The visionary media entrepreneur, diviner of virality, and one-time enfant terrible (he turns 40 on New Year's Day) explains the mechanics of how information spreads, his personal reading habits, and why it’s dangerous to be a slave to the numbers.
FORTUNE — People in the media industry are buzzing about entrepreneur Jonah Peretti, and it’s not just his latest venture, BuzzFeed, that has them talking. The Huffington Post co-founder has made a name for himself exploring the ways information spreads online, turning an academic interest into a highly lucrative series of businesses that are prompting people in the media industry to rethink the way they distribute content — especially online.
The inspiration for this conversation came from an interview of Peretti that I conducted at this year’s American Magazine Media Conference in New York. Here are the highlights.
Serwer: By now, everybody knows what BuzzFeed is, or at least has been to the website. What exactly, to your mind, is BuzzFeed?
Peretti: I think I maybe see BuzzFeed differently than a lot of our readers see it. I think a lot about how ideas spread, how information spreads, why is it that something you’re really proud of and you spend a lot of time creating sometimes doesn’t go anywhere, and something that you kind of do on the side, on a lark, ends up getting shared and passed around and having this big impact. And so I think of BuzzFeed as this platform that enables us to understand how people are sharing and distributing things like entertainment content, journalism, branded content, all these various types of content that we distribute on this platform that we built.
When we started, we really were a laboratory trying to understand how this stuff works. Now, we’re a team of really dedicated editors and reporters and people creating entertainment to distribute across the web on our site.
Where do you want to take BuzzFeed? You don’t characterize it as a journalistic enterprise per se. What is it most analogous to in the traditional media world?
I think that there are certain companies that, on the web, you can compare to a traditional entity. So you can look at eBay EBAY and say there were auctions in the real world, and now there are auctions online. Or you can look at Amazon AMZN , and you can say there were stores that people go to, and now there are online stores. In fact, at The Huffington Post, we talked about the “Internet newspaper.” It was kind of comparing the site to a [printed] newspaper.
One of the things that is counterintuitive about BuzzFeed is that there’s not a natural corollary to what we’re doing because it isn’t possible to distribute content through word-of-mouth in print. People rip out magazine or newspaper articles and mail them, and there’s some small amount of distribution, but we’re reaching 85 million unique visitors without owning a printing press or a broadcast pipe or anything. And every day we have to make content that people think is worth sharing, and we don’t reach any audience at all unless we make content that people think is so good that they’re willing to pass it on and share it with all their friends. That model doesn’t have an offline version of it.
That does make it more confusing. It’s easier to understand a business when you say, “Oh, it used to be done, you know, this way, and now it’s done more efficiently using the Internet or using technology.” What we’re doing you couldn’t have done without the Internet.
BuzzFeed is a commercial enterprise, a business. You are interested in making money, ultimately, and in growing it. Right?
Yeah. We became profitable this year, and we’ve been growing our team. We want to build a sustainable company. We like to think, What would a media company be if you created one from scratch today? What would a media company be for the era of social and mobile? [What would it be like] to build that company? Generating revenue is an important part of building a company. So is having great investigative journalism. So is having really entertaining content. We have a pretty broad purview — a pretty broad range of things we do.
You’re really all about leveraging the viral aspect of the Internet. That’s what you started out with at MIT, with that Nike iD thing where you were denied a request to put “sweatshop” on your shoes and the story went viral. And your friend said you couldn’t replicate that. That was your “Aha!” moment. So social platforms are, in a way, your lifeblood. Tell me about the landscape of social media, which social platforms are the most important to you, and which ones are you most sanguine about.
What we’ve found is that content spreads on different networks for different reasons. There are underlying human dynamics for social content. There are reasons why people share. But certain platforms are better for certain types of sharing. Twitter TWTR is very fast. Partly because of the architecture of Twitter, things flow so quickly and disappear so quickly that you need to post things that are of the moment. And so that’s why Twitter is great for things like live television, breaking news, and real-time events. Twitter is also much better for interest-based content. So people on Twitter can follow tech if they’re interested in tech, or business if they’re interested in business, or they can follow celebrities that they’re fans of.
On Facebook FB , because people use it as their actual network of people that they’re friends with in their real life — you have friends from college, you have friends from high school, you have friends from work, you have a diverse range of people that you’re connected to — you don’t really want to share things that only a very small subset of people would be interested in. So if you’re tweeting about running, you can write about running and people who are interested in running can follow you. But if you start posting on Facebook constantly about running, and most of your friends don’t care about running and actually kind of hate you because you’re a runner, or at least it makes them feel bad about them not being a runner or whatever, you know, it doesn’t really work that well. So Facebook is much more tied to broad human emotion and things that everyone can relate to, and things that connect people with the people in their lives. It’s not so much about the information in the content; it’s about how that content allows you to connect with other people in your life.
And then something like Pinterest is much slower. People are actually bookmarking things in Pinterest — making boards of DIY projects they want to do later. Things spread at a slower rate, but there’s a lot deeper engagement where you’re actually cooking a recipe or you’re building a dresser or you’re doing something that you have to wait for the weekend to have enough time to do.
And the list goes on. How do you tailor your business or your content to each one of these social platforms?
Well, part of it is the type of content. We think about that on a human level — we’re not like, “Look at Pinterest. Let’s make the perfect Pinterest. Let’s game Pinterest’s algorithm.” It’s more, “Let’s make something that people are proud to share. Let’s make something that people see and say, ‘This is going to be more fun with someone else.’ ” So things like humor, things like cute animals, things like breaking news, things where when you engage with the content, you immediately want to share it and bring other people into the conversation.
On one hand, you are incredibly data-driven. But you also rely on your gut, right? How do you balance those two things? How much is gut-driven and how much is data-driven?
People often say, “I go with my gut,” and they forget that their gut is informed by huge amounts of data and past experience. It’s like if you’ve launched five magazines that have failed, and then someone wants to launch a similar one, a sixth one. You’re like, “I’m going to go with my gut. This isn’t a good idea.” That doesn’t mean that you have some deep insight. You actually have these five very strong, painful data points that have informed your gut. That’s why when you look at someone senior in any company, almost always their gut is to do the thing that they had success previously in their career doing. They want to do that thing again and again.
It’s good to not always trust your gut, to have some skepticism about it, and to constantly have new data coming in and constantly question what you actually know. A lot of what we do at BuzzFeed is give dashboards to every person who works at BuzzFeed where they’re seeing how people are engaging with the content they’re producing: Is it going up? Is it going down? They can test hypotheses and theories and say, “All right, this worked in the past. Can I make it work again? Can I do something similar? Can I take one part of it and make that work?” And you start to learn through having a healthy skepticism and lots of access to new data.
You always hear people who say, “Well, we can’t be a slave to the numbers on the web because if we only follow the numbers, it will lead us down some path.” You don’t have that problem at all, do you? Don’t you only follow the numbers?
No. It’s dangerous to only follow the numbers. I think there’s a lot of over-optimization on the web. You see this sort of “side boob” trap or something, where you put some picture of a celebrity whose dress lets you see the side of her boob on the front of your website, and you say, “Wow! That gets a really high click-through rate.” If you were a slave to the numbers, you’d start putting more stuff like that and more stuff like that and more stuff like that. And pretty soon you would have a site full of trashy, salacious garbage, and you would say, “Oh, I’m just looking at the numbers,” but you would be hitting a local maximum, where lots of people would never want to read your site just because 10% of your readers are horny guys who can’t resist clicking, or women who can’t resist gawking at celebrities.
If you made something that was compelling and touched people in a deeper way, you might actually … there might be a mountain that you could climb, but if you’re just following numbers, you think you’re at the top of the mountain because every direction you go looks like it’s downward to a lower click-through rate. But there’s actually something there that’s much bigger. Optimization can lead to finding local maximums, but those local maximums often aren’t where you actually want to be. That’s why exploring lots of different kinds of content, and thinking about things on a human level and saying, “Is this actually good content, or are people just clicking this because they can’t resist clicking it because it’s a guilty pleasure?” is important. You need to have creative, experimental people trying lots of different things. And then the data becomes meaningful.
My college-age daughter loves BuzzFeed. What is your audience? What do you want your audience to be?
We didn’t really start with a demographic in mind. We just were focused on making content for the social web, with sharing being our distribution. Mobile and social really converged. So, it used to be that mobile was the thing that would stop things from spreading. I used to hate mobile. You’d get the email on your BlackBerry and say, “I’ll look at this when I get back to the office.” Now the most popular apps on mobile devices are social apps. When we see traffic from Facebook or Twitter, it’s disproportionately mobile.
So we were focused on making media for the way people consume media today, for mobile, for social. And then we kind of were looking at our contour numbers, and we were like, “Oh, wow! We have this super-young demographic. Why do we have this super-young demographic?” The reason was that we were making media for the way people were consuming content today, and young people are the earliest adopters of using smartphones to consume media, and of using the social web as their main source of content. So we kind of accidentally ended up with this demographic of readers in their twenties and thirties who are very active on social media, who are very educated, who have high incomes relative to the peers their same age, but not relative to their parents, and who share content at a really high rate and engage with content in a very voracious way.
Do you want to get older? Do you want to get broader? Is it something that you seek to do?
I think we will. I mean, it’s hard to predict the future. My guess is that a lot of the media consumption behaviors of young people will start to spread to a broader audience. I actually don’t think we’re demographically defined. It’s more a way of consuming content media, and I think there are older people who do consume media that way, and they’ll probably increase. That will be a growth area for us, just the same way that older people were a growth area for Facebook.
How many people go through the front door of BuzzFeed.com? Does that speak to the tension between search and social?
We have millions of people go to our front page and tens of millions of people go to our “B” pages and articles. Social has really become the starting point — the Facebook newsfeed, the Twitter stream — where most people are finding their news and information now. It’s become something that has become more and more prevalent, particularly for young people. And we are in those streams at a really high rate. We’re one of the most popular sites on Facebook and Twitter and other social sites.
People go to the front page of BuzzFeed partly because they’ve seen a bunch of things in their stream, and they’re like, “Oh, I like this site. Why don’t I go to the source?” I think that happens. But also people are going to look for something to share. So people come to the front page of BuzzFeed saying, “Oh, that article sucks. This one’s not for me. This isn’t my cup of tea. But this one is perfect, and I’m going to share it with all of my friends.” So the eclecticness of our home page almost is a benefit because it lets people pass over things that isn’t the right content for them, find the one that is, and put that into their stream and get the sharing going on the social web.
Do you care about search at all anymore?
We don’t spend that much time thinking about search.
Isn’t that amazing, that SEO is not an important part of what you do?
It’s a tricky thing. I mean, I think that search will at some point have to catch up to social, you know, to understanding how to index and present content that is getting shared at a high rate. I think there’s a problem now where Google GOOG doesn’t have a lot of that social data. The aggregators win on search where people cram in key words and do SEO and rewrite stories. But on social, you want to share the authoritative story — the Ben Smith scoop is going to get re-tweeted a bunch of times, not the rewrite of the Ben Smith scoop. But the rewrite is going to be what ends up getting a lot of the traffic in Google. That’s the thing that Google is going to have to figure out how to fix over the next years.
Interestingly, it’s not like Google’s suffering right now particularly. At one point, I thought that search was going to kill search. Can both pies grow?
Yeah. It’s always hard to predict. BlackBerry BBRY had record profits right before it all started to fall apart. Google could have another 10 years of rapid growth and success and drive us around in cars and tell us what to think and look at with glasses, or they could have real problems. It will be interesting to see.
What percentage of BuzzFeed content originates from BuzzFeed?
We have four types of content. We have news content created by our team of journalists and reporters. They really are not focused on aggregation; they are calling people, working sources, doing original stuff. We have entertainment, often characterized by our lists. And those lists are all original creations, although they often use material from, you know, AP and Reuters and Getty and image libraries and other sources. Then we have branded content, which is how we make our living, where our revenue comes from. Fifty of the top 100 brands used BuzzFeed’s platform to launch branded content. And we also have community content, which is where you can go to BuzzFeed, go to the community section, and create your own content and launch it. Four different types of content, all on top of the same technology platform that we developed.
Let’s talk about that branded, sponsored content. You treat that the same way you treat editorial in terms of optimizing the social component, right?
Yeah. A few years ago, our traffic was starting to grow, and we had a board meeting, and my board asked me whether we had any revenue. And I said no.
So I started thinking about how to generate revenue. The idea of putting banners on the site never appealed to me, in part because we were investing in this platform for distributing content on the social web. Our team was using it to make fun, entertaining stuff at that time — we didn’t have any reporters at that point — and we were looking at how to really detect whether we were connecting with an audience, and whether people wanted to share this content, and how to give people content that they were excited about to share.
So banners didn’t make much sense to me. It made much more sense to build a version that uses the same platform for branded content and advertising. And so we told brands, “You have to tell a story.” This is actually something the magazine industry has been great at over decades — making advertising that actually adds to the product. It’s something that websites have completely failed to do, and largely it’s the fault of trying to cram things into little banner ads. If you take a high-end magazine and rip all the ads out, it’s often a worse product. If you take all of the ads out of a fashion magazine, you lose half the photography, you know? So we really took the approach of, “Well, why can’t the web be like that? Why can’t we make great branded content, advertising, that has its own page that people want to click on and engage in and share and interact with?”
We created our own form of advertising that was totally proprietary. It was a real hard slog trying to sell it. People were like, “Well, look, it’s not IAB standard, and it’s not the way the industry works. And we don’t really like these banners, but it’s just sort of the way it works.” But we had a few people, including Beth Comstock, the CMO of General Electric, come to us and say, “Well, we want to innovate. We want to experiment.” General Electric GE is a giant company. At the time, we were very small. And we’ve been working with them ever since.
A bunch of other brands saw some of the early success we had and said, “Oh, look, if it works for GE and they’re getting 10X the click-through rate of banners and they’re getting sharing and they’re telling a compelling story, then we’ll try it out, too.” Once people try it, they renew and renew and renew.
What is BuzzFeed’s relationship with magazines?
We’re partners with some magazine websites. We have something called a partner network that lets other publishers use some of our technology to detect when something is taking off. So if you work at The New Yorker, for example — we are a partner with them — you can get an alert that shows that one of the stories on their website is starting to take off. They can have a dashboard that shows some of how their content is spreading.
If you look at the right-hand rail on the front page of BuzzFeed, you’ll see the publishers in our network, their most-shared stories trigger and show up on the site. So that’s one way we work with magazines.
Will you establish more partnerships like that?
Yeah. There are over 400 million unique visitors in the network now, and all of them are getting some of the benefit of our data science and our technology.
When you see magazines’ websites, what do you think is right, what do you think is wrong, what do you think they could do better?
The main way that I see magazine websites is someone who has tweeted a story or put a story into my Facebook newsfeed. The biggest challenge I see is that some magazines have a very bad mobile experience. You’ll click to see a story, and then there will be some kind of desktop-looking pop-up ad or overlay ad with one of those little x’s, and you’re trying to read the story, and it’s hard on a mobile device to find out where the little x is to get rid of the ad and see the story. At the bottom of the article, it will be paginated, and it’ll be like nine pages or something. So you have to find the little number 2 to click that. And when you click that, it will pop another one of those ads up. It has to be a really good article for you to go through that kind of [experience].
So, test all your stuff on mobile devices. Apps are important, and I think apps are part of the pedigree of the magazine industry, which is where subscriptions have been so important over many years, but mobile web is also really important. Because if I have an app for a magazine, and I share something, that’s going to Facebook. The person is more likely than not accessing Facebook on a mobile device, and if they can’t read the article, then you missed an opportunity to bring in new readers.
What magazines do you read, what magazines do you enjoy, historically or today?
I really like reading on the web because you get the best of all different sources. So really, Twitter and Facebook are the main ways that I find stuff. Sometimes e-mail. We have a long-form section on BuzzFeed that has, you know, 12,000- and 6,000-word pieces, and we do a long-form e-mail that links out to the best long-form pieces each week. That is another way I discover a lot of long-form magazine articles — through the long-form editors at BuzzFeed who are reading stuff. We have some original stuff, but usually it’s one of ours and three or four from other magazine sources.