Chris Johanesen on design, testing and launching new products at BuzzFeed
Kyle Ellis: You started at BuzzFeed as the company’s second employee. Tell us about your time with the company and how your roles have evolved over the years.
Chris Johanesen: I was hired as the first designer at BuzzFeed, and for the first few years we were a small team and functioned much more like an R&D lab than a real media company. Jonah [Peretti, founder and CEO] has always been a very product focused founder, but as we grew, and Jonah took on more of the CEO duties, I gradually took on more of the day-to-day product duties. It’s been a very gradual thing and there was never really a big shift in my responsibilities. The biggest change for me was hiring designers and product managers to work under me in the last two years.
As a designer I’ve never really had a big distinction between design/UX/product/technology. It’s always been about solving challenges and creating the best product experience possible within whatever technical, time, or resource constraints we were faced with.
K.E.: Aside from BuzzFeed, where do you get your news?
C.J.: Like most people these days I get most of my news from articles people are sharing on Twitter and Facebook. Now that we have a full news organization I’ve actually found myself checking our homepage to see what is going on in the world whereas before I would probably go to nytimes.com to scan headlines.
K.E.: Describe the structure of the product team, and how it work with other groups within the company.
C.J.: We have product leads who own certain products or product areas and focus on improving them. The product leads are in charge of taking in feedback from others in the company or outside of the company and planning new projects. There’s a design team, lead by a design director who works with the product leads to design new features. We also have a senior project manager who keeps everything running and a support specialist who helps keep our internal and external “users” happy.
Our team is still pretty small and we are trying to grow it smartly and in pace with the growth of our engineering team.
K.E.: How do you make GIFs, listicles and longform stories all work together in the same place? Those are such different content types that they seem at odds with each other.
C.J.: We’re not so much different from a TV network in some ways — we have our hard news, we have our entertainment, and we have everything in between. Our media is the internet and social media, not TV. We are the first large — by audience size, not head count — media company that could only exist because of the internet and social media. Most older media companies exist in spite of it.
Also, we think those things work together because people are complex and need lots of different things from their media. The idea that some people like serious reporting and others like fun GIFs is a false dichotomy. To borrow Jonah’s metaphor, you can be sitting outside a [Parisian] café discussing serious politics or philosophy and if someone comes over with a cute puppy you’re going to bend down and pet it.
Our site mirrors social media in that when reading your Facebook or Twitter feed you might see a serious article someone shared juxtaposed next to a cute picture of your friend’s baby.
K.E.: BuzzFeed calls itself “The media company for the social age,” how does it use technology to power the social distribution of content?
C.J.: We have a lot more technology than most people realize. BuzzFeed started as a technology company and it’s still in our DNA. We see our technology as giving our reporters and writers superpowers.
There are three main roles for our technology: make it easier to create great content, make it easier for content to spread on the web, and make it easier for people to find content they’ll love.
The first really focuses on building our CMS so the editors can focus on the hard job of creating stories. We work with them to implement new formats from lists to quizzes to long-form articles to games.
For making content spread we have developed our own sophisticated tracking architecture, which is geared around letting editors know, in realtime, how their posts are performing. We also optimize our article pages to make them fast loading and work great on mobile so there are as few barriers as possible to prevent someone from sharing something they like.
To help people find content, we have optimization algorithms that help show people stories that they are more likely to enjoy and share. The key is that all of our optimization is driven by sharing, which is something you can’t game. If people don’t like a story they won’t share it, so that keeps us true to our mission.
K.E.: Does BuzzFeed have user personas? If so, what are they and how do they influence editorial and design decisions?
C.J.: We don’t use personas in a formal way, but we are always thinking about our users and what creates the best experience for them. Among our users there are different groups we think about: the “first-time reader,” our dedicated homepage readers, our app. users, readers who comment and react, and community members who are posting their own stories. These groups tend to overlap but all have slightly different motivations and usage patterns.
K.E.: The homepage offers users multiple ways to browse and navigate content (sections, most recent stories, lol/win feeds), what led to that approach?
C.J.: When BuzzFeed first launched, the only navigation was a calendar-based archive. We had topical categories but they were never a focus. After a while Jonah came up with the concept of categorizing our content by emotional responses instead of topical categories, i.e. LOL, OMG, Cute, which really fits in with the way people share stories and became a big part of our brand and DNA.
When Ben Smith joined and began to build out our editorial and news teams we started adding more focused topical verticals like politics, tech, and entertainment. This was important for us to highlight the great reporting we were doing and promote the idea that we now have a wider breadth of content and news. It’s never been a big focus to divide up our readers by interest. We never put you in a box so you only see one type of content.
We also like to have fun with it, which is why we launched verticals around animals and nostalgia.
K.E.: Your homepage encourages and almost assumes lots of user scrolling. Do you have data that shows how far your users scroll? What tactics do you employ to encourage scrolling and browsing?
C.J.: We do measure this and find that most people get at least 80% of the way down the page. We encourage scrolling by not putting up barriers to scrolling like slides shows and article paging. I believe it’s just native to the medium to scroll (and even more so on mobile phones). Also having creative content helps. No one scrolls if they’re bored with the article.
Last year we leaned into it by implementing an “infinite scroll” on our homepage. If someone finds an article they want to read at the top of our page, that’s great. If they don’t, we want to keep showing them content until they find something. There’s no reason to cut them off for some arbitrary technical reason.
K.E.: How do you develop and iterate on new products and features?
C.J.: We have a very iterative design and development process. On the product team there are three tenets that we live by:
Do what’s best for users
Good UX makes everything else work; bad UX ruins everything. All of our metrics are dependent on our users being happy.
Always be iterating
“Minimal Viable Product” is a buzzphase, but it is a useful one. The faster we can get something in front of real people the better. The more complex a project is, the longer it will take and the more will go wrong. We’re always thinking about how to simplify, and we’re always thinking about how to improve products in small steps.
Combine intuition and data
Great decisions are always made using new data and past experience (intuition) combined. You can’t A/B test a bunch of bad designs to come up with a good one, but testing is a great tool to validate your designs and help you choose between different options.
It also helps that we have a great engineering and operations team that our product team works very closely with. We deploy code every day (often multiple times a day).
K.E.: How do you test new features?
C.J.: It really depends on the feature and the audience, but we use A/B testing as much as possible and limit rollouts to test features with a subset of users. For features on the site, we might push it out to 5 or 10 percent of readers, look at the data and feedback, tweak if necessary, and then test again. We only roll out to all users when we feel confident that it’s a good change. For things like CMS features we’ll give a new tool to a few editors to get feedback, then roll out to all of editorial, then eventually to community and clients once we’re really happy with it. We also do some informal, in-person user testing in the office; usually when things are in early design or prototype features, which is especially useful for user interfaces.
K.E.: What do you think the journalism industry could learn from BuzzFeed, and conversely, what do you think BuzzFeed could learn from the journalism industry?
C.J.: I’d say a big part of our success is our willingness to experiment and take risks. Jonah built this into our culture so it really drives our product and I’ve seen how Ben Smith really encourages this on the editorial side, too.
One thing we are figuring out now is how to have a big homepage. Before last year we really thought of our homepage as the social web, but after we grew our news team and started covering big news events we realized a lot of people come to BuzzFeed to find out what’s going on. We’ve been thinking a lot about our homepage (in terms of editorial, product, technology) and how to make them work better for a rapidly growing audience. (And our mobile app user base is growing as fast as our homepage so that is just as important for us to improve and grow.)
K.E.: What type of feedback do you get from your users and how does that influence design and product decisions?
C.J.: We get a lot of feedback via data … What things do people click? What types of content do people share? If we’re testing a design tweak on the homepage we’ll see how often people comeback, how many articles they read and share.
But we also get more direct feedback. For CMS features we’re lucky to have a big pool of captive beta testers in our editorial staff. And our readers and community do tend to write quite a bit.
K.E.: What applications do you have open while you’re working?
C.J.: Well, I don’t do a whole lot of designing any more, but I usually still have Photoshop and Illustrator open all the time. Sometimes I use Keynote for quick wireframes and mocks, and a lot of our designers are starting to play with new tools such as Sketch and some online prototyping tools.
The most important applications are communication tools. We use Pivotal Tracker to track tasks and projects that are in development, Basecamp to share and collaborate within the design team, and Slack to group chat among the whole tech team. And of course, lots and lots of email.