clock menu more-arrow no yes mobile

Filed under:

Full transcript: Scroll CEO Tony Haile on Recode Media

“There’s a Moore’s law of advertising, which is that ad formats double in frustration every 18 months.”

If you buy something from a Vox link, Vox Media may earn a commission. See our ethics statement.

Scroll CEO Tony Haile onstage Asa Mathat for Vox Media

On this episode of Recode Media with Peter Kafka, Scroll CEO Tony Haile talks about his company's not-yet-launched product that will let news consumers pay once for a clean, ad-free experience across multiple news outlets and across all platforms. Previously the CEO of Chartbeat and still an adviser to that company, he discusses why he left and explains why it still makes sense for media professionals to monitor real-time data about who’s consuming their work.

You can read some of the highlights from the interview here, or listen to it in the audio player above. Below, we’ve provided a lightly edited complete transcript of their conversation.

If you like this, be sure to subscribe to Recode Media on Apple Podcasts, Spotify, Pocket Casts, Overcast or wherever you listen to podcasts.

Peter Kafka: This is Recode Media with Peter Kafka, that’s me. I am part of the Vox Media podcast network. I’m here at Vox Media New York Headquarters. I am talking to Tony Haile. Who is Tony Haile? If you spend a lot of time listening to a podcast like this you might know that Tony Haile used to be the CEO of Chartbeat. If you spend a lot of time reading Recode you might know that Tony Haile is an occasional contributor and big thinker about the internet. Tony Haile is making a face at me. Tony Haile has a new job, he is the CEO of a new startup called Scroll. I spent some time with Tony a year ago trying to explain what Scroll is so I figured, it’s a year, let’s hear from Tony himself in his own words what Scroll is. Basically, the short version is you’re gonna save the internet, right? Or at least save journalism?

Tony Haile: That’s the attempt, yeah. I mean, well ...

How’s it going? Is it saved?

Any day now, I think, we should see the green shoots of recovery.

I forced you to talk about this a year ago.

You really did.

Well before you were ready. It seems like you’re about ready to start showing people what you’re doing now?

Well, I’ve been a little bit more open recently, yeah. It’s always frustrating when I see your name in my inbox because ...

You’re not the only person who says that.


Most people are my family.

Okay. Yeah, I think I’d only just put the company together and we were trying to be circumspect, even fly under the radar, and then suddenly you turn up and you’re like, “I’ve talked to everyone, I know everything. Come in here and tell me everything.” And I was like, “Oh dear.”

Yeah, that’s my trick. I only know half of it and ...

Exactly, and then I just kind of fill in the gaps like a fool. So, yeah, that was a year ago, and since then we’ve basically just been doing a lot of building, a lot of working with different people and ...

Let’s tell people what Scroll is.

Yeah, so basically the thing that I’ve been obsessed with for the last decade or so has been finding a business model that works for quality journalism. I kind of think it’s important for democracy and a whole bunch of other stuff.

Free press needs to be funded somehow.

Somehow. And we have a long history of philanthropy and entrepreneurs buying things or from having businesses that are unrelated to media supporting it like Kaplan did for the Washington Post for many, many years. But in the end it’d be kind of nice to have a business model that pays for itself.

Used to have one.

Used to have one in terms of ... Well, used to have a business model around selling paper that needed stuff printed on it.

There were newspapers and they made money. They were a good business.

They were a good business.

That went away, that’s going away.

Katherine Graham was probably the most successful CEO of media and probably one of the better ones of the last century. So yes, that was a great business. The print business is declining 20 percent year on year so that’s not going anywhere. Digital media is facing its own challenges, and the interesting thing for me was looking at the main problems. On one side you have the traditional business of advertising where you now have digital gross spend going predominantly to the platforms, to Facebook and Google, so in 2015 it was 72 percent.

People are spending money on advertising online but they’re spending it on Facebook and Google, not on a newspaper or the news sites.

Exactly, they’re spending it through those parties, and that’s going up towards 87 percent they think by 2020. So you have one problem, which is the old business model is looking very precarious and whilst some people are growing, they’re growing in a kind of zero sum way, which is they’re growing within a pie that is shrinking.

And then on the other side I saw this other really interesting thing everyone in Europe was particularly freaked out about, which was the rise of ad blocking. Because what had been happening was in our desire to capture more and more of this shrinking ad revenue, we were going for formats that were only bigger and more intrusive.

Jamming bigger and uglier and louder ads in front of you.

Yeah, and there’s a Moore’s law of advertising, which is that ad formats double in frustration every 18 months. So we go from the static image to the video pre-roll to the autoplay to the autoplay with sound on because apparently we want that, according to Facebook.

Still no autoplay with video, no autoplay video with sound on Vox Media.

Well done, you are ahead of the game — or behind, I’m not sure which.

I need to check every day to make sure it’s not the case, but we haven’t had it so I’m hoping that’s the case. You’re welcome, internet.

Thank you for flying the flag. So you had that issue, which was leading the ad blocking, and the rates over here are like — generally on desktop — are like 15 percent, but in Germany and so forth they’re 25 percent.

Wait, I know the answer. You get people to pay for content instead of relying on ads.

So sure, that is great, and people have been doing that successfully for the top 5 percent of their audience or so.

In big publications, the Wall Street Journal, New York Times, Washington Post, have convinced people to pay to varying degrees.


Smaller publications, Jessica Lessin with The Information has done this successfully. Ben Thompson has a newsletter, Stratechery. There aren’t many more examples.

Exactly. And this is the challenge. You have a few large news organizations, you have basically your top 2 percent or so who will pay for anything or nothing. So I think the New York Times is about 1.8 percent of their audience is digital-only subscribers and so forth, maybe, and maybe another 3 or 4 percent actually see the paywall because most people never see the pay wall.

Because they’re only grazing a few stories.

Yeah, the majority of people come once a week and read one story. So access models work really, really well for your super fans but they don’t solve the problem of how to replace casual fan revenue, which is the revenue that advertising has driven for the industry in general. And so the interesting question for me was, on the one side you have a precarious business model of advertising for casual fans. On the other side, you have a whole bunch of people, you have 236 million people who’ve put up their hands saying, “We don’t want this experience.”

Those people who ...

Who’ve downloaded ad blocking, using an ad blocker as they go. And so the interesting question for me was, instead of thinking of these things as two separate problems, why not think about one of them as a solution to the other. In that, let’s stop saying “ad blocking is evil and wrong,” I’m gonna try and stop it and stop this. We’re gonna treat it not as a problem but as a consumer signal, and our job then is to say, “How can we make ad blocking better?” And so Scroll, in some ways, is our attempt to say, “In the same way that TiVo was a consumer signal that led to Netflix, and Kazaa and Napster and Limewire were consumer signals that led to iTunes music store and Spotify, then ad blocking is a consumer signal that can lead to Scroll.” Which is, how do we get a better ad-free experience across premium media across the web? So that someone can pay five bucks a month and not want to stab their own eyes out every time they have to view media.

So the thesis is, you’re gonna allow people to pay someone some amount of money, I guess some amount of money, which you’ll then pass onto publishers, and then I go to the New York Times but don’t see ads? Or I go to Vox Media and don’t see ads?


And so Vox Media and New York Times takes the money that they get from you and they lose the ad money, cause they’re not showing that impression, but they’re getting money from me passed onto you.

Yeah, and that’s actually the hard part, yeah? Because the consumer side is, in some ways, relatively simple. There’s a bunch of areas where you can make ad blocking better. For example, the rates on desktop are like 15 percent or so. But the rates on mobile are about 1 percent. And the reason is not because people love ads on mobile — they hate ads on mobile even more than they hate them on desktop — it’s because people consume content within social on mobile, where ad blocking doesn’t work.

And so you can do things, we say, “Okay, we can make the ad blocking work on mobiles, a whole bunch of other stuff you can do to make that better.” The hard part is how do you do something in a way that works for publishers, too? Because you’ve had kind of two different approaches. You’ve had an approach which is kind of like the ad block plus or the kind of brave approach, which is, “We’re gonna try and do something in an unsanctioned way, get a whole bunch of users together and then use those users as leverage to try and persuade publishers to accept our terms.” And I believe that those people have never met a publisher in that regard.

And the other way to do it is to do the hard work and try to work out what it is that is a viable model for publishing going forward, put those deals together and then launch ...

The trick is for the publisher, “How do I figure out what it’s worth to me to not show these ads?”

Sure. It’s the opportunity cost of revenue.

Right? And I mean, by the way, you see versions of this in video where Hulu is $8 or $9 with ads, $14 or $15 without ads, and there’s two different things going on, right? One, they’re getting paid for the opportunity they’re not getting to show me an ad, and also they want to sort of establish with the consumer that it’s better for you to have ads, that it’s a better deal for you to watch these ads. They would still rather have that advertising business than not.

Let me put it like this. If you look at ... I don’t know about Hulu’s economics individually, so I can’t speak to that. I could tell you that with most other businesses, the money they make from converting someone into an ad-free is way more than they make for advertising. If you look at Spotify, the No. 1 reason why people convert to a paid Spotify premium is because they want it ad free.

Yeah the Spotify thing’s a little different because their ad business has always been a built on, it’s not ...

It was the thing they started with.

They started with it, but it was always gonna be paid subscription service, the underpinnings of that business have always been about getting people to pay. The ad business is sort of a tagalong.

The same is true of Pandora, of Twitch, all of these businesses.

Yeah, but the newspaper business has always had, “You pay us and we show you ads.” It’s a dual revenue stream.

Absolutely. And the problem, of course, is that when you actually look at the economics of that advertising, it’s pretty shocking these days. I mean, if you say ... Let’s take the top 80 publishers, the DCN publishers in the U.S. They make a total of $70 billion in digital ad revenue across about 234 million uniques. Which mean the total aggregated average revenue per user is about $2.50; all of them together make $2.50 a month per user.

From ads.

From ads.

Right. Not great.

It’s not great. And it’s also relatively easy to beat that with a subscription model.

So that’s the theory for what the publisher should do, and as the consumer, I’m a little confused. Because I can either pay for an ad blocker, have no ads. If I really love the New York Times, pay them $10 a month. I’m confused about the middle ground that you’re pitching to me.

Sure. So one, you can pay New York Times and what you get from the New York Times is access plus ads. There’s nothing in the New York Times that gets rid of those ads. And secondly, you can use an ad blocker and there’s a few things that will happen with that. One is, it’ll be great on desktop to a certain degree. But there’s a few things going on when you do that.

First thing, there’s a quality problem in terms of when you see ad reinsertion happening increasingly with publishers, people are stitching in ads on pre-roll and so forth. We don’t have that. You often have blank spaces or you have to wait for 15 seconds of blank air time before the video starts. You have the fact that the ad blocker companies make their money from inserting ads into your ad-free experience. You have that on the one side. On the other side, you have publishers increasingly going for content blocking, it’s like, “Hey, you’ve hit this site, you’re using an ad blocker.”

Spy thing.

Yeah, so you have a game of whack-a-mole, effectively, on that side of things. And then the big thing, which is like most people consume their content in mobile and most people are not able to ad block the content that they want on mobile.

Even though both Apple and Android have said, “Yeah, you can go ahead and create ad blockers.”

Sure, but look at the rates.

Yeah, yeah. So that’s all the theory.

That’s the theory.

Does the product do this? Can I use it?

Yeah, it’s not launched yet.

It’s not launched yet.

I’ve no intention of launching it until I think quite some time.

Well, I talked to you a year ago and you’d raised money.


So how far off is it quite some time before you launch it?

So we want to do a beta in Q1 of next year and a launch in the second half of next year.

So you got money from the New York Times, News Corp ...

News Corp, from Springer, from ...

All those publishers participated?

They have all been incredibly helpful in this side of things. We’re gonna do an announcement a little bit later, but I’ve learnt with you, Peter.


That you try to push me too soon to tell these things.

It’s almost like we have an abusive relationship.

It’s true, it’s true.

Thanks for coming at 9 o’clock, by the way.

I know. And not even any whiskey here, as well. So yeah, we’re gonna be doing an announcement on a bunch of publishers, I think, in the middle of next month.

And is this an app or I’m on the internet, I go, I click on a New York Time story, I’m on Facebook, I click on an airtime story, and because I’ve registered with you, you beta the Scroll?

Yeah, this is the other key when you look at this thing. Cause one, there’s the different challenges you have. One is the problem with beating the opportunity cost of revenue. So you have to be able to have a model that beats that. The problem with that is you have to be able to do it ... You can’t do it with micropayments, because micropayments and casual fans don’t mix well, and mainly because people don’t have a problem with advertising per se, they have a problem with friction, and micropayments cause friction.

So you need a single-payment solution which you can distribute to publishers, that’s one problem. The other problem is people are set in the past. Many of the corpses in the graveyard that I have to wade through on a daily basis who’ve attempted this thing have come at this by saying, “I’m gonna create this amazing app. It’s gonna be a beautiful app and all you have to do to use it is change everything about how you discover content.” And that to me is not a viable solution. And so the way that Scroll works is ...

You’re talking about “come sign up for this product, download this app which you don’t use already, and then things will be great.”

Yeah, and that doesn’t tend to scale. It can make a very beautiful experience that no one uses. And so the way that Scroll works is you don’t have to change anything about what you do. You sign up for Scroll, and then whether you come to content from Facebook, from email or newsletter, you hit that site, when that site loads it’s beautiful, clean, ad free.

“Hello, Peter, you realize you’re a Scroll subscriber, how awesome are you? You can enjoy this ad-free experience.”

Basically, except without the messaging, because that would be friction.

Good, all right. So we’ll talk in a month, when you save the internet?

We’ll talk at the right time.

All right, let’s hear from an advertiser who supports this free podcast and we’ll be right back.


And now we’re back. Tell me, what’s wrong with advertising? You used to work in the advertising business, right? Sharp Beat was basically a product that helped people who were in the business of delivering ads.

I don’t think there’s a problem with advertising per se, I really don’t.

Good, because that pays my salary — well, some of my salary.

Pays some of your salary. Ask me again in five years.

Pays for this room.

I think that people have a problem with friction, and one of the interesting things that we’ve seen has been the kind of transition in the kind of advertising moments that people have had to use. So there’s kind of three moments where advertising can be done. There’s advertising at the moment of choice, which is, “I’m choosing what content to consume next and I’m presented with commercial opportunities and editorial opportunities.” This is the ads on Google when you search for something, it’s actually the ads in News Feed — it’s why News Feed works so well for Facebook — it’s like I’m choosing what to consume, there’s an ad there, watch it while ... Same thing with Vogue magazine ads. Like ads at the moment of choice. Theses things are considered pretty good by consumers, people seem to be fine with that.

Then you have ads at the moment of adjacency. These are the kind of traditional kind of display ads, it’s like, “I’ve chosen content and whilst I’m consuming this content, there’s gonna be adjacent advertising.” And then there’s advertising at the moment of interruption, which is, “I’ve chosen what I want to do, and the advertising is now asking me to do something other than that.”

Increasingly, with a few things that have been happening — so, one, you had the move to mobile, which has kind of removed all the space for adjacency and so forth. You’ve had the rise of usability, where we’ve said we can’t hide the ads at the bottom anymore, they have to be kind of front and center.

This is advertisers and publishers getting together and saying, “All right, publishers are only gonna charge advertisers when people actually see ads.” Which is a funny thing to say in 2017 but it’s still an issue.

Yeah, it’s still a huge issue. And so you’ve had these things, and then you have the fact that video ads being the only ones that make any money right now. Which means they’re generally pre-rolls, they’re interruptive, and insightful.

So every ad, every kind of structural thing in advertising, has been pushing us towards more and more of a format where we say, “You have chosen what you want to consume. Now let us distract you from doing that.” And it is that friction, not the act of commercial messaging itself that, people have a problem with. If you look at the top two most annoying ads and you look at the research, the first one is pre-roll ads. The second one is actually those bounce exchange ads, they’re the ads where you think you’re about to leave the page and you kind of make a mouse movement and suddenly this huge popup comes up.

Right, ah-ha! Gotcha!

Gotcha, like put in your email address or check out this. Those two things frustrate people more than ...

Because they get more interruptive and more shout-y. Like if anyone listens to radio anymore, like the ads there are bananas, right? Because they’re assuming that you are completely tuned out so they’re literally shouting at you.

Yeah, they raise the volume on these things, they do all these things. So like that interruption from your flow, from the thing that you wanna do, is a problem. And on the other side of things, it’s not making that much money. It really isn’t. Like you look at the New York Times, who’s probably making more money than almost anyone else on ARPU basis. They make 16 cents a month from their users on an advertising basis. 16 cents.

Across all their users?

Yes. As an average revenue per user. Now there’s a curve of that, so the top 10 percent are probably responsible for, if they’re like any other publisher, 50 to 70 percent of revenue, maybe even more. And so you have to be able to kind of handle that. But what is clear is that you can find revenue models that are relatively inexpensive for consumers and can consistently beat the opportunity cost of revenue for publishers.

I want to have more big thinking talk with you about advertising and the state of media.

Go for it.

But I want to ask you about how you got here, first of all. I think a bunch of people know that you used to run Chartbeat?


If you’re listening to this podcast you probably have seen Chartbeat, are familiar with it, but it was providing real-time analytics for internet publishers.


Used to describe it to people who hadn’t heard of it as a sort of Pachinko, right? The screen with these numbers flashing on it and you would sort of see in real time who was reading — or not reading, unfortunately — your stories.

Yeah. It was a mirror to the editors.

Yeah, and there was a time when that was a novel idea, that even measuring your traffic to your stories for some people was a novel idea. And then certainly looking at it in real time was novel.

Certainly. When we started, it was considered crazy. A good friend of mine, Martin Nisenholtz who was at the New York Times in the early days, said that if he had tried to implement Chartbeat in those days he would’ve got punched by someone.

Yeah I remember telling someone at the New York Times what Chartbeat was and they looked at me like I was discussing some sort of weird tribal ritual. And I always had this question myself, which is, why would I want to see my traffic in real time? What as a writer, as an editor, as an advertiser, what is the point of knowing at this very second, 43 people, 500 people, 1,000 people are clicking on this story?

So do you think that editors have a valuable position in media?

I love editors.


Some of my best friends are editors.

Yeah, and so the thing that an editor does is when you write your story, they can say, “You know what? You’ve buried the lead in this thing. This bit is unclear. People aren’t gonna get to this part. Your headline is completely wrong, no one’s gonna read that headline.” They do all of these things based on heuristics of what they’ve learnt in their years. Chartbeat effectively does that in real time with the audience. Which is, it’s saying, “You know what? You did this story that you really, really cared about and no one’s reading past the first paragraph because you buried the lead.” Or, “No one’s clicking on it because your headline was aniline and boring.”

So I get the idea of measurement, this is a thing I’ve talked about over the years. I get the idea of saying, overall, hey that story you wrote that you’re proud of? No one read it. And either there’s a problem with the story or there’s a problem with distribution, or it’s a problem with the headline, or maybe you should get a different line of work. What I didn’t understand was the in-the-moment. I’m not gonna rewrite the lead in the moment, right?

Yeah, of course you are. People do that all the time. Actually, the GM of Vox, Mr Andrew Golis, back in the early days of Chartbeat, I believe that he was at Yahoo at the time, he would be looking in real time at the referrer traffic and he would say, “Okay, I’ve got a politics story here. If I have an audience that’s coming in from Politico, I know that we can kind of juice up this story with more acronyms, we can make it a little bit more inside based if that’s the main traffic source. If it’s coming from the HuffPo, we’re gonna kind of bring it, make it a bit more broad, we’re gonna change even the links that we refer to,” and so forth.

I’ve seen ... One of the most interesting slash fascinating experiences of my life was watching Glenn Beck do his radio show whilst watching Chartbeat data from The Blaze and adapting what he was talking about in real time on his radio show based upon the interest and behavior he was seeing on similar stories on The Blaze. The point is not ... Like for me the thing is, if you’ve spent and invested time in writing an important story, a story that’s important to you, it’s not good enough for you to say, “Well I just didn’t get ...”

I’m sending it out to the world, and ...

It’s done and my job is over. That’s not good enough.

I mean look, I’m an old man so I had to adapt to these ideas, but you know I’m comfortable with the idea that we’ll A/B test a headline in real time, and that headline performs better, let’s pick that one. But the idea of going in and tweaking a story, changing a story any significant way — other than sort of like the base packaging around it — doesn’t seem too practical.

And I get it also if you were back in the days when having a homepage really mattered, if you were a very big site, if you were Yahoo or Huffington Post, it might make sense to move stuff around based on this story’s doing better, let’s feature this more. We know this area of this page does better, let’s move this story into that, see if that works. It seemed — and still seems — like the notion of adjusting your content on the fly in real time is not a really productive way to think about making stuff. You’re ending up getting them sort of just hitting that bar enough.

I would say that the majority of behaviors is probably more around promotion and positioning and so forth as you go, and that stuff absolutely, as you say, has value in real time. The other thing that I would say, though — because this is something which we could debate for a very, very long time — is that I would argue that a journalist’s job is not to write the important stories. A journalist’s job is to communicate the important stories, and if you have not reached your audience with that story, then you have not done your job. And if you are given the opportunity to know that and the opportunity to change it, it is much more efficient for you to change or adapt that story so that it reaches its audience than for you to write an entirely new piece.

So now I feel like we should move into the college dorm session, maybe marijuana will be involved, but this is one of those things that could go on forever, right? Because you can argue yes, it’s 2017 and it’s not enough to just write the story, you’ve gotta take responsibility for figuring out a way to get it to an audience. And you can say this is ridiculous, we’re gonna chase our own tails, and this is why we all end up making the same food videos on Facebook in 2016.

And then the algorithm changes and it turns out we shouldn’t have been making those food videos, we need to move on to something else and we’re just gonna respond to these outside stimuli and whoever has a dominant platform will set those rules. We should make live video, no we shouldn’t make live video. We should make nine-minute videos, no we should make shorter videos. And what we should do is make stuff that’s good.

I totally agree.

And then figure out hopefully with the help of you or Andrew Golis or someone like that, the best way of getting it out there. But asking content makers to constantly on the fly adjust to whatever the new parameters are is gonna end up with bad results.

So I think you’re conflating two things there. Because one, I think that in general you’re absolutely right. I think people should write the things that fascinate them because that’s what makes good journalism. But I think in that context, metrics should just be one of a whole bunch of inputs that come in. And by the way, I don’t think you should ever pay people on metrics, I think that is a very bad thing that gets people down the path, especially around creating content. Around just creating the same thing over and over again, it kills innovation.

There’s a difference between metrics defining what you choose to write about and the effectiveness with which you write. So if metrics is saying only do food videos, only do cat stories and so forth, then absolutely you should ignore the hell out of that and you should try and do the things that you’re passionate about and so forth. If you said like, “I’ve written this amazing thing about ISIS but no one is reading this thing, and this thing I spent two weeks writing and if I can spend an hour just reshaping it, understanding with fresh eyes why I see it’s not working, why people are breaking off, it’s worth it to polish.”

Of course, of course. And this is where the very old school “I never look at metrics and I think the idea of measurement is ridiculous,” that makes no sense. Why wouldn’t you want to know if someone’s reading your work?


I mean, I know why you may not want to know that, right? Because it’s very disheartening.

It can be depressing.

But you know and you should think about what those signals mean and maybe you’re measuring the wrong thing. But and then you brought up ISIS, because I think in, what, 2015, 2016, that was the most heavily read story, most engaged story that you measured across Chartbeat.

Yeah, well, this is the interesting thing here is, which is that the other factor in this is of course the metrics that you choose. And one of the things that Chartbeat got a lot of heat for back in the early days was the fact that we didn’t show pageviews. We didn’t show click data at all for seven years. We didn’t do this. We only showed data around engagement because we had a very clear kind of ideology that we wanted people to be hearing about the content that people are actually consuming, not just clicking on.

So when you start to look at those kind of metrics, you get a very different story than the one that is just like everyone should be doing X and so forth. Because people will click on those stories but they won’t necessarily read them. And when you start to look at things like total engaged time, then yeah, the top story in 2015 was what ISIS really wants by the Atlantic, which got 55 million minutes of engaged time. It was phenomenal.

Your point was, that’s a good thing, this shows you that people will read heavy long stuff, can’t guarantee it but there’s a thing. By the way, we’ve talked about this a few times on this podcast but there’s no measurement, there’s next to no measurement right now for podcasts.


Which is an interesting ...

Starting to open up a little bit.

It’ll open up a little bit, Apple’s gonna start telling us a modicum of information and someone who’s spent most of their career making stuff online, and where stuff is measured all the time, it’s a weird place to be in. It’s uncomfortable, but pleasantly uncomfortable.

Do you think the quirkiness of the podcast industry is in part because of the lack of metrics?

Yeah. I think ... You know, look, if we all decided that almost no one listens beyond minute 32 of the podcast, right? These hour-and-a-half, two-hour podcasts probably go away. And that’s probably not the worst thing in the world. Or they’ll tweak the advertising. It’s gonna change eventually, but it is a little bit freeing right now to know, “All right, let’s just make a thing I like and because it’s almost literally impossible to tell if people like it, let’s make more.”

Yeah, I think everybody in the world is happy with that, apart from advertisers who pay for it.

No, look, I would rather make things that people are listening to and if I eventually see data that says, “Five minutes into this podcast people are done,” then maybe it’s a four-minute podcast.

I highly recommend the British accent, by the way.

Tony, that’s why we have you on. How did you get to Chartbeat? You didn’t start with the company, right?

I consider myself to be employee No. 1. There was a brilliant man called Billy Chasen who was the original creator and then ...

This was out of ...

Out of Betaworks.

Incubator called Betaworks. Still around.

Still around, still going strong. They recently ... One of the more recent companies they did was Giphy, which is a very, very big deal now. So yeah, they’re still going great. So yes, so Billy Chasen was this incredible engineer and also kind of artist and he created the early prototype of Chartbeat. And then I joined him and he relatively swiftly went on to kind of create new things.

He made Turntable FM.

He made Turntable FM.

Which was a huge deal for a week. Two weeks. I loved it.

It was a great, great service. Turntable FM, by the way, also shows one of the challenges of dealing with content providers when you start with the users and don’t have the deals. Because yeah, Turntable FM was fabulous, and then he went on, did Sticky Bits and now is doing great stuff at Facebook. So he was the creator. I always think of him as the founder of Chartbeat.

And what were you doing prior to that? Your Twitter handle says ArcticTony.

So the main thing I’d been doing for the four or five years prior to that was leading and managing polar expeditions.

How do you get into the leading polar ... you don’t mean this metaphorically?


You mean like putting on the gear.

Yeah, putting the big heavy sled behind you, couple skis on your feet and walking slowly into whiteness, yes.

I’m looking at you like you seem like an odd person. But you seem normal.

I had a very weird 20s, basically. I came out of university, I competed in an around-the-world yacht race, then 9/11 happened and I can speak Palestinian and Arabic and so I went to work on a Middle Eastern International terrorism desk doing kind of like security consulting.

For ...

For a company called Control Risks, which is kind of like a, if you get kidnapped anywhere in the world, you’ve gotta call us.

You were an action hero?

I was not an action hero. I was the sandal-wearing geek.

You were the guy behind the action hero saying, “Go here.”

I was the guy telling the action hero not to go into this forest in Nigeria.

Then after that, I got into polar expeditions and had lots of fun up in the cold for many years.

So not the traditional digital media startup route?

It’s true.

It’s true, Peter. That is a true and obvious and stupid thing you said.

You could argue that, like, when I started at Chartbeat I don’t think I’d had zero normal business experience. I don’t think I’d also made more than $20,000 a year in my entire life at that point in time, but I had an amazing amount of fun and it turns out that actually a lot of the things that are relevant for polar expeditions just transfer perfectly to startups. So one, you have to raise an awful lot of money.

This sounds like a Medium post.

Maybe. I haven’t written it yet, though. I’ll see if Ev will let me, put it behind the pay wall, maybe. So yeah, you have to raise a whole bunch of money on something which seems bizarre and may or may not work out. You then have to plan with other people that are sometimes flaky, like dealing with Russian helicopter pilots and so forth. And then you have to deal with ... basically everything is risk mitigation as you go. It’s like, “How do I make sure that I’m still alive two months from now?”

Did you feel like the fact that you had a radically different background than the sort of stereotypical Stanford CS student who’s then funded by Sequoia, whatever that traditional sort of Mark Zuckerberg-ian route, go Stanford. Was that evident to you while you were running the company? That you had a different life experience and that was manifesting in the way you were running the company?

I think it helped in a bunch of ways, and even like the around-the-world yacht race was one of the best things for learning how to lead a manned team, deal with a high-stress situation and so forth. So that side of things, I’ve never had a problem with stress, I’ve never had a problem with anxiety, I’ve always been able to know roughly on the team side of things.

But also, there was a bunch of stuff I didn’t know as well. I remember, this will be the first time I’ve ever publicly talked about this. In the very early days of Chartbeat, 2009, I’m dealing with media companies, I don’t really know anything about media at this point in time, and I’d go into a ...

You’ve created this thing you knew a media company would want.


But you didn’t really know a lot about media companies.

Yeah. That tends to be a fairly normal thing, I think. So a media company would say to me something like, “Well, I don’t know how this is gonna interact with my DMP.” At the time, I had no idea what a DMP was. I didn’t want to look stupid.

Tell us. Tell us what a DMP is.

It’s a data management platform.

Ad tech.

It’s ad tech. And so rather than saying, “I’m sorry, I don’t know what this means,” and my nascent start of getting kicked out of the room, I’d do something called the broad church answer, where I’d say, “Well, you know, DMPs are kind of a broad church. What do you mean by DMP?” And basically did that repeatedly for the first year until people told me everything that I needed to know about some of the ins and outs of the industry and so managed to kind of catch up pretty fast on that. But yeah, I felt like I had a very orthogonal skillset to someone who’d grown up through marketing or through sales or through engineering or product.

You did that for how long?

Seven years.

Seven years, company did not get sold. There was not an exit.

No, still going strong.

Why’d you leave?

So I think two reasons. One, I missed the sense of risk. I had this fabulous executive team, I had a great COR in John Saroff who’s now CEO of Chartbeat, and my job had been less about taking the big risks and more about kind of like setting okay asks for the quarter. And that I still enjoyed ... I still loved the team, but that was less fascinating to me as a kind of experience.

And secondly, I’d had this idea and it was the idea that I couldn’t get out of my head, the idea that you think about in the shower. And it would’ve been wrong for me to have stayed at Chartbeat whilst I was still passionate about this idea. So I called up my board members, and I spoke to the exec team. I said, “This is the thing, this is where I’m at.” And they were incredibly supportive.

So you ended up, you basically signed up ... Everyone in digital publishing has a Chartbeat account.

About 80 percent.

At Vox we have it, even though we were all our own, but everyone’s got it. So I think it’s, in one way, you guys were as successful as you could’ve possibly imagined. The flip side is when you were raising money last year and I was reporting about this, I was asking some folks and they said, “Well, the problem with Chartbeat, we all love Tony, everyone loves Tony, but they didn’t really build that business as it should’ve been. It should’ve been a giant advertising business. Those things should be something that WPP and Omni Com all use or own. This thing eventually should have been a big ad play and the fact that it wasn’t is a failure.”

So I would say two things to that. One I would say ...

Which I told you before.


It’s not the first time you’ve heard that because I’ve told you that before.

It’s true. I’d say two things. One, so I’m still an adviser to Chartbeat and I can tell you, the last two quarters they’ve had have been their best quarters ever. So they’re still growing very, very nicely as they go. And the CEO’s doing a far better job than I ever could have in that regard. But the second thing is that Chartbeat is a mission-driven company. There was definitely an opportunity for us to go and become a Tabula or Outbrain, and I don’t think that would’ve been particularly good for work.

Because you have all this data.

We have all the data, yeah.

So why not use the data to serve up awesome ads or at least lucrative ads or at least ads?

I’m familiar with a company who’s had similar levels of data who’s been able to serve up awesome ads. The thing that we cared about was, how did we make sure that the journalism we care about today is still around tomorrow? And so the things that we tried to do, and the bets that I made, were around how do you try and change the economy? Not just in terms of how do we sell a different kind of ad, but how do you actually make it possible for the industry in general to get better? And that was where the whole idea of time-based selling came from. We were the first to do that, we built the systems to be able to do that, the FT ...

You’re saying you ... Yeah, the engagement, right?

Yeah, selling on attention, selling on these things.

Which was a great idea. Hasn’t taken off, right?

Yeah, I mean like the FT just ... I posted today, I think, saying they’re doing pretty well on it. But yeah, you’re right, it’s ... The idea was sort of the data is definitely there for it. The challenge is that you have a kind of chicken and the egg problem. Which is, you can go to an advertiser and they’ll say, “Great, how many publishers are doing this? And then if all the publishers are doing this, then I’m in.” And then you go to the publishers and publishers say, “Wait, how many advertisers are kind of campaigning with this quarter?”

So there’s that chicken and the egg problem and in the middle that you have the agencies and the media planners who enter kind of really optimistically. So I can build a great city, but I cannot add a line to an Excel spreadsheet in it for a media planner. It is an incredibly difficult task to do and you have clicks, you have impressions.

Right, you can step out, you can say, “Here’s my new novel thing.” It’s a good way to get in the door, I think, for a lot of people. “I’ve got a new thing to show you.” But to get someone to sort of re-up and pay you over and over for a new metric that no one else is using ...

To change the currency is an incredibly hard thing with so many moving parts. There’s also the right thing to try and do, and I’d probably approach it in a different way, but if I had the opportunity between trying to choose something which I thought was lucrative but bad for the web, or trying something that was more of a Hail Mary but if it worked would make journalism more secure, make the things that I cared about more secure, then I’m gonna do that every time.

So in addition to being an arctic explorer and a startup CEO twice over, you’re also an unpaid columnist for people like Recode?

Occasionally. When you publish my stuff.

Love having you. Was that a jab? Is there something that we should’ve published? So you can look for Tony’s stuff on our site, people love reading it. You wrote something for us a couple years ago, it was a multi-part series about Facebook and there was a picture there of the planets, and the idea was sort of Facebook was the biggest planet that was dwarfing us all. And then more recently, you had a piece that said, “Facebook may finally have to compromise its user experience in order to keep growing.”


And I thought, when I looked at them both just now I was thinking, “Is Tony saying actually that there might be a limit to Facebook’s power in publishing? To Facebook’s unchecked dominance or ...” Was I missing that?

It’s less about dominance, more about Facebook running into some of the same constraints that have been bedeviling publishing as it starts to reach certain limits. So whilst ... So Facebook’s revenue growth has been driven largely by two things within Facebook proper — forgetting Instagram and WhatsApp and these things for the moment. It’s been driven by user growth in the West predominantly, which is where the advertising revenue is.

They have two billion users worldwide.

Yeah, and it’s been driven by News Feed, which has been the most successful ad unit. And with those two things, they’ve risen to undreamed-of heights. Now the problem is this: User growth is about 1.2 percent a quarter in the West now, and so if you’re just relying on user growth you’re not gonna see the kind of revenue growth that you want to do.

Amazingly though, they post amazing numbers every quarter.

They do, and they’re in some ways as amazed if you speak to them internally, they’re as amazed as other people are. But the way that they’re doing that is by increasing ARPU, okay? Which is their average revenue per user, they’re increasing the price of their advertising.

Which, by the way, if they’d sketched it out, would’ve been exactly what they said: We’re gonna keep growing and eventually we’re gonna make each one of these users more and more valuable.

This is actually the ... There’s kind of three stages for any company going through this, okay? The first is user growth, cause you see the most flexibility, you can be easiest on like ad load and everything else.

It’s the standard internet company business plan.


Grow, then make money.

Then you go for ARPU, and you try to maximize ARPU. And then after ARPU it’s about trying to capture as much value from the value chain. So that’s the point we’re saying, and this is the kind of enticing foods analogy, which is, at a certain point, when you’ve maximized the ARPU that you can, you say the way that I can do this is I can use my leverage to make sure that out of the total margin across this value chain, I’m getting the majority of it. And that’s where you ...

So from egg to sandwich I want to ...

I want to have an awful lot of chicken farmers who make just enough to survive and not much more than that. And so that’s the kind of ... Facebook isn’t quite there yet. But that’s the obvious kind of third stage as they go.

Because the Facebook story, right — which they’ve been telling for a long time — is just wait until we get to TV, there’s $80 billion in TV and we have a Super-Bowl-sized audience and eventually, one day, all that money’s gonna come over here.

And it’s hugely problematic, in some ways. I mean, they have in Fidji Simo, they have probably the best product manager in the company working on this.

It’s the woman who runs the video for them.

She’s brilliant. But, faces a number of challenges around video being the kind of revenue driver for them. The problem on the News Feed side of things is that pre-rolls are probably the most successful video ad format, but News Feed, unlike YouTube, is a passive discovery mechanism. And so if you were to do pre-rolls, you would literally just have a feed of pre-roll things. Which no one would consume.

And so they have to go with mid-rolls in News Feed. But the problem is that the mid-rolls, people don’t tend to get to or they switch off as soon as they go to and just move onto the next thing, and so we’ve seen reporting in the newspapers recently, in the trade press, around how little money people are making from the mid-roll side of things.

So Facebook knows that it has that problem, which is why it’s created Watch, which is this whole new tab where they can say, “Now we can do what YouTube does, which is we can do intent-driven video. And with intent-driven video, we can start to do pre-rolls, we can’t start to make some real money around video.”

Have they said they’re gonna do pre-rolls for those?

They’re going to do pre-rolls.

You’re intuiting it? Or?

I’m intuiting it.


But like, mid-rolls are not the way to drive revenue.

If you remember a few years ago when Facebook first started playing videos, it said, “We’re gonna run the ads after the video.” Which, by the way, sounds good.

Great for the user.

For user I like the idea, it seemed like ... Because you can go to the advertiser and say, “Look, anyone who’s watching this at the end, they watched the whole video, they love the product, they feel great about it.”


Turns out advertisers didn’t go for that.

And users don’t act like that, especially if — and again, this is the kind of challenge of the News Feed, yeah? Which is the next piece of content is just a scroll away.

So does the fact that Facebook is bumping up against these limits, does that mean that for the publishers who’ve been watching Facebook and Google — the sort you were just describing, eat more and more of the pie, whatever metaphor you want — is that good news for the Vox Medias, the New York Times, the Minneapolis Star Tribunes of the world? Oh, finally there’s an opportunity for us. Or is that not good?

So I don’t think so. I think that what you see here is you see a challenge on the user experience side for Facebook to try and overcome, and they’re very good at overcoming those things as they go, but it’s not something where it’s going to change the trajectory of ad dollars. You’re not gonna see WPP suddenly saying, “Wait, you know what? Facebook is starting to mess with its user experience in some way, so now we’re gonna go back and just use traditional media and just put all our money there.” That’s not gonna happen as we go.

So their problem does not create an opportunity for the rest of us.

And, well, this is the kind of interesting thing. There’s two schools of thought. There’s a school of thought which says that the platforms are the new cable companies and we all are the new cable channels on this side of things. And that we can all be as successful as media companies, as the kind of early CNNs were and everything else. And that is one school of thought.

The other school of thought is to say that that is completely the opposite of what’s really happening here, which is with cable TV you had intent-driven consumption around ...

And really limited distribution, right?

Yeah, but with real brand affinity, there were high costs of entry, there were non-competitive business models.

Right, maybe there were 500 channels, really there were 50.

Yeah and like to start your own channel cost a lot of money as you go. And the reason why ESPN gets paid so much money is because of its marginal trend contribution, like if you don’t have ESPN then you have a problem. And so you had all of these things where ... And the cable companies themselves didn’t really provide any value themselves, they were pure pipes.

Now with the platforms is you have almost the opposite of all of those things. You have ... It’s not bad-intent driven with high brand affinity, it’s a passive-driven experience where people aren’t even aware of the creator of the videos they’re consuming most of the time, or the content. I think it’s 64 percent of the time they’re unaware of the brand behind the content they’re consuming. You have almost zero barriers to entry and you have the fact that this content still really makes up a minority of the total content that’s being consumed on the platform.

Right, because we’re all consumed with “What’s Tom doing on Facebook?” But really it’s still cat pictures, cat pictures.

Or it’s pictures of your friends and so it’s a tiny percentage of the stuff that actually kind of drives the business. And so all of those things mean that the leverage that any media company has is vastly diminished compared to how the cable channels were in the cable era. And by not having that leverage, over time, margins are crucial, the point of control. If you have no control, you have no margins.

After the first round of the internet, where a lot of newspapers had played around with the internet, put their stuff up for free, maybe they charge, and then things sort of went south and then folks said, “The original sin was that we devalued our content right away, we made it free or we made it too cheap. We’re not gonna do that again.” And then Facebook shows up, and kind of slow motion, they’ve been saying, “Give us more stuff, give us more stuff, give us more stuff.” And everyone’s been saying, “Our eyes are wide open, we’re fully aware of what’s going on here and we’re gonna be very wary, we’re not gonna make that mistake again.” But yet it looks like ... Was there any way for publishers not to have done what they’ve done, which is basically sort of cede a lot of their business to Facebook?

It feels a little bit like Greek tragedy in many ways, in both kinds, in that you can fight and you can have heroic efforts.

I’m nodding because I’ve seen “Time Bandits,” and there’s some Greeks in that. I’m not really familiar with Greek ...

That’s true, it’s all Greek to you. The gods define it. The structures define it as you go, and one of the challenges you’ve had for publishers is you have a model, which is basically around, like, can we capture someone’s attention and can we put an ad next to it? You found someone who was able to do that more cheaply and someone who had the traffic and there’s structural problems in terms of being able to avoid that as you go.

Now some people with niche content can say, “We’re gonna put up a hard pay wall” and you can do like a Ben Thompson thing or what Jessica’s done very successfully with The Information. But in general, I would argue that content is a commodity for 95 percent of people. Content is what happens between 2:50 and 3 pm whilst you’ve got a break between meetings and putting up friction while trying to say, “We’re gonna charge for that when other people aren’t.”

Right. So the anomaly was sort of that you were able to build a big business on that prior to this. The fact that now it’s being commoditized is sort of actually where it should.

If you think of it like this, the business has always been around distribution, okay? The newspaper business was around trucks and newspapers, like when Ben Franklin started one of the first newspapers, he did it because he just needed more stuff to put into print. It wasn’t because he decided to start a newspaper. And so we went from trucks and papers and barrels of ink to the new distribution, which is the social platforms and so forth, and that’s always been where the money is. That’s not changed.

You’re gonna change it. Scroll.

I’m trying to.

All right. You were cagey but you’re gonna make some announcements still this year? And then there’s stuff for us to play with early next year?

Yeah, I mean, when we finish this podcast, I’ll show you it live.

Maybe we’ll turn the video camera on for that.


All right.

The challenge for us is this, which is in order to make something valuable for consumers, you have to have a critical mass of publishers on board. People don’t pay for an ad-free version, they wouldn’t pay for an ad-free version of Vox, very few people would. And if they would, they’d pay for something else even more money as they go. In general, people have tried single site ad-free models, they’ve ...

How many publishers do you need for this to work?

So I don’t think of it in terms of publishers, I think of it in terms of content, consumption and coverage, okay? So this is actually — you’re smiling, but there’s actually ...

I don’t wanna hide it.

Yeah, because the thing you could do is you could do like — I could have 1,000 tiny sites and I wouldn’t be covering that much in someone’s consumption. So the way I think of it is like this, if you look at the top 200 domains, which if you look from a precode of consumption is the majority of content consumption, in the U.S., and then you break that out by how much of the total consumption they cover, someone like ESPN is around 14 percent, CNN is around 8 percent, the New York Times is around 4 percent and so forth. And so I think for a beta, I probably need about 50 percent of content consumption.

Five zero?

Five zero. For general launch, you need about 80.

Are you there? Are you close to 50?

I’m confident now. I’m more, like for the longest time I thought I was absolutely crazy, I no longer think I’m crazy.

All right.

Other people might.

Yeah, but you got the English accent so it sounds convincing.

It does help.

Tony, thank you for your time and your patience, this was heavy but good. Again, you can read Tony, he’s available in many places including, Arctic Tony on twitter. Everyone here knows how to use google, they’ll find you.



They’ll figure that out as well. Thanks again for your time, thanks to you guys for listening.

This article originally appeared on

Sign up for the newsletter Sign up for Vox Recommends

Get curated picks of the best Vox journalism to read, watch, and listen to every week, from our editors.