BuzzFeed CEO Jonah Peretti: 'It's not just a site, it's a whole process'


BuzzFeed announced today that it’s acquiring the talent team behind mobile app developer Hyper IQ, a significant investment in building native apps for a publisher that’s traditionally been web- and social-first.

That’s a big tech story, and I spent an hour talking to

BuzzFeed CEO Jonah Peretti about the media business, what technology actually helps BuzzFeed do, and what’s next for a company he says was a “cat site” three years ago.

Nilay Patel: It sounds like you have some pretty exciting tech news going on.

Jonah Peretti: I think that one of the interesting things about all the stuff we do in tech is that in general technology is harder for people to understand. Some of the stuff that I'm most excited about I don't get a chance to talk about as much. It seems exciting to me and is probably interesting to readers of The Verge, but there are a lot of people who love BuzzFeed for entertainment and news and how it gets to them isn't really as exciting as what is getting to them.

Obviously we have a very big tech audience at The Verge, but I spent six months earlier this year working on the team, and it was actually really interesting for me to step back and watch how our publishing platform helped that team scale so fast and hit a different big audience. We think about our platform so much that it almost circles back around to taking it for granted, and I suspect that that is true at BuzzFeed as well.

Well, there's certain technology that you want to have disappear so it makes things faster and easier. Ev Williams likes to talk about technology that lets you skip a step — something took five steps now takes four steps. The benefit is something that isn't there. Then there also are other things where technology doesn't recede to the background, like a lot of the stuff we do with new post formats or slide-y things or quizzes or rubable GIFs. That’s all more in the category of fun technology that you get to play with.

And now you’re buying a company to help you build native apps?

Yes. Hyper IQ is based in Minneapolis, and it's a pretty exciting opportunity for us to continue to expand our tech team. As we get bigger, we've expanded the company to a lot of different cities around the world now. There was a period where that was frustrating, because we had everyone all in one office, and then we didn't, and now it's moved to another mode. We have a network forming, we have offices with critical mass that can build parts of what we're doing. We're doing so much more than we've ever done before, and there's nice big meaty things for people to work on. We'll grow to 30 people in Minneapolis, and this acquisition is the first step.

How many other cities are you in?

We're in, let's see… we're in London and Berlin, and Sydney, and DC. We're opening up in San Francisco right now. Sao Paulo, in Brazil. It's getting up there.

Hyper IQ has been building your app thus far?

We have a larger team in New York that's working on the apps, but [Hyper IQ] has been working with us. Part of the reason we wanted to acquire the company is that they've done such great work.

Maybe this is a question that only a media nerd asks a media nerd, but it seems like a great deal of your investment in technology so far has been at the platform level and for content creation, while distribution has all been about making things really sharable and really great so people will share them. But buying a company to make a native app is a distribution side move. A native app puts your icon next to the Facebook icon on my phone. How do you think about that relationship?

I think that things have really evolved fairly quickly. I think the classic model of a media company is that you have content and you have distribution, and usually they're pretty separate. One company makes content, another company does distribution, and they often have a deal, a partnership, to distribute the content.

BuzzFeed has invested in a model that's more vertically integrated. We're building the technology and the site and the CMS, and also the brand and the content, and building this full-stack startup. What we're seeing now is that's become more complicated. You're starting to see more network-integrated companies — we have teams that are making great news, and entertainment, and lifestyle content. We break that up into News, Buzz, Life, and then video. That content is part of a vertically integrated site that lives on the web. And that same content is going to an app. And to native content formats on Facebook, on YouTube, on other platforms.

And technology — I'm sort of describing a diagram in my head — technology is important at every single level. Technology is important for how you create the content: do you have tools to build games or quizzes or beautifully laid out longform articles or breaking news? The way the CMS works and the way that the format works is driven by technology, so having great technology for making content is important.

Then having great technology for distributing content on your own site matters. How do you recommend articles? What you show in the thumbnails, what kind of recommendation system do you have? Then technology really adds to distribution on the other side — do you have a good way of tracking which of your YouTube videos are getting lots of views, and using that data to inspire teams to make things that will connect with the audience?

So I think that when you look at the role of technology it's everywhere in this network-integrated model. You have content creation requiring or benefiting from technology, you have content on your own site and owned and operated properties, apps, and websites that benefit from technology, and then you have technology that is even really important when you're publishing to other platforms because there's APIs and ways of tracking what's happening to other platforms and understanding what that means on other platforms.

That makes sense, but let me pull this out. This sounds like ‘BuzzFeed Everywhere.' You're pushing BuzzFeed content into every platform and you're making kind of every bet. You're making a web bet, you're making a ‘social-content-driving-back-to-your-website' bet, you're making a ‘we're putting an app on your homescreen' bet. Where does BuzzFeed live? Where do you think of as BuzzFeed's canonical home?

I think it's becoming much more distributed. And people are still figuring out what that means.

I think mobile apps are increasingly important to us. We care most about what’s best for consumers, so it's not like we have an ideological or religious view on where content should live. Should it be a native app, or on the web, or mobile web? It really should be dictated by what's best for the consumer and what's best long-term for building a media company that can thrive with how people are consuming media today instead of how they consumed media 10 years ago.

When you think through that lens, things are becoming much more distributed. Our video business was over 800 million video views last month, we're on pace to do even more this month, and it's very distributed across other platforms. Our videos are on Facebook, on YouTube, on AOL, and on Yahoo. But we're able to use one dashboard to track how those videos are doing on all different platforms, and Ze Frank's video team in LA is able to learn from all of that data.

"It doesn't matter where the content lives, but it does matter that you're able to get some knowledge and learning from it."

So the key for us is publishing to wherever consumers really want to engage with our content — but can we get data and knowledge back so we can learn and get better? Also we should get money back, which allows us to fund doing more of it. If you have that, it doesn't matter where the content lives, but it does matter that you're able to get some knowledge and learning from it. That's the thing that traditional media sort of lacks. It’s a huge advantage for a tech-oriented media company that’s on a digital platform.

If you're a traditional production company and you make a television show, you sell it to a network, and you get Nielsen ratings back. You don't really know much about which parts of your show people liked or didn't like, or what people thought about it. It's really hard to link producing content to a real relationship with your audience. I think with the technology base underneath everything we do, we're able to get knowledge back and have a closer relationship with our audience and serve the consumer better. That matters to us more than asking if the content is on mobile web or in a native app, or if it lives on our own properties or on some other property.

So the news here is Hyper IQ. When you're thinking about BuzzFeed as distributed on every platform, what led you to actually buy the shop that's helping you build the native app?

Native apps are extremely fast, and they are a great way to get to news and entertainment and video. Native apps can be a really great experience. The other reason we're really excited about native apps is that you can get data back in a really rich way about how people are engaging with your content. So that helps us make better stuff and be engaged with our audience. So that's a big reason.

[Hyper IQ] has really talented engineers who are doing great work, located in a city where there’s a lot of great engineering talent. We think that we can provide really great opportunities for them to do really interesting work that they might not be able to do or find elsewhere. I think Minneapolis is an underrated tech town. There's a lot of really great engineers there.

The other thing we see, and lots of publishers see, is that the users of our mobile apps share at a higher rate. They visit more frequently and read more. It's also a place where it makes lots of sense for us to invest.

The Verge used to have an app, and we actually just shut it down in favor of responsive web. One of the reasons was that we weren't able to innovate as fast as we were on the web. How are you going to manage that as you're more native?

So this is a real challenge.

If you come up with web development as your core, you get used to the ability to deploy multiple times a day, to test features with 5 percent of your audience, and only rolling out to your audience once you see that it works well. A lot of agile development and flexible testing is much harder to do in an app environment. It's harder to do for technical reasons, it's harder to do because of app store approval, and pushing your app is really a big deal. Pushing a deploy to the website is a trivial thing that really any engineer is able to do because we've built all kinds of tools to make deployment easier on web. That's great, because you can do more faster and have shorter loops and shorter cycles of development.

But mobile apps are somewhere between the web and old-school desktop software development, where there’s the big release once a year and the patch to fix bugs. So it is a challenge, and that's primarily why we wanted to hire 30 people in Minneapolis and many more than that in New York and LA and San Francisco and other places. The tech team is over 110 people now, and we want to go to 250 people by the end of next year.

It's a big investment to develop for all the different platforms that are out there. We feel that it's worth it. We won't be able to develop as quickly as we do on the web and have as short a dev cycle, but you can start to figure out ways. It's a kind of different mindset of how you build things — if you have a big enough team you can start to really push features that aren't even possible to the web, because they're native.

"Pushing a deploy to the website is a trivial thing that really any engineer is able to do."

When you look at our app now, we spent really a year working on underlying architecture and making sure that it loaded quickly, it didn't crash, and it was designed in such a way that we could build some cool features. It would seem crazy to spend that amount of time on the web. Now we're ready to start building some of the cool features that will be more noticeable. We went from having an app that crashed a lot at the beginning of the year to having an app that was in the Best Apps of the Year on the Android and iOS App Stores.

So it is a challenging thing, and my natural impatience which stems from being used to the web is definitely [tried] sometimes. Sometime I want to be able to think of an idea, test it a week later with a small percentage of the users, see if it works, and then roll it out. And [with an app] you have to think much more strategically and build things on longer cycles that you know or hope will have a bigger impact. So you're doing things, you're taking longer strides, which can be frustrating or scary.

Let me get back to the idea of putting an icon on the home screen. It's something I think about a lot. You’re putting an icon next to Facebook. Once you start delivering better and better experiences in your native app, especially for something like video where the best experience is native, how do you think about competing with Facebook which drives all of the traffic to your other media types?

Well, it's really similar to BuzzFeed's homepage. It's not that it's competing with Facebook, it's that we have a dedicated audience that comes to BuzzFeed. Whether they come to our app or our homepage, they're the ones who initially share things to Facebook. Our content wouldn't even be on Facebook if it wasn't for people who thought it was worth sharing with their friends. The app really will be a smaller number of visitors, but they'll be the ones with the biggest influence on what BuzzFeed stories initially are getting shared.

I see it as part of a larger ecosystem of content discovery, where apps created by media companies are actually pretty important because there's a place people can discover content they think is worth spreading across a social network. So in a way our app is a service to allow people to find stuff that they want to share into other services.

I spend a lot of time thinking about homepages, because we have a big one. The Verge is older in that way. How big is the BuzzFeed homepage? Do you guys spend a lot of time thinking about it? How do you think about programming the first screen of the app?

So we're actually kind of deep in this right now. Part of adding additional tech resources is really building collaboration between engineers and editorial people. That practice is core to what we do on the web — there's folks on the engineering team working closely with editorial people, and they go back and forth, and they develop ideas. It's kind of a rare thing to have a tech team at our scale that also gets to work with the people making the content. We can point out a new format that inspires an editor to make something, or an editor has an idea that's inspired by new technology, or there are new ways to display content. That's something that we want to also do with apps. And we haven't done it on the same level with apps. Sometimes the app is getting programmed by what's happening on the web, and that doesn't really make sense. People visit the BuzzFeed app many times a day more than the average person is looking at the homepage of the site. So it’s something that we're pretty excited about. We've got some star editors and star developers working together to define how to program the front page of an app, and how much personalization to have, and how to say this is important news. All that has to be a real partnership between great editors and great developers.

So there’s personalization that happens in your app?

There is some light personalization that we played with in an earlier version of the app, where you would choose what you liked and didn't like. Now we're working on some subtler, more automatic personalization. Apps definitely lend themselves to personalization in a way that is easier than the web — even if you don't have a logged-in user, if they're using the app, you have a sense of the person. But I don't think we want necessarily want [the app] personalized. You probably do if you’re Facebook or Twitter, but if you're a news organization, you also want to sometimes take an editorial stance where you tell people something because it's important.

Once you go down the road of having a native app, having the content right there, having some offline capability, and then doing some algorithmic personalization based on interests, you’re basically creating the next news feed, right? I hate to keep hammering in the idea that you're competing with Facebook, but it's an easy thing to draw the broad strokes of.

We don't really have any plans to do personalization on that level. We will always have a weird, eclectic mix of things that we're creating. We don't have a universe of a kajillion stories to personalize for each individual person. We have the stuff we're making, and even though we have a large editorial team and reporters around the world, and all kinds of great content, it's still content being made by one organization, and not the entire world.

The primary way that people get a personalized content experience with BuzzFeed content is the Facebook newsfeed or their Twitter feed or Pinterest or YouTube. But when you come to BuzzFeed, directly to our homepage or our app, the kind of personalization we would want is much lighter than that. It’s like, say you hate cute kittens because you're a monster with no empathy. So you don't see any cute animal content. Or you don't like sports, or you love sports, you have a certain category. Some people really care about what's up with Taylor Swift and the latest in celebrity stuff. Other people find celebrity news enraging. We realize that there are some people who may want a hard news experience and some people who may want mostly video and love all the stuff we're doing with video, and there's some people who will like the mix of stuff but they don't want X, Y, or Z, or they want a little more of A, B, or C. That's the kind of personalization that makes sense for us.

The heavyweight personalization that social networks do either through something like the follow graph on Twitter or through more algorithmic approach on Facebook is designed to give people the stuff that they want. What we're focused on is giving people stuff that they think is worth sharing with other people. We want the stories we're doing to have the biggest possible impact. So if we do personalization, it would be more of a personalization about what you’re most likely to share or discuss with your friends.

You can do video much better in the native app. Is that going to be an immediate focus, or is that going to be something that you build towards?

Well, our video business is pretty distributed. We’re getting the native video viewing experience by distributing our video into other native players. So we could have said we're really investing heavily in video, let's put this all in our own player, and let’s make people use that player. The problem is that even if we had an iOS native player and experience in our app, our video wouldn’t be be native in Facebook's app. Right?

We want our video to be native when people open in Facebook's app, because it will have the silent autoplay and all those things that are native to that environment. So, it's not this crucial, pressing thing where people will not be able to watch BuzzFeed video or that people won't have a good experience watching BuzzFeed video because it's mobile web, because most of the views now aren't happening in mobile web. They're happening in native players of other sites where we take our video, or other services where we take our video.

Sure, but I mean, that raises a pretty obvious question of how well can you monetize your video when you're distributing views across a million different players?

I mean, our video business is in the hyper-growth stage. There's about 150 people in a four-acre studio lot in Los Angeles that Ze Frank runs. We've built it essentially as a break-even business while in hyper-growth mode. So, we probably should be investing more in it, but we make branded video. Things like Dear Kitten, which we made for Purina, is a great piece of content that’s also an advertisement for cat food. It works as content and also has amazing effects on purchase intent. And we can take what we know about making great video and make a branded version of that video, and use that as a way of monetizing. And then we also have all of the monetization that is part of YouTube, which is kind of light monetization but with the amount of views we have, it can actually add up to something pretty significant.

"Our video business is in the hyper-growth stage."

Video is really interesting to me. One thing I think about the multiplier effect of technology. The way you scale and the way we scale is that our platforms allow us to scale for cheap or free. You can add one more reader for free. And if you're really good, the platforms enable you to create content at a lower cost as well, because it's easier to make one more quiz or the format lends itself to something more popular. You can get more efficiency out of one more employee. That is not true right now with our platform, I’ll be honest, or I'm guessing your platform when it comes to video creation. Because you make a CMS that enables you to write or publish a quiz really fast, but you don't make a video editing software that allows you to make animated explainers for YouTube the way that After Effects lets you do it. So you are constrained by the old way, and technology doesn't provide nearly as much of a multiple in video.

How are you attacking that problem? I see that as a big problem for every technology company that wants to go into video, and I say that as a man who works at a technology company that invests in video. I'm very curious about this answer.

Well, some of it is the reason we have that LA campus is that it allows us to build a set, and we have green screens and facilities. All of that is an upfront cost, and then the variable cost of making a video is actually very low because we have all of that in place. So it's not really of what you would think of as technology, but it's partly why the old studio lot model makes sense — you end up having the ability to make video at a lower cost because you have resources and facilities and talent all together in one place.

Then all the data that you get back helps you learn which formats work, which actors work, which topics are people interested in, and that furthers creative development. You're not just outsourcing it to a production company — you actually have the people making the videos learning based on how the videos are being distributed. Then I think there are some technology factors where you build technology to help you syndicate video, to help you track it, help you understand potential technology benefits to make workflow faster, better, make it easier to translate videos.

I think that usually there's ways to generate some advantages from technology or from organizational or operational initiatives, but it's more expensive to make video, and the cost of failure is higher, and the way video is shared is different, and it's all these things that make it challenging. It isn't it just any improvement based on technology and operations.

I read a couple days ago Ben Smith saying that in three years he doesn't think BuzzFeed will exist in its current form. Can you tell me what Ben was talking about and what you think that means?

He was talking about all the stuff we've been talking about. It's hard to predict three years out, so part of it was saying, "Who knows what'll happen in three years, what the web will be like in three years?" We've been based on a model of continual change. Three years ago, BuzzFeed had no reporters. Two years ago we had no video. One year ago we didn't have foreign correspondents around the world or an investigative team. Three years ago we were a cat site, an internet meme site. So a lot has changed in three years.

It's an out-of-context quote — Ben was talking about the changes that have happened in three years. We went from the traditional media model of content and distribution to the vertically-integrated model of content distribution technology to the network-integrated model of technology helping at every level. Technology helping with content creation and then that content going on our platforms, distributed across the web, potentially going to traditional platforms like television or print. We don't really have plans to do any print.

But there's a possibility of having something that you look at and think that this isn't a site, this is a global media company. It's not just a site, it's a whole process for distributing news, buzz, life, on the web, mobile, native apps, and it looks very different than it looks today.

When you think about that process of going from one stage to another to landing at the network-distributed model, in terms of consumer experience, where does the technology have the most impact? Where's the place that you're delivering to them the best experience?

So the media bundle, if it’s the newspaper, it lands on your doorstep, you go through it, and see all the different sections, and it’s the complete, canonical version.

And I think what you see now is technology and engineering is having a bigger impact on what the bundle is. Instead of one bundle, we ended up with lots of bundles. We're some percentage of a Facebook newsfeed bundle, we're some percentage of Twitter's bundle, we're some percentage of Pinterest's bundle, and so there's all these personalized bundles that are being delivered by technology, and we're a part of it.

And then when it comes to our app, and to our homepage and to our properties that we're developing, most of that is our content, so then the app becomes a bundle that is 80 percent BuzzFeed content. Maybe there's some other links or other content, but it's mostly our content. And so you end up with multiple bundles. Depending on who you ask you'll get a very different view of what BuzzFeed is.

We’re moving a world where there will be multiple bundles, and that's why technology is so important, because engineers are able to understand this dynamic bundling, which is a different way of thinking about the media industry. They’re able to build better bundles. And brands will matter across this distributed landscape. BuzzFeed needs to stand for a certain kind of quality and sensibility — we need to make things that deserve to be in other people's bundles, and we'll have our own bundles that we create. Which is part of what the team in Minneapolis will build with the app they'll work on.

Most of the media industry regards BuzzFeed's technology advantage as some sort of dark magic — a mixture of platform and social prowess that's terrifying for an older generation of media companies. And for good reason: they likely can't change their culture or develop their platforms fast enough to compete. But what scares BuzzFeed in the same way? What don't you guys understand?

I don't understand the traditional media business, so the feeling is mutual! And it is still huge and powerful, in scale and especially revenue. We are very small compared to Disney, Viacom, Time Warner, and the other big guys.

"It's not just a site, it's a whole process for distributing news."

BuzzFeed excels when we can combine art and science. This is why we focus on platforms where creative people can try ideas, get feedback from the world, and use it to learn and get better everyday. This is why developing the technology for our own site and apps is so important to us and why we invest so much in data science. And it is also why we invest in platforms like Facebook and YouTube that give us data back to incorporate into our creative process.

This isn't possible in print, broadcast, or traditional films which may be why the media industry is such a dysfunctional place. Executives make huge bets based on gut, it’s hugely expensive to take risks, and most projects fail. It’s enough to drive anyone a little crazy. I don't understand that world, and it scares me. And I also get scared whenever a tech platform wants to evolve to be more like cable TV or more like the media industry, by locking up the data and tightly controlling media consumption. In the long run, I think the platforms that share data with creators will win and so will the media companies that hire lots of great engineers to help make sense of that data and put it into action.