The Keys to Successful Digital Experimentation

MIT Sloan Management Review: Tell me a little bit about your role as senior vice president and digital development person.

Kimberly Lau: First, I’ll just clarify, our parent company is Atlantic Media Inc., and I specifically work on The Atlantic brand.

Essentially all the areas of our business that are not sales or editorial report in to me. That includes the product team, which is composed of our developers, designers, and product managers, who are designing, developing, and strategizing about our website and all our digital products. It also includes our consumer marketing team, which historically was focused on print subscriptions but is increasingly focused on developing new products and growing our paid consumer base. And lastly, it includes the magazine production team, which is a very small team that literally handles magazine production, printing, buying paper, all of that.

This year, we’re building out a data team. This will be a group of people charged with servicing edit, product, consumer marketing, and sales and marketing. The team will both build out the infrastructure we need to have a complete customer database and create an infrastructure that will allow us to take in data and synthesize insights and really deliver those insights in an ongoing way with each of those different groups.

I’ve been at The Atlantic for five and a half years, and before that I was at Hearst Magazines, where I led business development for six years. I’ve been dealing with digital transformation for my entire career.

When you think of media companies, the story line is that they’re struggling with digital, that they’re having trouble reinventing themselves. But that doesn’t seem to be the case here. What is it that makes The Atlantic different — that you’re in a stronger position than you were five years ago? What have you done differently than competitors?

It’s been about 10 years since we started really embracing digital and focused on that as a key point in our transformation. And we truly are a digital company — the vast majority of our revenues are coming from digital sources. We have certainly crossed that path.

I believe one of the big things that helped us as a legacy media company was that the brand had been struggling when it chose to pivot to digital. That was helpful because it made us somewhat cavalier — we weren’t overly worried about what was going to happen with print revenue.

In fact, when I joined The Atlantic in 2012, one scenario I was told to consider was that we’d be out of print in three to five years. So, when The Atlantic initially embraced digital transformation, it was an imperative. There was no other option. What makes that different from a lot of other media companies is that those companies had brands that were still bringing in meaningful print advertising and consumer revenue. And although they could see that things were not growing and would begin to decline, there was no scenario where they could just walk away from that revenue. When you’re trying to maintain one thing and embrace something else, that makes the switch a lot harder.

It sounds like a strength that you were struggling so much, because there were no eggs in the old basket and you needed to move quickly.

Much of this transformation began well before my arrival at The Atlantic, but my impression is that there was a willingness across the board. Everyone knew they had to embrace a new plan in order to survive, and that shared understanding fundamentally shifted the culture and made that transformation a lot easier.

It sounds like it was really the fact that survival was palpable to everybody, which made them more willing to make that change.

I think that’s accurate. I mean, again, I’m speaking to this as sort of the lore that is part of the company. What I can say is when I joined five and a half years ago, we were modeling scenarios where print advertising would be gone in three years, and that was very much part of our operating focus.

Today, we’re in a different position and absolutely don’t believe the print format is going to die. In fact, we reached our highest circulation in more than a decade in 2017. The market is changing, but we are most definitely not forecasting print’s demise.

What changed? Were you just wrong, or did something else shift that made you breathe new life back into the print side of things?

I think we were wrong. But I also believe that our successful embrace of digital helped introduce The Atlantic to new audiences.

The brand has grown tremendously in the last five years, which has helped shift the conversation we’re having. The industry has also changed. We’re at a point now where consumers are much more accustomed to paying for content. That has been an ongoing shift across the board in the last 10 years.

You talked about transformation: shift in mindset, shift in culture. Can you tell me a little more about that? What were the big changes that were necessary, and what were the hardest ones to make? I realize you’re speaking a little bit from lore, but as companies are dealing with this, what are the challenges they’re going to face?

One of the ongoing challenges is understanding the difference between fast content and slow content, for lack of a better frame. When you’re publishing online and you’re publishing at a high frequency, there are a lot of different beats.

To go from a magazine process that involved researching stories and fact-checking and editing things for months, one of the biggest changes is embracing new formats for storytelling and being open to experimenting with new approaches.

One of the shifts happened in 2008, when The Atlantic started hiring bloggers. The editorial team was willing to expand its views and bring them into the fold and experiment with those new forms. Today it’s a lot more nuanced, but there are constantly new formats and new experiences to engage with.

Thinking about how you define your brand is core to the debate, and it applies to everything, whether it be video or podcast or other formats. You’re always looking at it and saying, OK, how do we maximize this format or this distribution mechanism while staying true to our brand?

We talk about the difference between, say, print and digital. But digital is not homogeneous. So how do you keep innovating? How do you stay on top of things to make sure you’re not just relevant today but you’re going to be relevant tomorrow?

It’s a good question. Part of that is hiring smart people, whether it be my product team, which is a bunch of people who live and breathe digital every day, or whether it’s just the mindset of people you hire. You’re looking for people who are energized and excited by solving complex problems. That’s the core, whether it be on the product side or a developer or even a marketer; that’s ultimately what I’m looking for in people.

Culturally, it’s about having a culture that is comfortable with experimentation and encourages an entrepreneurial nature. We don’t have a ton of approval layers or overhead required to try new things. In fact, testing and thinking about things in new ways are encouraged.

We still have to be good at picking our priorities, and that’s an ongoing challenge. Sometimes we get it right, and sometimes we don’t.

This is one thing that legacy companies are really struggling with. In theory, they like experimentation, but they’re so afraid of failure that they have trouble developing a culture where failure is understood as a natural part of experimentation. How is that handled in the company?

I’d say it’s handled honestly and directly. We’ve done many things that have failed over the years.

For me, a lot of experimentation is about more transparency, being clear about what your goals are and where you’re trying to go. This means when you’re not getting there, it’s an easier conversation because people are more aware of what’s happening or what’s not happening. By the time things have failed, the communication of that failure is not a big problem.

But the organization talks about it openly, right? You talk about, “This didn’t work because of XYZ.”

Yes. There are always exceptions, and I’m sure there are ways that message is softened for different things. But there’s a very good understanding here. This is partially also because of our size. You have to be focused on where you make your bets. We have to choose carefully, because we’ve got limited resources.

And for me, that leads to an imperative that you work fast and you move on quickly, because there’s always somewhere else to put those resources.

You had said one key is making sure you hire smart, passionate people. Does the organization do anything to make sure they stay smart and passionate? Once they’re working at The Atlantic, how do people stay fresh?

It’s a good question. I don’t know that I could speak for all the different teams — within an organization you have different cultures such as within the editorial team, within the sales team, etc.

From the product perspective, we are operating in a dynamic environment, so the technology we use today, the things we experiment with today, may be totally different tomorrow. It’s really just part of the job, part of the day-to-day of trying to respond to challenges and making sure you’re rising to the call.

A few years ago, ad viewability was a big thing. It fundamentally requires relooking and changing how you do all your ad surfing and how you think about your pages. Last year, ad blockers were a big theme, so everybody’s trying to scramble and think about what they can do. For my team, it’s about bringing these problems to them and asking them to solve and shape how we continue to move forward.

I try to give them as much transparency as possible because I find that people are more engaged when they know why they’re doing it. If we’re pivoting or shifting focus, that can be really disruptive. But it can also be energizing if the team understands and buys into why you’re making those decisions.

I always say to my team: None of us are experts on anything. I’ve been doing this for years, but everything changes.

Every time we start something new, it’s the questions we ask as we’re approaching it that matter: How do we do this faster? What’s the right way now to do things? It’s important to constantly force everybody to question what they know and to bring that to the table.

It occurred to me as you were talking that The Atlantic has really good data on your products — the “how it happens and how things are happening in real time” — and responds accordingly. Is that access to and availability of data important in helping you make these decisions, and be bold and responsive and entrepreneurial, etc., because you have that feedback?

I do think access to data is important. But no matter how much data you have, you never feel like you have enough. Data tends to lead you down paths where you have more questions, and you quickly get to a point where you can’t answer them.

We had not historically invested in our own data stores and data teams, and we’re only now starting to build out those capabilities. In one sense that’s an advantage, because we don’t have a lot of legacy stuff in place, so we have the ability to start from scratch and use the best technology of today to build it.

I often compare what we do to The New York Times. I love The New York Times, but it has a huge, huge team. I had a conversation with the person heading up the video team earlier this year, and there’s a data team just for the video team. And the data team for the video team is bigger than our whole video team.

So, when we’re designing products, we might look at The Times, and say, “The Times is doing this, and we are fairly confident that means they’ve done the research.” We use the data that’s available to us — when we design new products or work on new things, we’re always looking to see what our peers are doing and trying to pull as much information as we can. We’re obviously looking at whatever we can pull from our core data sources as well. But at some point you have to take what you have and make your bet. We make the most of what we have. I’m fairly certain that even after this year, when we are investing more into data, I’m going to get to the end of it and say, “Oh, I need 10 times that.”

It sounds like enough data but not too much. And I love the story about watching what other people with data are doing and assuming they are doing something smart.

They’re all data points, right? We’re constantly trying to triangulate where we are and what we can do, and also just using our intuition. But we have some people in some groups that are more into data than others, so going forward, there’s a continual aim to try to bring insight — less data, more insight — to more people.

We’re finding that more digitally advanced companies emphasize collaboration not only across boundaries within the company but also outside the company. Is collaboration an important part of what you do, and if so, can you talk about how you collaborate internally, as well as if you collaborate externally with partners?

Yes, it absolutely is. At the core, that goes back to when you’re trying to solve problems that nobody has the answers to, it generally works better when you have a handful of smart people working on those questions together.

Internally, as we develop new products, we’ve tried a few new things in the last year, and we continue to revise our own processes. We used to be more of a waterfall shop, where requirements would get built and designed, and then they’d end up in developers’ hands. We’ve been actively moving toward a more fully agile approach that, at its core, is just having the developer, designers, and stakeholders sitting and working through the problems together, getting to an end result, and then iterating on that.

And it’s been extremely important to not design things that can’t be built, to make sure you’re asking and dealing with the right questions, and part of it is the complexity of what you’re building for. For instance, when you’re building websites for multiple browsers and multiple bandwidths, you need to take into account many different elements. It gets increasingly difficult to design a comp when you have eight different breakpoints on the page and you need to synthesize all these ways people might consume or experience that.

The other internal collaboration tool we tried last year was the Google Ventures sprint process. One of our main challenges was that we have superlimited resources across the board, so trying to solve problems to big, new, open-ended questions was something we really struggled with.

The Google Ventures sprint process basically outlines a five-day, Monday-through-Friday process where you have a team that’s dedicated to answering a question or solving a problem, and it ends with a prototype that’s getting tested. So, ideally, at the end of the sprint, you’ve made significant progress on a product.

Are there any barriers that get in the way of collaboration, either internally or externally?

Collaboration generally takes more time, and it’s harder to get people involved. It takes more coordination. In some ways it is a skill set to be able to actively solicit and pull in more people and hear their thoughts. You still have to make decisions, but making decisions with a group is sometimes harder than doing it on your own. I’d say time, resources — and sometimes politics — can be barriers.

Who actually led the original transformation you talked about 10 years ago? Because one thing we’re finding in our research is it matters who helped then.

It definitely came from the top. For The Atlantic, that meant our owner, the company president, and the editor in chief.

And then the senior management team was all aligned and/or hired to explicitly push those things. In fact, my role when I joined five years ago was really brought in to elevate product and these other functions that had never existed in an organization that historically was mostly editorial and sales.

Transformation definitely comes from the top.

We talk a lot about culture as a part of transformation, and the media industry in general has had to go through a culture change. Journalism is not what it was 10 years ago, right? And you actually have a fair number of journalists at a time when lots of media companies are losing them. Can you talk a little bit about this cultural shift in journalism, what it means to be a journalist today, and how that might be culturally different than it was 15 years ago?

I’ll caveat all of this by saying that I’m not a journalist, so this is my own perspective. My impression is that today the feedback loops are instantaneous.

Today there’s instant feedback. Our journalists constantly worry about distribution, how they structure their story, and how they write their headline. Everything that goes into the production and the publishing of a story ultimately has some impact on its ultimate distribution and reach, so the roles that a journalist has to play are significantly increased. It’s not just finding good stories and being rigorous in your reporting.

With the different skill set requirements and the different things you’re asking them to do, are you asking them to make this shift, or are you hiring people who have that disposition?

It’s probably a little bit of both.

To a large extent, this is how a product team can help the journalist. The tedious parts of workflow development and making sure journalists know that our evolving best practices are ultimately things that, if done well, can be built into a product, so they don’t have to remember their checklist of things they have to do to make sure a story is ready to publish — the tools help them do that.

I wouldn’t say we are doing that 100% today, but the ambition is to have the journalists just think about their story and not have to explicitly think, OK, how do I publish for maximum distribution? That should be the easy part, where the tools just guide them through it. That’s the ambition.

In that way, I think it’s partially about what tools you give them and how you train them — but not in an explicit way, when you sit in the training room and work through and learn about this module. Instead, it should be an easy and intuitive process. So, I think it’s both.