(Bloomberg / Getty Images)
Ten years ago this week, when Facebook was still a place where college students poked each other, the service’s users — all 10 million of them — logged on to their favorite website to discover the News Feed. It was the most significant invention in the history of the social web. It reimagined our relationship with personal information, changing every picture, poke, or status we posted into a stream of data for our friends — and then advertisers — to parse. At first, people hated it. But much as they complained, Facebook’s data logs suggested they came around quickly. The News Feed caused more people to spend more time on Facebook, a trend that picked up speed during the years that followed. Meanwhile, nearly every other social web service, from Twitter to Instagram to Pinterest, adopted the format.
From the start, the News Feed’s central feature was that it didn’t show everything: it curated the information your friends posted to surface what was most relevant to you. Ever since the feed’s launch, the company has been making adjustments to this process to get better at giving people the information they most want. At first, Facebook’s tools were primitive, but over the years the company developed and improved its algorithm—and then replaced it altogether with machine learning augmented by qualitative research. Now, when you go on Facebook, the News Feed considers thousands of factors to present a feed that is personalized to you. The more you use it, the more it learns to show you more of what you like.
Now that a seventh of the world’s population is on Facebook, the News Feed has has become a prominent source for news and information about our friends, sure—but also about the elections, an imminent tropical storm, a recent terrorist attack, and anything going on in the world more broadly. It has, as John Herrman recently wrote for The New York Times, “centralized online news consumption in an unprecedented way.” Every tweak to the technology that powers the News Feed has consequences for the people and businesses that attempt to harness it to win people’s attention. Along with this power comes a growing tension over how decisions get made about what information belongs in that feed. News organizations, many of which have seen their businesses decimated — or bolstered — as their readers turn to Facebook for headlines, express concern that Facebook acts as a giant managing editor for the web.
Facebook insists it’s a technology company, not an editorial operation. “When you think about a media company, you have people who are producing content, who are editing content, that’s not us,” founder Mark Zuckerberg told a Rome audience in August. “We exist to give you the tools to curate and have the experience to connect with the people, businesses, institutions in the world that you want. So every person gets to program their own Facebook experience.”
Two of the three engineers who helped launch News Feed are still working at Facebook. Chris Cox is now the chief product officer, where he is in charge of guiding changes to the big blue app. Andrew Bosworth, or “Boz” to his friends and colleagues, is vice president of ads and business platform, where he oversees how ads are made. (The third of the original trio, Ruchi Sanghvi, left Facebook in 2010.) I spoke with Cox and Boz to see what they remembered from the early days — and how they think about the News Feed today.
Chris CoxChief Product Officer
Credit: Facebook__Tell me how News Feed came to be.__
I started at Facebook in November 2005. There were less than 50 people. When I interviewed, the people that interviewed me explained that this is what they wanted to do.
If you remember using Facebook before News Feed, [the site] would say “welcome back” and that was it. You would click around people’s profiles, and the profile was a terrible way of keeping up to date. Facebook wanted a homepage to solve that problem.
Did you think News Feed would be a game-changer for Facebook?
We’d already been using it for six months before it launched externally. When you’re building something, a lot of what you’re relying on is your own love of something. We loved it. We definitely knew it would provoke a reaction and we thought everybody was really going to like it. And we were wrong.
Yeah, I remember.
We started rolling it out, and I think that Facebook had around 10 million people on it. We did it late at night because there was less load on the server. If you wanted to make changes, you would do it at midnight or 1 a.m. We flipped the switch and we were like, “Alright, let’s watch the feedback. Everyone is going to love it!” And we just watched the most vitriolic feedback: “Turn it off! This sucks!”
Did you freak out?
It was really, really disappointing. Our jobs were so all-consuming. For example, we didn’t have an infrastructure team. I literally used to sleep with my computer next to me with the error logs up, so that I could wake up every few hours and see that the error logs were being checked.
But, we had this conviction that if people used it, they would learn to love it. Mark was also very heavily involved, both in the development of News Feed and the decision that we had to give this a shot. After users freaked out is the first time I can remember that Mark addressed users and apologized.
We had not done much education [before we launched the product]. That was a lesson for us.
Also, that is the first time that I can remember a public conversation about privacy on Facebook.
Immediately afterwards, we released a whole bunch of privacy controls specifically for controlling how stuff works. We didn’t anticipate all of the types of controls that would be asked for, but we immediately went and built them.
Did you guys understand that what you were building was a new method for distributing social information, and others might subsequently adopt it?
It was interesting to watch the early stuff start to grow. So there were two groups. Each grew over a million people, I think. The first was called “turn off News Feed” or something like that. And the second one was to raise awareness for violence in Darfur. If you remember, in 2006 was one of the really awful humanitarian crises. Each group had over a million people and Facebook only had 10 million people on it. It was a really cool, early indicator of the kinds of things that may be possible one day.
As soon as we translated Facebook into Spanish it was pretty quickly [after] that in Colombia a guy named Oscar Morales organized a group called “No Mas FARC.” This was in 2007.
He said, “You guys have no idea. You have absolutely no idea how important this will be.” He was from a tiny fishing village in Colombia. He had these aerial photographs from Tokyo, New York, Oslo, Sydney, and Bogota. Each had huge amounts of people marching on a single day, carrying a sign that he had designed. He said, “I didn’t spend any money on this campaign. I just used free tools on the internet. I started a group, and 12 million people marched on a single day.”
If each feature that you add has such a profound impact on the things that people can do, how do you go about making adjustments?
We watch what people are trying to do. The News Feed itself was born out of watching how people were trying to use Facebook. And even reactions, which was one of the more recent changes, was just born out of watching what people were trying to do and listening to them.
How has the process for making adjustments changed?
Well, I think in the early days, It was pretty anecdotal. The woman I sat next to processed all of the tickets that came in overnight. She would relate to me what feedback we were getting. It was kind of like coming to work every morning and reading the mail.
Now, obviously, it’s a lot more dialed into a system. I mean, we do a ton of international research. We have people all over the globe. We are developing for thousands of different combinations of devices. And we’ve gotten a lot more sophisticated in developing tools to measure what’s going on, and good at doing research and understanding it.
The most important data that we get is feedback at scale of what people find meaningful. And we have, over the years, developed ways of understanding thousands of people every single day. We can find patterns of where we’re getting something wrong.
There’s what people want in their feed, and then there is what people say they want in their feed, and then there’s what people should see in their feed. What’s Facebook’s role in making those decisions?
We want people to have an experience that they consider to be meaningful to them. Think of it as if you were to go home and have dinner with your family or drinks with your friends, and just observe what they’re talking about.
How sophisticated are you guys at showing me different information than you might show someone else, and how is that evolving?
Okay, we’ve come a really, really, really long way. The biggest difference is that if you look at Facebook 10 years ago, you weren’t connected to that many people and there wasn’t that much going on. If there’s one thing that characterizes what’s changed, it’s just that there’s so much more information.
Also, in 2006, it was a wall of text in a browser on a webpage. Today, you can move your phone around and see a 360-degree photograph of the Dolomites that I just took. It’s something that is moving and something that has sound. We have always wanted to get to this world where, as much as possible, I was sending my experience to you. Over time, it’s getting closer to that.
Let’s talk about the News Feed in the context of the political debates right now. If you’re giving me more of what I want, the reflection of that, for me at least, is that I end up seeing a heck of a lot of stuff that reinforces my political stance. Is that what Facebook should be doing, or what you intend to do?
A lot of what you’re seeing on Facebook is coming from weak ties, which are not your close friends. If you look at the world without Facebook or before Facebook, a lot of what you were talking about was with the people you were around day-to-day — the people you were sitting with. Your best friends. What Facebook surfaces, that wasn’t surfaced before, is all the information from all these people you don’t normally hear from, and at least for me, that surfaces a lot more opinions from sources I wouldn’t read.
I think sometimes it gets a little bit lost that a lot of what Facebook surfaces is not the people who are super close to you.
So what you’re saying is that people are seeing more perspectives on Facebook?
I’m just saying that if you just look at the math, most of what you are seeing from Facebook is coming from weak ties.
Is the News Feed intended to surface all the news and information somebody wants?
The way we think about it is the stuff that matters to you most. Most of those things for me are going to be about friends. Some will be about what’s going on in the world, in my city, in my company. Some are going to be about my friends’ children. Some are going to be tragic. Some are going to be amazing.
Andrew “Boz” BosworthVice President of the Ads and Business Platform
Credit: Facebook__Tell me about the early days of News Feed?__
In a lot of ways News Feed brought me to Facebook. It wasn’t a well-developed idea yet, but it was a problem that needed to be solved. It was Chris Cox, myself, and Ruchi Sanghvi, and we sat in a little pod by ourselves. Ruchi was the product manager. Chris built the front end. And I built the infrastructure and rankings.
Despite how incredibly hard it was to find the things that you wanted, people were coming back [to Facebook] over and over again. So they have this valuable stuff and people wanted to get to it. We just needed to figure out how to make it easier to find out things.
When this was happening, Myspace was the category leader. Myspace didn’t have anything like that. Did you draw from any outside inspirations?
We didn’t. I think if we had been more connected in the valley or more experienced, we might have. I wasn’t even on Myspace. I wasn’t cool enough.
Then News Feed launched, and it wasn’t received the way that you expected.
That was hard. We really believed that we were solving such a serious problem that people would just love it. It was a real surprise and really eye-opening for us.
That was a really important lesson for us in terms of how to approach change and how you talk about the value of software and make sure that people understand what you are trying to solve. It was very humbling. It was so far from what we anticipated that we really had to think, “Okay, well what is this change and do we believe in it? Why is it good? And how can we make sure that we addressed the concerns that people have?”
Did you have an impulse to change it immediately?
We had one affirming piece of evidence. Almost overnight, people’s engagement and activity with the platform shot up. It never came back down. And in a cruel twist of irony, even the success of the protest groups on Facebook was the real testament to the product that we had built. There were groups that had a million people in them when the site only had 10 million people total. That simply wasn’t possible before News Feed, and it was exciting to see, in the very fabric of this protest, the evidence that this technology was working.
Were you doing any advertising at all at that point?
Yes, we had flyers. And we actually did have ads in News Feed in the first couple of years. Back then, they were actually sponsored pages and sponsored groups, so it’s not like today. You would actually sponsor the Page [and then it would appear in the News Feed]. Our first big deal was with Victoria’s Secret. Apple was another big deal.
And then flyers lived in the right hand column. We ended up focusing exclusively on right hand column ads for five years before ultimately reintroducing ads in the News Feed.
What were the early development turning points for News Feed?
There are a few. The initial launch set the benchmark for how important ranking and relevance was. We actually played with the idea of moving away from ranking and relevance to sharing all the content. We tested it. It turned out that ranking and relevance were really the right thing.
Another big shift was when we went to real time. You used to have to wait 30 minutes before a post would show up. The company had matured a lot. We had more funds available to us, and of course RAM had gotten cheaper.
And I think that the third major point in the evolution was the reintroduction of ads into the News Feed in 2012. It was something that we approached cautiously, but we are on a path to build really great experiences for people who advertise. If you look at the last couple of years and how successful that is from a business perspective, which frankly is secondary for us, we have been able to maintain the quality of the feed experience.
How has that challenge changed over the last 10 years?
The scale of Facebook brings both challenges and opportunities. If you think back to 2006, we sent a survey to the whole company. That was maybe 40 something people. We asked, “On a scale of 1 to 10, how much do you like seeing photos? How much do you like seeing text updates?” The original ranking of the Facebook algorithm was based on a survey of 40 mostly college grads who worked at Facebook in 2006. Of course, within a couple of weeks after launch, they were completely out of the algorithm as we got new information.
Throughout the last decade, there have always been developers and companies that have harnessed the News Feed’s algorithms to their own end. How do you combat that?
One thing that we have done and started to talk about in the last couple of years is engaging qualitative measures. It’s fundamentally a check-and-balance against any kind of machine learning run amok that gets us into these situations where someone is content farming or exploiting a loophole in human behavior, in the case of clickbait. And I think that qualitative process makes it much harder for any long-term gamification to be profitable.
Was there any larger understanding that what you would create was not just a feature for Facebook but a new design element for the way that social information was distributed on the web?
No. In the aftermath of the launch of News Feed, one of the narratives that was in the media was how bold of us it was to move beyond page views. That was the de facto currency for all websites at the time. I don’t think we’d had a single conversation about that. We were really just trying to solve a problem that people had.
We believed in relevance and putting people first. We always believed, as a mission-driven company, that that would change all software and ultimately the world and the way the world connects and interacts.
[Revealed: Facebook's Project to Find Out What People Really Want in Their News Feed - Backchannel
*Every weekday beginning at around nine, 30 contracted white-collar laborers arrive at a generic office park in the…*backchannel.com](https://backchannel.com/revealed-facebooks-project-to-find-out-what-people-really-want-in-their-news-feed-799dbfb2e8b1#.19dhguozj "https://backchannel.com/revealed-facebooks-project-to-find-out-what-people-really-want-in-their-news-feed-799dbfb2e8b1#.19dhguozj")
