How to save a magazine, with The Atlantic’s Nicholas Thompson

0:00As a founder, you're constantly wearing multiple hats, usually at the same time. but the most successful leaders have learned to hand a few of those hats off to the right people. That's where Upwork shines. It's a one-stop shop to find trusted freelance talent across IT, web dev, design, and admin support. And it's built to be simple. Register for free, connect with freelancers who match your style and pace, and hire them with industry low fees. Upwork makes the whole process easier, more affordable, and way less stressful. If your business is ready for that next level, Upwork has the talent to help you get there.

0:39Visit Upwork.com right now and post your job for free. That's U-P-W-O-R-K dot com. Upwork dot com. This is fun, by the way. Thanks for having me on. Thrilled to do it. Did you run up here, by the way? I was going to run. It's like a little, like then you stink during the interview and it's like not great. Yeah. I should have done it. I just. Should have done it. You want to re-rack? Yeah, yeah. Can we reschedule this for 1130? I'm going to go home and run. Nick Thompson didn't run to the studio for our interview, but he runs almost everywhere else. He runs marathons, ultra marathons, and most days he runs to work at his job as CEO of The Atlantic.

1:20When you saw subscriptions actually drop, was there a moment where you had a crisis of conviction about this? Yeah, totally. I was terrified.

1:58This is Masters of Scale.

2:29Nick, welcome to Masters of Scale. I've been looking forward to this conversation for a while. You are now the CEO of The Atlantic. Why take on the challenge of turning around this iconic American institution? So a couple of things are important that somebody who just looks at my resume or just sort of knows me from my Twitter handle wouldn't know. One is that when I graduated college when I was young, I didn't intend to be a journalist. I did not intend to be a journalist. I just didn't know what I was going to do. But then I eventually found my foot in and I got some great jobs. And so I work at Wired, I work at the New Yorker, I go back to Wired to run it.

2:59And then the opportunity comes to be the CEO of the Atlantic, which is a big shift. And the way I thought about it was, I'm a pretty good writer, I'm a pretty good editor, pretty good manager. But there's this concept of comparative advantage where what matters in life is not what you are best at, it's what you are best at relative to your peers. And I sort of thought that my comparative advantage was figuring out business models for serious journalism organizations, right? And if that's my professional compared advantage, that's the thing that I'm best at relative to my peers. I may not be, certainly not the best in the world at it, but compared to everybody else in journalism, I'm like in a higher tier, like at the New Yorker, a ton of people who can write better than me, right?

3:40So many of them, right? Like, and how many of them are better at figuring out business models for serious journalism? Well, I don't know, fewer. And so the opportunity he came up. I was like, well, I should really test and see how this works. So that's why I made the shift five years ago. I met with David Bradley about this job when it was open. I think I was maybe the 712th person he met with along the way. He made a much better choice and I wasn't available for the role at the time. But I walked away really struck by two things. One, what an opportunity. I mean, this is one of the most important institutions in American media and American culture and the chance to take it in decline and build it back up is off the charts.

4:20And at the same time, holy shit, this is a big hill to climb. This is not easy. You've got a legacy institution in a sector that's in decline and all of it. When the opportunity presented itself, when you were sitting there deciding, am I going to do this? I'm sure you're excited about it for so many reasons, but how daunting was it? It's funny to go back to that. So this was 2020. I took the job. It was announced in December of 2020, right? In the middle of COVID. And there were two really salient facts. One, I knew the economics of the Atlantic were terrible, right? And they'd been reported that they were losing 20,$30 million a year.

4:56They'd laid off 20 % of their staff. They shut down the whole live team. They'd fired the whole video team, right? David Bradley had written this effusive memo. It was very complicated. This is the time of George Floyd. It's the time of COVID. So I knew it was in deep financial trouble, right? It's a real cultural touchstone for me. But the other salient fact was that the journalism was amazing, right? And I knew that because I'd been competing with it, right? And I'd been running the newsroom at Wired while we were competing with the Atlantic on stories on COVID. And so I thought, you know, if the journalism wasn't amazing, not a fixable situation, but journalism is amazing.

5:33I know how to run a playbook where the New Yorker, we built a digital subscription model, we built a thriving website, advertising business, all kinds of naturally business that like really worked. We did the same thing at Wire. Like I know how to work on building plausible economic models. So I thought, I don't know, I bet you we can do this. So I said, yes. Plus David Bradley. So I mean, charming, a remarkable human being and very hard guy to say no to. I'm sure, where every time you take a new job, it's a covered dish, right? You think you know what you're getting and you open up. I didn't know there were raisins and almonds in this thing.

6:07Right. So when you got there - It's actually when you take a bite that you realize the raisin and almonds. Right, right. And cumin on top of it. When you got there, what did you see? What did you learn where you were either like, holy shit, I can't believe we have to deal with this too. Or we were like, oh my gosh, I didn't realize how big this opportunity here was. Well, it's interesting because when I came in, I saw a couple opportunities that didn't matter, right? I was like, wow, they haven't digitized their archives, right? That's crazy. Like go back to 1857, like the archives aren't readily accessible.

6:40Like, let's just do that. It'll cost$2 million to do it. And then we'll generate like tons of traffic, Google SEO, right? Like we'll be able to run like an archival playbook like we did at the New Yorker. And we digitized the archives. It's a great project. So glad we did it. Does it generate millions of dollars? No, it does not. Right? Right. The second thing I was like, well, you know, Wired, we build this incredible affiliate business model, right? We review gadgets and people buy, we get an Amazon commission. I was like, we'll do that through books at the Atlantic. We'll figure out, we build that.

7:05It generates like $7 ,000 a year, right? It doesn't matter. What did matter was, okay, wait, we have this paywall and we have this great journalism. Let's figure out the math, right? And then we'll just do hundreds of tests. I think we ran 230 tests last year on different variables involving essentially the paywall. This is actually comes from covering the tech industry, right? Where you watch Amazon and like, Hey, they have this brilliant idea. And part of the way they execute is they just, every variable gets tested a thousand times. And if something is 1 % better, they change it. And so we start this culture, data science, iterative testing, and turns out that like changing the math on the way the paywall works and the way the paid marketing works and the offers work, like suddenly our subscription growth goes, you know, from, you start to see the hockey stick.

7:52We see a hockey stick. And that's what takes us from losing a lot of money to making money. The Atlantic is easily in my top tier, most valuable subscription. I mean, I'm in that cohort where if I woke up tomorrow and you're like, yeah, we're doubling it. I'd be like, well, I'm in. Okay. Great. Test me in that pool, please. Hold on a note to self. I'm writing this down. But was there an inflection point on the subscription piece where you said, oh, this is going to work. We can really do this. When you're the Atlantic, you don't publish that many stories, but you publish a lot of hits and you have people out on MSNBC and all over the radio talking about it.

8:26It turns out that actually you can have a very hard paywall. Now there's a trade-off there though. You have a hard paywall. A, it hurts your advertising business. And B, maybe it hurts your brand. And C, maybe it's not great for democracy, right? You're in this business because you want people to get your information. And so you suddenly have to square all those things, right? And we had this meeting bringing together all the stakeholders. And I was kind of like, you know, this is what the numbers show we should do. And we didn't go to zero because all these trade-offs, but we went much tighter and we built a very sophisticated paywall that the goal when you have a paywall is if someone's not going to subscribe, you want them to read the article, right?

9:03And if someone is going to subscribe, you want them to pay to read the article. And problem is you can't know, you can't like interview every person about their subscription propensity. And so you have to guess and you guess based on what the story is, based on what the referral source is, based on all the information that you have about them. And if you can set your paywall rules so that you guess correctly more often and incorrectly less often, you're in really good shape. Maybe you set different rules or people are coming from Google search and people are coming in from Drudge. What about if they're coming in from Twitter?

9:28If they're coming in from at the Atlantic, what if they're coming in from somebody else? And so you can kind of like make a bunch of rules based on that. And in year two, when you saw subscriptions actually drop, was there a moment where you had a crisis of conviction about this? Yeah, totally. I was terrified. So what happened? I was just worried. I was like, well, maybe I'm not good at this, right? Like I came in with all this confidence. We could run the New Yorker playbook, the wired playbook. But what I knew is like, I knew the team I was building was amazing. The team was getting better and better.

9:54Right. And like morale wasn't dropping. Like it wasn't like people were starting to fight it. When an organization goes up, it's kind of like, everybody's happy, right? Like you, you have to remanage it. You have to figure out like reporting lines and flows of information. When an organization is going down, you can have a great team and like everybody's mad. Right. And we weren't, we didn't have that. And I also had said, it was like my first all hands. I said to the staff, I said, we want to get to a million subscribers, really hard goal. And we want to get to profitability, really hard goal.

10:21Right. And when we get there, everybody gets a 10 ,000 or 10 % bonus. So I started in 2021 and 2022 was like ominous. And then 2023, it starts to pick up. I think it's the summer of 2023. What happened that summer? The test started working and like maybe it just took enough time to gestate. Maybe it was some changes in personnel. But when I was there a year and a half, that's when you could start to see when you look at kind of the leading indicators. Oh, wait, actually, like the numbers everybody else sees are bad, but the numbers I see are good. And then by the winter of 2024, it was like, oh, wow, now everybody can see the numbers are good.

10:57And then by the end of the year, we were like profitable and had reached a million subscribers. So everybody gets paid. And everybody was incredibly excited about that. But there's like there's a period, I guess it's about 18 months from when I start to when the indicators look good. And it was probably about a year, that sort of the year to 18 month period where I was like, God, huh? Maybe I should have stayed a journalist. Yeah. Well, and incredible where you guys are now. The goal was a million subscribers. When you came in, you were at? Remember 800, but we were not growing, right? Okay. Yeah.

11:21Right. And you were losing 20 to$30 million. And so to be where you are now. And now we're at 1.46, right? And we're making, I can't say exactly how much we're making, but we're making good money. I mean, we take the money we make and we hire new people. So we don't, The goal of the Atlantic is really interesting. We're owned by Lorene Powell Jobs, who obviously doesn't need revenue from the Atlantic as her principal source of income. But she doesn't like being taken advantage of, right? And she liked the things she owns to make money and be sufficient and be serious and not to treat her like she's an ATM.

11:51am. And so the way I think of it, not everybody in the organization agrees, is you have to be profitable and you'll do everything to be profitable. And if you're not going to be profitable, then you make hard choices and all that. If you are going to be profitable and you're going to make$2 million, well, then you can take 1.9 and put it back into the most viable things in the organization, ideally to grow and have more influence and hire more great journalists. You're going to make 4 million, maybe put 3.9 back in. And so that's the way we're running it. One of the decisions you made is to do a big licensing deal with OpenAI.

12:19There are a lot of media companies that are doing these deals. There are some that are hesitant to do them. Why did you do that deal? This was interesting because I had the most hate and the most anger internally, right? We ran a story. Like I signed this deal and then our journalism side, which is separate from the business side, runs a story called the deal with the devil, right? And we run all these stories that are kind of attacking our deal and there's a lot of commentary and I do this all hands, which is quite aggressive. So I never doubted that it was the right thing, 100 % confident and remain so.

12:48So the reason we did the deal is that the AI companies came and they scraped all our data, clearly in violation of our terms of service, probably in violation of copyright law. Of course, you're going to sort that out, certainly in violation of kind of every sort of ethical standard. And then they built competing products against us, right? So they did this, right? And we weren't savvy enough to catch them and stop them in the beginning. So then the question is, all right, what should you do about that? The thing that's not useful is just kind of like stamp your feet and be frustrated. And unfortunately for many of my peers, the strategy has been stamp your feet and be frustrated.

13:18And so I basically said, look, we would like to be compensated. So we're either going to sue and pursue this legally, or we're going to make deals. And with OpenAI, we said, hey, here we are. Let's talk. And they said, great, let's make a deal. And you'll become a partner as we build this new part of the internet. And to me, that's a much better deal than not doing that. Now, I can see the New York Times argument. Let's sue. Go get him, right? And we are party to another lawsuit and who knows, we've reserved all rights. There has to be a fair exchange of value, right? And if there isn't a fair exchange of value, well, then we'll use whatever methods of pressure we can, whether it's lawsuits or whatever.

13:52And so that's how I have approached it. So if you look at the last 20-ish years, there's sort of two directions that it seems like this has gone. Three. One is the NBA where it's like, we're leaning into YouTube. We're going to figure out how we build a global audience here. And we're going to work with the platform to make this happen. Two is Viacom just sues YouTube. Like, we're not going to play ball. We're going to sue you. And then three is where the traditional media companies just became arms dealers to Netflix, licensing content to Netflix, where all of a sudden, eventually Netflix is displacing and replacing them and now buying them.

14:26Yeah. How have you informed this decision about doing the OpenAI licensing deal with this data from the last two decades or so? What we've seen in the relationship between the media companies and the tech industry and the tech platforms. And I view these guys as like the new tech platforms, right? And they will probably have a similar role to what Facebook and Google had. None of these partners can be trusted, right? Not because they're bad people, they're good people, they're friends of mine, right? But they're capitalists and they're self-interested and they would rather they make money than you make money.

14:53And if you can both make money, they're happy, right? But they don't actually really care that much about you. So you shouldn't rely on their goodwill, but you should rely on their economic incentives. So as a result of that, we've got to build direct relationships as much as we can. We expanded our print magazine, right? in part to have direct relationships with people. We work really hard on our app to have direct relationships with people. We build our newsletter product, right? Those are all about direct relationships. So there's no algorithm in the middle, but with the algorithmic players, the key is to develop an ecosystem that is mutually beneficial.

15:20And in the old internet, we had a mutually beneficial ecosystem, right? Google scrapes the internet, scrapes us, but then it has links, right? There are all kinds of different ways where you can have a mutually beneficial relationship. With the AI companies, we don't have that yet. and them paying one-time licensing fees, that's not it. What has to happen is the AI companies, whether it's Gemini, whether it's ChatGPT, whether it's perplexity, whether it's Claude, someone has to figure out a model where there's a real exchange of value. And it could be modeled on what ProRata has done. So what ProRata has done is they've built a search engine.

15:53You search for something, they identify, okay, the answer comes 6 % from the Atlantic and 4 % from the Guardian. So therefore we'll give 3 % of the revenue to the Atlantic and 2 % of the Guardian, right? There's like a direct payment relationship. That's a really good model. You can imagine a different model where Gemini says, you know what, we're trying to make the Atlantic whole because they've given us all this data and they've given us all this stuff. We're gonna run our algorithms and we're gonna identify who's a high propensity subscriber. And if there is one, we're gonna put like a call out to the Atlantic.

16:21We're gonna put a link to the Atlantic subscription page. That's an amazing exchange of value, right? Perplexity could do that too. Claude could do that as well, right? So you can imagine all kinds of different models. What about subscription integration, right? What if you're an Atlantic subscriber and you authenticate it with ChatGPT? And then when you search in ChatGPT, you either get the Atlantic to inform your answer or you don't. And what if we only do a deal with ChatGPT or what if we only do a deal with Gemini, right? And then there's a way that we're making money and they're building a better product for their consumers.

16:48Now, what hasn't happened is any of those things, right? And what I hope happens in the next year is that one of those big companies says, oh, you know, right? Like, we're going to figure this out so that we can build a model for the future internet that works as well as the old model did. How do you get them to see that it is in their long-term interest to work with you rather than just cannibalize you? Yeah. Okay. Amazing question. So part of the problem, part of the reason we have less leverage than we did in the last round is that they all think that their ultimate in value is going to be infinity, right?

17:18So now the question is, how do you get leverage, right? And how do you get them to pay attention? And there are a couple of ways you can do that. So one, and I wish the media industry had done this sooner, is block them from scraping you, right? And so initially we were like, gosh, they're scraping us? And then we're like, okay, well, we'll put a robots.txt file up and we'll say, hey, please don't scrape. And that's like a no trespassing sign, right? You know what? You know how much they care? Zero, right? So then it was like, okay, then Cloudflare comes in and they're actually good at blocking bots.

17:43So now we have a nice dashboard and we can say, no, no, no, no, no, no, no. And so we block them. So that gives us a little bit of leverage. So now what are the other points of leverage? Well, you could move up the stack and have people somewhere else in the stack of information supply, right? You can imagine an ISP, right? That's one point of leverage. Second point of leverage would be legislation. It's not going to happen in the United States, but it might happen in Europe. It could happen in Australia, right? And then once there is legislation, then things change. Another point of leverage would be lawsuits.

18:12Anthropic just had to pay out billions of dollars to authors. What if we start winning, right? And like the case law has been a little mixed. so far, but you can imagine we start winning lawsuits. Well, then that changes it. And then another lever is public pressure. My view is that once one of these AI companies comes out and actually does the right thing for the publishing industry and the right thing for content creators and the right thing for Hollywood and the right thing for authors, once somebody does that, like maybe they'll get the reputation as like the ethical AI shop and then things will be much better, right?

18:42And that'll actually be good for their brand. And you can imagine a discussion like, huh, wouldn't it be good to have the reputation as the ethical AI search engine as opposed to the unethical one. Still ahead, how Nick Thompson learned to run faster than ever in his 40s and what it can teach the rest of us, even if you are an extremely reluctant runner like I am.

19:09This is Mike Nicholas, Capital One business customer and co-founder of Ansett Uncles, a plant-based restaurant and community space in Brooklyn, New York. and he's telling us how they started a product line. We already had a space in the community. The food was an extension of our lifestyle and our values. So we know we wanted to create something that was an offshoot to that. Our pepper sauce. That's my grandmother's recipe. That's like a liquid gold, right? If someone wants to approach us and ask us, what do we do? We provide flavor. Growing a product line is no small investment, but Mike and his wife and co-founder Nicole were able to manage with the help of their Capital One Business card.

19:47Working with Capital One Business, we're able to leverage our limits and utilize those points, making sure we can continue the scale at the speed that we need it to go. To learn more, go to capitalone.com slash business cards. Welcome back to Masters of Scale. You can find this conversation and more on our YouTube channel. And be sure to check out the link in our show notes to subscribe to our newsletter. If you spend any time on LinkedIn, you've probably seen Nick Thompson's videos there. Even after moving into his new job at The Atlantic, he decided to keep publishing his popular daily series, which has evolved from the most interesting thing in tech to the most interesting thing in AI.

20:33I think I stopped for like two weeks when I took the job. And then I remember the first one I posted, one of the first comments was like, dude, you're CEO of The Atlantic. Why are you still doing this? And then I remember thinking, well, you know what? it's actually a really good learning mechanism. And so it's a way to like force myself. You become CEO, you can spend all your time, you know, doing things that are helpful to your job, but you're not learning about the world in a way. That's what's best about being a journalist. So it's like a learning and forcing mechanism. And then I actually realized it's good for a business because you know, who's watching these videos, CMOs, and they're really important people to reach.

21:05And so you go into a meeting with the CMO and you're like, Hey, you know, I think you should advertise in the Atlantic. Tell me more about it. By the way, I liked your videos. Right. And it turns out to be like a really good marketing channel for The Atlantic. Now, I still do it as a journalist. I'm not talking about our advertisers. I'm not talking about our partners. And sometimes I'm upsetting them. I'm just saying what's on my mind. And as soon as I stop doing that, then the things are worthless, right? As soon as it's canned, rehearsed, as soon as it goes through a comms channel, then it's dead.