and came back to Facebook. And so we collected that information and depressed the sharing of those clickbait-y sources.

But then again, the minute you wack-a-mole one of these problems, another problem shows up. And so it’s a constant evolution, an arms race, if you will, between the system that we have set up, the way that content companies use that system, and the way that we then change the system in turn to depress bad actors and elevate good actors.

originally posted on pbs.org