Mariah Craddick is the Executive Director of Product at The Atlantic. She began her career as a journalist and a marketer before transitioning to audience engagement and development while at Crain Communications. From there, Mariah eventually became a product manager owning the audience product roadmap for the organization. Before her current role at The Atlantic, she worked as a product manager for audience at McClatchy, a media company, and as a senior product manager for storytelling at The Wall Street Journal.
In our conversation, Mariah shares how product management differs in the media industry and how The Atlantic is working to break silos, collaborate closely across departments, and innovate with new technologies. She also talks about how she foresees the media space transforming paid experiences, overall content and reporting.
I’ve only worked in product within the news industry, so most of my understanding about product at other places comes from having conversations with different leaders in those fields. But, for product in news, the primary function of a product team is to develop all of the experiences that allow readers or listeners to access, engage with, and support our journalism. Many product teams in news companies are also responsible for internal-facing tools, such as the content management systems reporters use to publish content.
With that said, a few things are unique to the news industry when it comes to product management. For starters, the function itself didn’t exist until about a decade ago, so it’s relatively new compared to the tech industry. A lot of news organizations are still making the transition to being digital first and more audience-centric.
Also, product management within news organizations can vary from place to place. When I first became a product manager in 2018, the product team I joined was not very mature. We were primarily taking requests from around the org and putting them into tickets. There wasn’t a large understanding that product management is not just a way to get engineering work done, but a whole strategic element. At The Atlantic, I’m lucky to have a much more mature product team that helps shape strategic vision and direction for our owned and operated properties in addition to execution.
One big challenge is that it’s still predominantly a siloed function. A lot of news organizations struggle with the product team acting as a sidecar to the rest of the news operation, and that limits a team’s ability to be really effective. We have to work extremely hard to foster collaboration and make sure that communication exists across the organization.
Further, newsrooms have historically made a lot of product decisions. Especially in the early days of the internet, an editor decided what the homepage might look like, which elements should exist on the site, etc. They were the ones coming up with the design for how a story is treated. This shift toward thinking like product people or working with product teams to make these decisions instead has been tough for some organizations. There’s a cultural shift to let product people in to help make some of these decisions.
Lastly, it’s tough to attract product managers and engineers to the news industry — we’re not known to be innovative or tech-focused. That’s largely a PR problem and not necessarily the reality, though — many news organizations are solving sticky, interesting user problems.
Before I worked in product, I was a reporter and also served in different marketing and audience development roles. The question I always had as I moved around was, “Why don’t we talk to each other more? We have information that the other team can use. Why are we not collaborating all the time?” At The Atlantic, we’ve had good success in establishing open lines of communication, though it’s always a work in progress and something we’re always trying to improve.
One thing we do emphasize is over-communication. This means being extremely transparent about everything that we’re doing, from ongoing projects and status updates, to what we learned from an experiment. We have many different touchpoints to facilitate that, such as bi-weekly shareouts and monthly roadmap status updates.
When it comes to ideas or initial concepts, we also try to bring that to stakeholders early in the process so that nothing surprises them later. In a lot of cases, we need their support to either market the initiative or add editorial perspective. We aim to build that openness and transparency even into the development process through various recurring touchpoints that sometimes include stakeholders as well. Because things often change in the middle of development, we might have to reduce scope or make a new decision. As such, it’s crucial to have their involvement directly in the process.
A lot of this is really about building a culture of collaboration, which makes us more effective as an organization. The best products and features are made from collaboration where silos are broken apart. The end goal is to be integrated and work together on solving all these various problems.
We try to lead decision-making with thorough insights and stay very focused on who we’re trying to serve. We have fantastic data science, audience research, and customer care teams that all help us understand how our audience uses our products, the problems they face, and what they wish we would offer. That’s a critical piece for us.
Three or four years ago, our teams were structured and oriented around platforms. We had product managers who owned specific platforms and focused on evolving them over time. Now, our product managers are oriented on outcomes and objectives. They focus on features, but more importantly, what impact did those have on the audience? We lean heavily on data because of that, which has been a positive shift for us.
We’re always looking at what our peers are doing. But The Atlantic is slightly different from our peers in that we’re not a breaking news publication. We come with a lot more voice, perspective, and ideas from various writers. We have a product strategy and experience that tries to mimic and support that. We’re not necessarily trying to be The New York Times or The Washington Post, which are leaders in breaking news and publish a massive amount of news stories on a daily basis. Instead, we try to be a resource where, if someone reads something in The New York Times, they can come to us to make sense of it, get more context, or hear various perspectives.
In general, our site and app experiences have a lot less “stuff” on them. It’s a much more dialed-back experience. We’re very thoughtful about what we put in front of users, especially because we know they’re coming for journalism. Of course, we have to promote subscriptions and features, but we do so in a way that doesn’t disrupt or distract from the journalism.
One example is our investment in audio experiences. Two years ago, we hardly had any audio experiences. We had a few podcasts but that was about it. We love long-form journalism — we pride ourselves on the written word — but we would hear from our readers, “I wish I could listen to this story instead.” And we also were seeing the rise of the podcast format and audio across our industry, as well as outside of news. So we knew it was something we should invest in, knowing that it could be a way for us to not only amplify our journalism, but also be a way to more deeply engage with it.
We started to ask ourselves, “Could we actually offer a broader audio library? Could we get more of our content in the audio form?” In comes ChatGPT and generative AI, and we, like everyone else, were trying to understand what we could do with that technology. We ended up hosting a hackathon and one of the ideas that came from it was using AI to help us narrate content in a way that doesn’t sound robotic. We tested it out in the hackathon, and it was eerily good. It wasn’t perfect, but it was pretty good.
From there, we developed it further. We eventually got it to a place where we thought it was really high quality, and could produce an audio version of an article shortly after publishing. We could also still have a human in the loop to edit it if the tone needed to be adjusted.
At the same time, we started to invest even more in our podcast strategy and expanded that suite with several new shows. Our audience really appreciates the “behind-the-scenes” vibe of podcasts — being able to listen to our writers talk about topics they might have read about.
So, with the investment in narrated audio articles and podcasts, we now have a suite of audio content so that our users can listen to The Atlantic. And my team has done a lot of work to bring that experience to the forefront, particularly on our mobile app, knowing that listening is often a thing people do while on the go. We’ve extended the audio experience on the app, making it easy to continuously listen while multi-tasking, building a queue of things to listen to next, etc.
And we’ve seen signals that those who listen in the app tend to come back to the app more frequently. This is a positive sign for us, because we know that usage frequency often correlates with long-term retention, which is our ultimate goal.
This is all still very new to our industry, and we’re all adjusting, experimenting, and putting the right guardrails in place. Just recently, The New York Times unveiled how they officially asked the newsroom to use internal AI tools they’ve built to help them with certain elements of reporting.
I’ve seen great use cases from across the globe of newsrooms using AI tools in this way. Generative AI can help reporters more easily go through large bodies of data and pull out potential leads for stories. It can help reporters translate handwritten text from an image. It can help transcribe interviews, which is a godsend for what used to be a pretty painful process. And there are lots of other ways AI can help make the entire reporting process much more efficient.
The Atlantic is in a place where we’re trying to figure it out for ourselves, too. A lot of the areas we’ve been exploring focus on things that generative AI is particularly good at. Our writers are actually the best at writing and putting forward novel ideas and opinions, so we don’t want to use AI in content generation at all because it’s an area where we really excel. We’re also not interested in the idea of summarization, which is something other news publications are experimenting with.
With that said, we do have an experimental site called Atlantic Labs, which launched a couple of months ago. It’s meant to be a place where the product and technology team can experiment with AI and other emerging technologies safely and separately from our core owned and operated products.
We have several experiments up on Labs right now. We have one experiment called “Companion,” which is an AI chatbot that can assist with finding articles on any given topic within our vast archive. As we’re learning, AI and large language models are particularly good at sifting through and organizing large swaths of data really quickly. And we have a lot of that since we’ve been publishing since 1857.
So many of the experiments up on Labs have been making use of this, and we’ve been exploring different ways to find and present relevant content on the wide variety of topics we’ve written about over the years. But we’re not just experimenting with generative AI, we also use the Labs site as R&D and are beginning to test out new games that, if people show interest, could potentially come into our core products.
I feel like we’re at an inflection point for our industry. With generative AI changing how people relate and find content, we’re at a time when publishers are going to have to get really serious about their business model and value proposition in this new era. How do you create products and experiences that people value and will support in some way? Otherwise, many likely won’t continue to exist.
Many of us are going to need to diversify what we offer further so that we’re less reliant on only advertising and reader or subscription revenue.
The New York Times has done this well with its Games and Cooking products. Even though we all can’t do Cooking, we can all get rigorous and diligent about figuring out what our audiences want, need, and desire to help develop what we bring to market. How do we figure out our unique value proposition and diversify our portfolio based on that unique position?
That’s my main prediction — that we’ll all need to figure out the unique value that we bring and invest in our owned and operated properties so that people have a reason to come to us directly. We think that we offer what people want, but we have to continue giving it to them in a way that feels valuable and enjoyable, and makes them want to pay for it or support it in some way.
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?
If you get AI personalization right, you’ll see higher customer retention, increased trust, and long-term revenue growth.
Kevin Erdner, Senior Director of Product Management at Labcorp, shares the lessons improv comedy taught him about product management.
Isaac Szymanczyk talks about how storytelling and effective communication are essential in getting people to rally behind change.
Rick Agajanian talks about the real impacts of working on mission-critical software for the professional service industry.
Learn more →
