Pod Insights a16z Podcast a16z Podcast | For Buzzfeed Sharing is the Metric that Matters About this episode Jonah Peretti is building BuzzFeed to “inform, inspire and entertain” in a world where news and entertainment is increasingly passed around on social networks and consumed on smartphones. Chris Dixon, who is on the board of BuzzFeed as part of Andreessen Horowitz’s recent investment in the startup, sat down with Peretti to talk about building a media company from scratch. If editorial success is driven by digital word-of-mouth, and mountains of data about what people like to read and watch, what does Peretti do differently? Why do BuzzFeed’s lists work so well? Does video finally make financial sense? And why, even if it’s a traffic monster now, is Peretti not religious about any particular format -- including lists. (For more on Dixon’s take on the media business, and BuzzFeed’s “full stack” approach check out this recent post: cdixon.org/2014/08/10/buzzfeed/)
Transcript 0:00 Speaker 1 Hi, this is Chris Dixon. This is the A16Z podcast. Today I'm here with Jonah Pretty, the uh founder and CEO of Buzzfeed. We announced I guess a few weeks ago now that we invested in BuzzFeed. So we're very excited about that. And thanks for being here.
0:13 Speaker 0 Thanks for having me.
0:14 Speaker 1 I guess we'll maybe you could start by just sort of for people that don't know explaining the premise of Buzzfeed. Why you started it and
0:22 Speaker 0 Yeah. So the idea the idea behind BuzzFeed is that you know people still want to be informed and inspired and entertained, and the media industry has always had that role. But the way people consume media has really dramatically shifted in the last 10 years with social and mobile and digital video being becoming the dominant ways people consume media. And so BuzzFeed really is a media company for the social mobile world.
0:46 Speaker 1 And so and so you know, someone might say, okay, great, people read stuff on their mobile phone now, they click on stuff on Facebook, but how does that really change what a media isn't a media company? I mean, a skeptic might say, you know, don't you still want to write the same kind of stuff you were writing before? Like how does that change what you do as a media company, the fact that people consume it differently? Aaron Powell
1:09 Speaker 0 Yeah, so there's a lot of differences. I mean, from a business perspective, there's differences in cost structure. So you know, printing and broadcast. You
1:16 Speaker 1 don't have a trucks and
1:18 Speaker 0 we don't have timber interests and pulping mills, and we don't we don't uh have to you know have satellites or you know, cable networks or broadcast networks in the in the same way. Um so so distribution is different. But the the biggest difference is that you get a tremendous amount of data back about what consumers are reading and what they're sharing. And for us, sharing has always been the the biggest metric because it shows someone thinks that a piece of content is worth passing on to a friend. And so you've always had this word of mouth. People watch a television show, they really liked it, they'd go tell their friend around the water cooler, I really like this show, but basically the distribution was was tied to these massive industrial era uh models, and then you have a little word of mouth as extra, and now it's been kind of inverted where the word of mouth is the distribution.
2:05 Speaker 0 Uh and and so you're getting so much data back about what people like and what people share, and that can immediately inform the media you're creating. So instead of just guessing and making something and hoping it's a hit, you're you're able to be in a more direct relationship with with readers and what
2:21 Speaker 1 does that mean for how you work? Like, does that mean that you see that people like some topic and then you write more about the topic? Does it mean you change the formatting, the headlines? Like what is it what what is it what is it in terms of how does the data translate into actions?
2:33 Speaker 0 Well, it's it's a the way to translate data into action is a huge part of what we do and what we think a lot about, and it's it's not totally straightforward. Like you can you can use data to get people to click a headline, but then not deliver on the story, and that actually has a long-term negative effect. And so and so some people use the data to optimize for the very short short term. Our goal is to use the data to understand on a deep level what kind of media people want to share and consume and and to to have a bigger impact on people's lives and a positive impact on people's lives, informing them, entertaining them, inspiring them. And so that means you have to have a more nuanced and skeptical view of what the data is telling you. Um just because someone is clicking at a higher rate doesn't mean you're doing a good job.
3:20 Speaker 0 Um just because someone's sharing doesn't always always mean you're doing a a good job. And so we try we try to um avoid local maximums and and and focus on on things that seem like more meaningful long-term signals.
3:33 Speaker 1 So one um line of argument today is that media companies, like print uh news news organizations in particular had kind of or or printed news had local monopolies in the past, and that created very good businesses for them. Um the New York Times or whatever, the local New Jersey newspaper or something. Um that now with the internet, that's destroyed these geographic boundaries and some people think has destroyed the economics of that of the news business. Um obviously you don't think that. Like how how how do you know how do you create a dominant news company in the modern era? And what are the barriers to entry?
4:14 Speaker 0 Yeah, so there used to be local, there used to be loc geographic-based monopolies and spectrum-based monopolies. So if you were one of the three television networks, you know, you had a huge advantage in that there was only two competitors and you had a slot on the dial or with radio, uh, you know, they both had spectrum monopolies and local monopolies. And so what happened is if you got an early start in the beginning of radio or television or newspapers, and you you amassed a big enough audience in a local market, it became very hard for anyone to compete with you because you already had the the big printing press and the trucks, and then you had a lot of good editors and reporters, and so then you'd attract more editors and reporters, and it would became very difficult for someone else to like create another newspaper.
5:01 Speaker 0 Um the argument is on the internet that's not true, and there's a million blogs and everyone's on equal footing, and and so you can never build a a defensible business and you can never generate great profits, and that this is a terrible thing for the news industry. And I mean it it's funny actually when you hear people say it's a terrible thing for the news industry because they're saying we need these monopolies that are impossible to compete with, and that's what's good for society. Um it actually um is more much more complicated than that. And in fact, um I think when you look at the at the way that technology can provide competitive advantage for companies today, and it's it's actually um makes it in some ways different, but bigger barriers to entry. Having a site that runs quickly, having an advertising model that works on mobile, um, having um a good data science team that can give you better insights in terms of what people are sharing and what people are engaging with.
5:56 Speaker 0 Um certain network properties, once you have um a certain level of scale, there's there's um things like preferential attachment and network theory that that suggests that it becomes easier to to uh to to reach an audience. And so all those things together, people I don't think understand very well, but it does make it possible to build a defensible business, and that's why we've been able to grow so quickly generate profits and in while simultaneously investing in foreign correspondence and investigative news team and other things that people thought weren't possible for internet companies to invest in.
6:31 Speaker 1 I mean, the other uh obviously the big change is that the addressable market is now there are I think the last numbers I saw two and a half billion people with internet connected smartphones. That number a lot of people think are is going to five billion very soon, um, which is which is you know two uh hundred X what uh the scale that a media company in the past could have reached, right? I mean so even if you assume the margins, let's say drop you know by let's say 3x or 5x if the scale of the opportunity goes up a hundred X, it's more than compensated for.
7:04 Speaker 0 Yeah, and people misunderstand that piece a lot too, where they they look at particular companies that were have print publications and then they go online and they're like they had two million subscribers in print and they generated that this much revenue, and now they have 20 million or 30 million, and they're generating the same revenue from that that they did from the two million in print. Therefore, print monetizes well and the internet doesn't. And what they're missing is that it's it's actually possible to reach a much larger audience. And and if the
7:30 Speaker 1 two million were sitting down and reading it for an hour in the morning, the 20 million, it's they're like unique visitors, which means somebody might have passed through and read a paragraph in the entire month. Right. They're not they're not commensurate measures. So
7:43 Speaker 0 yes, for sure. So they're not the same, and also it's possible to reach more people than ever before. And so even if they didn't, if even if even if things were monetized at a lower rate, the fact that you're able to reach 100x more people means that you can build a really big business. Aaron Powell
7:57 Speaker 1 So you see some really interesting stuff. You've always, you know, I've always gone to you to get little insights into social networks and things because you as a content creator um sit on the cutting edge of this and see you know the rise of Pinterest and WhatsApp and all sorts of other interesting things. Um can you talk a little bit about that? You even go so far as to organize BuzzFeed, right, according to which which social network the content is is created for.
8:24 Speaker 0 Yeah. So we see the the social networks almost like the new cable networks, and we're the new you know, MTV or CNN where we're making content for these these networks. And the you know, we this kind of gets back to what we're talking about earlier, but the way the the way you make content changes depending on the distribution. So when CNN started making making there wasn't 24-hour cable news, and that was a different kind of media that was only possible because of cable and doing something in the half-hour evening news is very different than 25 years. Yeah, and the point is
8:55 Speaker 1 that yeah, 24, you don't just do the half hour over and over. You do something very different. Trevor Burrus,
8:58 Speaker 0 Jr. Yes. And you can do things like go live and follow a story, and you don't have to break in. And so there's a whole bunch of different things you can do that that um gave them an advantage because they understood that cable was different than broadcast. With the internet, it's even with social and mobile and digital video, it's even bigger changes where understanding understanding how these different platforms work give you uh uh uh a huge advantage. Um we are continually trying to under learn how to make content that really is native for these platforms. People think BuzzFeed does you know native advertising for BuzzFeed. We actually do native advertising for other social platforms for Facebook or Pinterest or um you know the the mobile web, the social web. Um and so we're constantly learning trying to understand and learn what how how do people consume media in different contexts on different platforms.
9:50 Speaker 0 And so um we you know, I I've been shocked that that our video consumption. We did four 400 million video views last month and about half as mobile. Like I thought video wouldn't be as big on mobile just because There's sound and a smaller screen and really mobile has sucked up lots of video consumption. The same for long form. We we had a 6,000-word piece where the average time reading it was 22 minutes on mobile. And so people are sitting on their couch and the phone is good enough for them to read the long form article. And even though they probably could get up and get their laptop or uh an iPad or something, they're the effort of getting up isn't worth it when your phone is pretty.
10:29 Speaker 1 I've started reading books. I actually like I've the Steve Jobs book, for example, when it came out, whatever it was two years ago. I read the whole thing on my iPhone. Um you get used to it pretty quickly. And
10:39 Speaker 0 I didn't expect it. And
10:41 Speaker 1 just the convenience of being able to like you get in the car and you can still read or whatever, you know. Um makes up for the screen size. You get used to the screen size too.
10:50 Speaker 0 Yeah. Um the other thing that surprises a lot of people is that Pinterest sends double the traffic to BuzzFeed as Twitter. Yeah, a lot
10:56 Speaker 1 of people don't because you do it, you have a big DIY section, you'll have a recipe with a photo. Yeah,
11:01 Speaker 0 we had this clean living challenge, which was uh 30-day recipes, original recipes, breakfast, lunch, and dinner for 30 days. We've had over two and a half million people read that. And the vast majority of that is coming from Pinterest. And you do
11:16 Speaker 1 things like the lists, uh, tell me if I'm wrong, aren't like some of the lists in the DIY section, you can pin just one of the items as an example, right? As opposed to pinning the whole list. Aaron Powell
11:24 Speaker 0 Yeah, you can pin an individual item, and then also we make these title cards for for longer articles that are very Pinterest friendly that let you know there's more, so people on Pinterest can say, oh, this is a recipe, and I can click to get the full recipe.
11:38 Speaker 1 It was funny. I was reading the New York Times the other day, and it was like the weekend one, and they had a thing on like summer, you know, summer salad recipes or something. And I read it, and it was sort of like basically they went through and listed five summer recipes, but they did it in prose format the way the New York Times does. Um, and you think about it, you can do it like what one question people always have, you know, people associate BuzzFeed with lists, listicles, and uh and you know, why do you like lists so much? I think one tell me if I'm wrong, but one reason is you can take the different if it's you know, as opposed to pros, you can take the different parts, share them separately, discuss them separately on Facebook. It's a much more sort of social format. Aaron Powell
12:13 Speaker 0 Yeah, and and also mo a more mobile format because slideshows were pretty dominant, which would game CPM advertising and would there slideshows are just particularly on mobile very difficult. Like with with a list, you can just with one motion you can slide through. Yeah. And then also you know when you're done with the story, you know how many how what you're getting into, you can skip things that are boring to you and go to the things that are interesting to you. Um but it's not that we we have any religion about around lists. I mean, we we really don't see ourselves as a company that is wedded to any particular format. And what's so exciting about not having a legacy business is that is that we're not tied to the constraints of any particular medium. And so I think when you look at even the first generation of online media companies, they they tended to mimic print, even though they didn't have to.
13:04 Speaker 0 So articles tended to be the same length as as newspaper articles, or sometimes they do a long one, which is like a magazine article, and and you know, video tended you know, to you know, to sort of mimic television with like Hulu and even Netflix, where it's kind of looking like more like television. And what's what's interesting to me is that none of those things are predetermined or but when when you're making media for the web. So you can make a list, you can make a quiz, you can do something super short, you can do something super long, you can do the little slidey things that we started doing recently where you can s you can slide between two images and compare them. Um then you can keep tweaking and changing the formats to fig figure out new new ways of communicating that don't don't that aren't constrained in the same way that they were with with this very standardized industrial media distribution.
13:55 Speaker 0 Um so I think that is going to be a really exciting thing looking back, you know, ten years from now when you say, oh, look at all the different forms of media that emerged, you know, in the next one.
14:07 Speaker 1 It's funny what people people really get so accustomed to certain things, and that becomes that becomes okay. So for example, you know, you have horoscopes and you know, dating columnists and advice columnists and crosswords, and those are all of course little games and diversions and comics and things like this. Um yet because those have existed in newspapers for you know decades, it's considered okay. But then you know, you have slidey things and lists and everything else, and because it's new, it's like, oh God, the web is destroying Western culture.
14:39 Speaker 0 Yeah, so we we sometimes get get criticized for things and and the argument goes along the lines of how can I take you seriously if you have a slidey thing on your site or if you have a list or quiz. Um yet when you look across all traditional media, there's always been a mix of serious things and crosswords and cartoons and things like that. But when things are new, they they stand out more. They they stand out more for sure. Trevor Burrus, Jr.
15:02 Speaker 1 Let's talk about native advertising. You call it branded content. Um just can you describe what that means to you first, I guess. And
15:09 Speaker 0 yeah, I mean, back in my previous company, which was the Huffington Post, we we I didn't know anything about advertising, neither did my partners and and when we first launched, and we did all this tech development in to build the platform for Huffington Post that the editors would use. And they had all these special tools and all these things that they could do. And um the advertisers would come to us and they would just get banners. So all our tech development did not go, did not benefit us as a business at all. Um at BuzzFeed I wanted to avoid that, where I wanted one unified platform for distributing media, and that unified platform could be used for news, it could be used for entertainment, it could be used for branded content. Um and then we still created separate teams to you know, separate team works on the advertising then works on the editorial, but they get the benefit of the data science and all the cool formats and all the different innovation we're doing on the platform side.
16:03 Speaker 0 Um so the idea was to do really what Google did with their product, which is make the advertising a version of the editorial product where where it's powered by the same thing, but have a church-state separation. So the the organic search results are relevant results for a keyword query and the branded um results are relative. Just the critical thing
16:22 Speaker 1 to understand is that Google ads are relevant in the sense that they relate to what you search for. Um but they're not about tricking the user. They're clearly marked. And now, of course, when they first started off, people didn't fully under. I mean, it said sponsor but didn't fully understand. Now over it takes years and people get trained and they understand they're separate. But for the most part, like I click on Google Ads all the time, they're actually really relevant. It turns out that an auction is a good way to get relevant content to a keyword in addition to your algorithmic results.
16:51 Speaker 0 Yep.
16:51 Speaker 1 And similarly, like your in your case, it's not you're not trying to be relevant, you're trying to be compelling, right? With entertaining lists and videos and things like this. Um I don't know. I think it's funny, like people in Silicon Valley, uh very widespread adage is to make things that people want. Yeah. And yet somehow when it comes to monetization, that they they forget that and they put up a teeth whitening banner ad, you know, or whatever. These really for the most part, I think is is is uh one of the you know is uh is a very poor experience is the is the banner ad kind of monetization layer of the web.
17:25 Speaker 0 Yeah, I think part of the part of the reason that that Silicon Valley sometimes has that problem is that they're so allergic to ever paying anyone to create content, you know, just because the the holy grail is is this purely scalable business where nobody is actually making content and you're just making the uh the platform entirely. And so as a result, you end up with the some third party making teeth whitening ads and pumping them through your platform, and then you kind of do a policy level of saying, oh, we're gonna have some ad quality control or whatever, and you try to limit it. At BuzzFeed, we say we'll do a full stack of helping the brand make compelling branded content, and they often want to do things that aren't good for the user and aren't in their own self-interest and would be bad for the BuzzFeed site and bad for them as a brand, and we tell them don't do that. Here it will help you do something better.
18:12 Speaker 0 Um I think in a way it kind of goes back to you know, the uh uh a magazine like Vogue has really great editorial fashion photography and then really great branded fashion photography. And if you ripped all the ads out, the magazine would be a worse magazine, and the the branded content actually in the advertising actually adds to the product. Um that's a really high goal. You know, the Super Bowl is kind of another example people use, like the ads actually add to the entertainment value of the of the Super Bowl. And so
18:39 Speaker 1 TV is the original need of advertising, right? Or at least at least one of the original ones, right? Because the ads are compelling, they're the same format, they're video, you know, 30-second whatever videos. Um,
18:51 Speaker 0 for sure. I mean it it's it's not it's just it just feels like a historical anomaly that you had this weird fish and foul kind of thing on the web where banner ads looked not didn't looked and feeled and worked totally different than the the rest of the content on the site and just felt bolted on. Um and
19:09 Speaker 1 it's in my view at least, and I don't know if you agree, it the decline of that very bad business model is going to accelerate on mobile because on mobile you simply can't fit all that stuff. It's a terrible, it's even a worse experience. Um I would wager when you look at the click-through rates on mobile banner ads, it's probably mostly accidental clicks. It's like 0.2% or something, and a lot of it's like fat finger kind of stuff. Yeah. Um so basically it's just not working and it's going to collapse very soon. And then you're left with either what are you left with? You're left with a paywall model, which the New York Times and the Journal are, you know, another and the FT maybe are experimenting with. There's some other experiments going on with like crowdfunding and things. And then there's native advertising, and that's it.
19:51 Speaker 0 Yeah, and the payroll.
19:52 Speaker 1 Or hope a billionaire buys your your newspaper, because that those are the only models, I think. Aaron Powell
19:56 Speaker 0 Yeah, and the paywall model is unfortunate. When you think about news being a in the public interest. Where you know every time there's really serious news, the news they have to take the paywall down.
20:09 Speaker 1 The rest of the time it doesn't matter. But
20:12 Speaker 0 your news, if news matters, you want it to be out there. And you know, by the way, like you know, we're not even selling in terms in jacency the same way that that traditional media has. So you can have branded content that finds its own audience and and that doesn't even doesn't even necessarily ride along with news content. So everything is kind of unbundling and spreading along these social networks. And ideally with great technology and great content, everything is reaching the audience that it should reach.
20:45 Speaker 1 Which I think somehow have become synonymous in some people's minds with native advertising. Um I think that'll change very quickly.
20:53 Speaker 0 Yeah, it I mean uh we find sometimes uh people criticizing BuzzFeed are actually criticizing someone who's copying BuzzFeed, which is frustrating.
21:02 Speaker 1 What so the next couple of years, what are you most excited about?
21:06 Speaker 0 Uh there's so much to be excited about. So uh video you know is is some is an area that that has just exploded for us. We didn't have a video business two years ago.
21:16 Speaker 1 Um can you just just very high level explain why you know we're hearing video more and more in Silicon Valley and just generally on the web. Like what is it just the transition from TV? Like why video?
21:27 Speaker 0 Well, for us, I I was I I have been uh frustrated with video for years. Okay. And I've been frustrated with it because you we haven't found a way to do the rapid creation of content and then constant learning from the audience. And that's something you could do with text because it was cheaper to do. And so you could make more content more quickly. You could like play more hands of poker or I don't know what the right metaphor is. Um and video was so expensive that it was diff it was difficult to try ideas and experiment. And Zay Frank, our our our president of BuzzFeed Motion Pictures, you know, he he joined through an acquisition of a small startup company and he went from four people to 100 people and build it, built out a really tremendous team that has figured out how to um make video production much more web web friendly than ever before.
22:13 Speaker 0 So we have a studio lot in Los Angeles that's a four-acre studio lot. We're built we've built out sets for all different kinds of of scenarios. So a school, an office, a cafe, a dungeon, uh, you know, whatever you want. And then we have a team that can that brainstorms ideas for concepts they want to do and then quickly can shoot videos, release them, and then learn from how how people respond. Everything from Facebook and YouTube comments and sharing and and and um you know how many views it gets, but also the way people engage with it, and then learn from that and then immediately go back and make something else. And so it's become all um we we have the the fixed costs of our studio already spent, and then the variable cost is almost nothing. So it flips the Hollywood model where nobody, you know, the studio lots essentially become banks and and they will try a project and 100 million dollars will be spent to to make a film, and then people come from all over make the film and then they disperse and there's no learning and they go off to you know their separate ways after the project's done.
23:13 Speaker 0 We've been able to figure out a way to flip that so that we can try lots of ideas, experiment and learn through time. Um that's one reason why I'm super excited about video. The other is Facebook video, um, mobile video view consumption, all these sort of trends are starting to converge where social, mobile, and video are really converging and it allows us to reach a much larger audience with the video that that we're creating. Um so video is really exciting, international is exciting. Uh um we you're one of the biggest sites in the UK now and and have the the the if you if you want to see what the UK site is like, there's the you can toggle and look at the UK version. It's like I looked at the
23:52 Speaker 1 rugby and we don't we don't do so much sports but the
23:55 Speaker 0 it's a little it's a little smarter, a little snarkier, uh the sense of humor is a little different. But I I really like love the British the British site. Um and we're opening up other offices um and and um what else? I don't know. There's a lot, you know, a lot of tech a lot of tech things and data science things that are exciting. We're starting to really be able to map out how content spreads across across networks. So one of the things that's there's kind of blind spot for social networks, which is that they have huge amounts of data, they have huge number of logged in users, but they're only seeing sharing within their network. And what ends up happening is sharing within their network is heavily influenced by sharing that happens outside their network that they don't even see. You know, so if something is blowing up on on Facebook, they might not realize that a lot of the people sharing it are discovering it on Pinterest or discovering it on Twitter and then sharing the content they discover on one network with another network.
24:46 Speaker 0 So you're
24:46 Speaker 1 one of the few people with the scale and the cross network reach to see that.
24:50 Speaker 0 Yeah, so we can start to see, oh, we're we're we're actually able to understand how how um something taking off on one network bleeds over to another network. Often Facebook's a beneficiary of it. Something will you know, take off on on a social network and then um quickly start to be bigger on Facebook than it was on the original place where it took off. Um but but that kind of cross network sharing where you're really looking at social networks um that that not uh social network in the sense of a tech company is a social network, but the actual human social network of sharing that goes across many networks. Um so that's another interesting area where I think um both for making bet content that is better quality, so you're not ever looking at um one social network as the as the arbiter of quality, because if you do that, you might be there might be a quirk in an algorithm or a flaw in a social network.
25:38 Speaker 0 But if someone is sharing it across three or four, you actually realize okay, this is this is something that people like in different contexts, so it's a stronger quality signal. So it allows us to make better content, but it also is powerful for advertisers when they're making content because they they want to you know understand how their content is spreading between different networks.
25:59 Speaker 1 Okay, great. Uh we're we're out of time. Thanks for thanks for chatting.
26:03 Speaker 0 All right, thanks.
Previous Company Building in Crypto: A Conversation between Brian Armstrong and Chris Dixon Next What's Next for Education Startups Page updated August 25, 2026 · transcript and summary are machine-generated