Cristina Fuser is VP, Product at BuzzFeed. She began her career as a marketing analyst at Glu Mobile before joining Electronic Arts (EA) as a product marketing manager. Prior to joining BuzzFeed, Cristina held various product and leadership positions at Arkadium, Eko, pq by ron arad, Refinery29, and VICE Media.
In our conversation, Cristina talks about how being tech-led and data-led can be a strategic advantage, particularly in the media space. She discusses how she promotes innovation via “logical creativity” — where anybody should be able to logically follow the context and expectations of end users when implementing an idea. Cristina also shares examples from several of her roles of times when technology and data have transformed the user experience.
BuzzFeed is one of the first digital media companies. Our founder and CEO, Jonah Peretti, is an MIT Media Lab graduate who’s always been fascinated by “contagious” and “viral” media. Back when HuffPost, which he co-founded, was thriving, he launched BuzzFeed as an experimental lab to push the boundaries of what could be done with media and data.
BuzzFeed grew alongside social media and the company figured out, very early on, how to hack growth through the viral spread of information on social media. Most millennials and younger generations grew up with Buzzfeed — taking quizzes that told us exactly who we are, revealing fun things about our identities, and sharing cultural moments and pop culture.
Today, like every other media company, we’ve gone through many transformations and iterations. We’re currently in one with AI. BuzzFeed is the most resilient of the media companies I’ve worked at — one that knows how to reinvent itself at every pivotal moment.
We’ve always been “pro-data” at BuzzFeed—we get excited about it, and that mindset starts at the top. I mentioned that Buzzfeed started as an experiment. As our data and engineering capabilities grew, we leaned more into the growth hacking mindset that’s typical of startups and tech companies. Jonah has always seen data and tech as a huge competitive advantage in media.
From the beginning, we’ve been quick to spot opportunities in the data — those interesting growth signals — and then use tech to amplify them, turning them into sustainable traffic or revenue drivers. Jonah pushed data as a key to growth and a full feedback loop with our audience, especially via social media. We adapted our strategy based on that. Data has always been a key component, and tech was a big investment from the start, so they worked hand-in-hand.
I’ve worked for several media companies, and tech is an essential part of this type of business. In media, designers, product managers, and engineers are building the tools that make it possible to create the core product, which is content, connect the audience with it, and distribute and monetize it.
In most media companies, editorial teams set the priorities, and tech exists to support those goals. There’s room for impact, but tech is usually limited by the content strategy it’s supporting. Of course, tech enhances the experience, increases distribution, and differentiates the brand, but it can be more like a support function — a reaction to the editorial strategy. It’s also common for editorial and tech teams to have some degrees of separation, with strategies and roadmaps that can feel disconnected or even conflicting.
A few years ago, when AI became the most disruptive force in the media industry, tech was the one department that was positioned to lead the way. Our engineering and product development teams have taken the driver’s seat on this. Being tech-driven means that tech and product development aren’t just about efficiency anymore — they’re driving entirely new editorial opportunities.
To adapt to this, we reorganized so that tech and editorial now report to the same leader, our publisher Jess Probus. This setup has broken down a lot of barriers to collaboration. We’ve also embedded editorial team members into every tech team. It makes a huge difference in how decisions are made and how we build things. In fact, some of the best prompt engineers we have are actually writers. Ultimately, being tech-led is about embracing a “tech way of thinking.”
When I was working at a previous company, two engineers in our organization were frustrated because they knew what was possible with big data, but they didn’t have the infrastructure or investment available to make a difference. I feel lucky that BuzzFeed has always been invested in tracking, measuring, and storing data, and we use it to push innovations. So, when big data became a (big) thing, we were ready to take full advantage of it.
We had a data warehouse in BigQuery with massive amounts of user information, so we were able to build predictive models and machine-learning capabilities on top of that data. We knew what content did well, how it was distributed, and how it resonated, so we could then use those models to reproduce the type of strategy and processes that someone would have to manually think about.
What made this particularly powerful was that we weren’t just automating the process of pushing stories to platforms like Facebook, Twitter, and Instagram — plenty of third-party tools can do that. Instead, we built a custom model that used our editorial strategy to predict what kind of content would perform best on each platform.
Yes — we developed POUND, which stands for Process for Optimizing and Understanding Network Diffusion, in 2015. This was a fascinating initiative that tracked how BuzzFeed stories spread across the social web by following their journey from one sharer to another — even across different social networks and through one-to-one sharing like chats and email.
For example, we had a handful of super viral stories. You might remember “the dress” and the argument of whether it was black and blue or white and gold. It exploded. The data wouldn’t always tell the whole story about how from that initial seeding on Twitter, we ended up with millions of views. POUND gave us a much deeper understanding of how content spreads online and how to optimize it for maximum reach.
There is nothing more daunting than a list of creative ideas that don’t connect to the goal of benefiting the audience or the business. To contain this type of unbridled creativity, I prefer to support something that I refer to as “logical creativity” — where anybody should be able to logically follow the context, whys, and expectations of end users when implementing an idea.
That’s why I’m such a big fan of using hypotheses to guide problem-solving and decision-making. It’s what keeps me sane in the chaos. A strong hypothesis starts with understanding the audience, making some key assumptions about their needs and behaviors, and predicting how a new feature or change will impact user behavior or key metrics. From there, you can test it through experimentation.
I make sure the teams’ roadmaps are built around the handful of key hypotheses that they want to test each quarter. It’s up to tech leadership to review and approve these during quarterly planning, and at the end of the quarter, teams are evaluated based on whether they tested their hypotheses and what they learned. This approach keeps everyone innovation-focused and ensures that product teams are always learning and improving through hypothesis testing.
We’ve also borrowed the concept of hackathons from the tech world, but we call them buildathons. We want to make sure teams actually build something at the end of it. These are 2–3 day sprints where teams pitch ideas for projects they could complete faster with dedicated focus. Unlike other companies, we don’t schedule buildathons on a strict cadence. They’re encouraged year-round as long as the idea meets two criteria: it’s aimed at improving a metric we care about and it enables us to learn something faster.
At some point, my team taught me a lesson about this. During a meeting, I said, “I remember that we already tested this. Wasn’t it a failure?” One of my team members said, “I don’t think so, we learned a lot.” They nailed it. Ideally, that knowledge gets you one step closer to success.
When we launched an AI chatbot on Tasty, BuzzFeed’s cooking app, we called it Botatouille. The idea was that our audience might enjoy asking questions about food and recipes. We thought having a “personal chef assistant” could make the app feel more customized and helpful, especially for people trying to improve their cooking skills. It was one of the very first consumer-facing applications of AI we’d crested. We got Botatouille to ingest our entire recipe library on top of the broader internet knowledge you get from ChatGPT and similar AI tools.
Users interacted with it a fair amount, but it didn’t really move the needle. We learned that, for our audience, having a back-and-forth with a bot isn’t how they want to discover recipes. We tried a few iterations to make the feature simpler and more seamless, but in the end, we stopped investing in it. More often than not, with experimentation, metrics stay flat or may go down. At that point, the key is to dig into the data. Figuring out what happened and what to do next is the hardest — and most valuable — part of the process.
Yes. I want to start by saying that we are very lucky at BuzzFeed. I joined at a time when the data team had been built out to the extent that every member of my product team has a dedicated product analytics partner. Our data team is the largest I’ve worked with at any media company.
Our PMs work hand-in-hand with product analysts throughout the entire testing process — from setting up experiments and defining success metrics to interpreting the results. They go far beyond just crunching the data. They really want to understand and offer their interpretation of data, or, if neither they nor the PM understand it, they will start brainstorming different ways to go one layer deeper and get to the answer, which we always do. That is the most amazing thing.
Our analysts also play a big role in strategic planning. They help PMs prioritize initiatives by digging into past performance data. For example, after a test wraps, they’ll analyze how a feature might have impacted our metrics if it had been implemented six or 12 months earlier. That insight helps us forecast future impact much more accurately. Considering that most media companies don’t have A/B testing tools, I’d say our data infrastructure and talent are a huge competitive advantage.
That said, data doesn’t always tell the whole story. At the end of last year, HuffPost launched a program inviting users to donate money to support free journalism. Intuitively, we felt this was the right move given the politically charged climate and the loyalty of HuffPost’s audience. But we were also cautious since a previous paid membership program (offering ad-free access and exclusive content) hadn’t been very successful.
This time, we were asking readers to support us financially with nothing in return — no perks, no exclusives. From a data perspective, we didn’t have much evidence that it would work. But it did. The program has been a huge success, surpassing all expectations and becoming a meaningful new revenue stream for the company.
I don’t believe in creativity just for the sake of creativity. Sometimes, it’s hard to say no, especially when things are coming from the top down and people fall in love with an idea. I prioritize taking a pause and asking, “Who am I building this feature for? Why do I believe it will work? What exactly is it going to do?” We don’t have just one BuzzFeed audience or Tasty audience — we have many different audience segments with various motivations as to why they come to us and use our products.
With Tasty, we had a challenge that was similar to that of Netflix: decision paralysis. It’s so easy to just keep scrolling. On Netflix, you can watch six trailers, read the descriptions of what each one is about, and then you’re exhausted and don’t even want to watch a movie anymore. We saw that trend in people taking too much time to look at a recipe itself, watch the video, and learn how to make it. Our goal was to shorten the amount of time between opening the app and clicking on a recipe.
We decided to focus on beginner cooks because we know that’s a large audience for us. We saw, through data, that this segment favored shorter recipes that didn’t take a lot of time to make. Information about prep time wasn’t clear based on the image and title, so we decided to add the difficulty level and amount of time it takes to make it at the top. That gives people a sense of how complicated the recipe is.
We tested these hypotheses by surfacing that information on the home feed and saw that we were making it faster for people to do what they were already doing. This wasn’t innovation for the sake of innovation, it was putting the user at the center of the thinking process. Ultimately, it all comes down to ensuring the changes we make serve the user’s needs first, because that’s what drives satisfaction and, by extension, long-term success for the business.
LogRocket identifies friction points in the user experience so you can make informed decisions about product and design changes that must happen to hit your goals.
With LogRocket, you can understand the scope of the issues affecting your product and prioritize the changes that need to be made. LogRocket simplifies workflows by allowing Engineering, Product, UX, and Design teams to work from the same data as you, eliminating any confusion about what needs to be done.
Get your teams on the same page — try LogRocket today.
Want to get sent new PM Leadership Spotlights when they come out?
Jann Curtis talks about understanding your audience’s purpose and what they hope to get from the conversation.
Scaled agile is an approach that allows you to extend agile principles across multiple teams, projects, or business units.
“Disagree and commit” is a management principle that encourages team members to voice their opinions during the decision-making process.
Drew Doman, VP of Product at Apptegy, talks about leveraging tenets in product management rather than strictly relying on processes.
Learn more →
