On this episode of Recode Media with Peter Kafka, Eugene Wei — an early Amazon employee who went on to work at Hulu, Flipboard and Oculus — talks about how all companies have a ceiling to their growth, but the ones that can figure out what that ceiling is can adapt and keep growing beyond it.
You can read some highlights from the interview here or listen to Recode Media on Apple Podcasts, Spotify, Pocket Casts, Overcast or wherever you listen to podcasts.
Below, you’ll also find a lightly edited transcript of the full episode.
Peter Kafka: Eugene, welcome.
Eugene Wei: Thank you.
You’ve got a long, crazy bio, which is one of the reasons I want to have you here. What’s the best way of describing you in this state today? I don’t want to call you the former head of Flipboard product. My worry is that the podcast label is going to say, “Eugene Wei was the former head of product at Flipboard.” That’s you, but I want to call you a thinkfluencer.
Okay. I haven’t heard that term before.
That’s not polite. Oh, no. People actually use it. Well, the fake Jeff Jarvis uses it.
Smart guy, guy who’s worked at a million media companies that I’m all interested in.
That’s better. Probably better than thinkfluencer.
What’s paying your bills today? You said you’re in New York for meetings. That implies a job or some kind of professional …
No, no. I don’t have an official job right now. I’m free and footloose and whatever the saying goes right now. Trying to figure out what’s next.
Okay. I will fill in some of the bio for you.
You are an early employee of Amazon. Somewhere in the 300 range.
Yeah, 300s I think it was.
Yeah, that’s pretty early. Worked there for a bunch of years as a business analyst. Is that the correct term?
I did that for about a year and a half. Then I moved over to product to work on the video store launch and then stayed on product, worked a bit on the third-party seller platform, worked on the apparel launch. Then the last thing I worked on was the Amazon Web Services team.
You did a million things at Amazon.
Left Amazon, went to film school for a couple minutes. Then had a variety of jobs, including product at Hulu, product at Flipboard, like we talked about. Something called Airtime, if you have a deep knowledge of failed internet startups. You ran video at Oculus for a couple years.
Now the reason a lot of people have heard about you … I was emailing you before. I feel like you’re my indie band that blew up. You’re not a professional blogger because you’re not getting paid for it.
But you’re dropping all this deep knowledge at your personal blog, which is called …
I guess it’s titled Remains of the Day.
Remains of the Day.
Many of you will have read this blog. If you haven’t, you’ll go read it after this conversation. Don’t stop and then come back. Just listen to us talk. It’s this great combination of Eugene’s personal musings about culture and tech and product that’s all mashed together. It’s like going to college for free from Eugene’s brain. There was a post you did a few weeks ago that I think got everyone’s attention maybe over a slow news weekend. It was Memorial Day weekend. That was part of the reason.
That’s the key.
Bill Gurley gave it a shoutout. This is the one where you walked through the flaws of all the existing social media products.
Instagram, Snapchat, etc. I want to talk about that for a second, but I just want to ask you, because it seemed like that one popped in a way other ones haven’t. What is it like when a blog post you put up for free grabs everyone’s attention? People like Bill Gurley who runs Benchmark. What happens after that? Does your inbox get flooded with requests? Are people wanting to come pick your brain?
Yeah, it’s a mix of all of the above, probably. It’s not like it happens that frequently for me. That one probably is the thing that I’ve written in my life that’s been the most read and shared. It was new for me. Probably for a lot of your listeners who have a lot of followers on Twitter, they’re used to just their phone being a constant stream of notifications.
People want things from you, right? I would assume. Besides just, “Come blog with me. Come podcast with me.”
Yeah, some. Certainly it’s gotten me a bunch of interesting proposals and offers and certainly a lot of companies were curious about the ideas and wanted to talk about them in depth. I was already spending some part of my time advising different companies that would just bring me in every so often. I just got more of those.
I don’t want to oversell this, but I’m going to really hype this. This is great writing. It’s super interesting. It’s the kind of thing you could charge money for. It’s the kind of thing Ben Thompson at Stratechery sells on a weekly basis for $10 a month. Tens of thousands of subscribers make real money from it. You’re doing it for free. Sometimes, especially in the earlier days of blogging, you’d see people like Fred Wilson from Union Square would blog a lot. It was very actively promotional. He wanted to talk up his business, talk up individual products, startups he’d invested in. You don’t seem to have that book to talk up, other than yourself. It’s just literally something you’re doing for giggles? Or is there a professional reason to do this?
It never started out as a professional thing. I started my blog in 2001. This was back in the days of … I started on Blogger. There were Live Journal. Nobody was reading it. The audience …
You were still working for Amazon.
I was. It was just a personal thing for me. I can’t say that I was ever in it for some professional purpose.
Now doing it now. Doing something like the posts we’re going to go through in a little bit. It takes you a long time to write. It’s obviously well thought out.
Yeah. I think I enjoy putting ideas out there now. Well, first of all, all the posts that have really popped for me have been on tech or company things. It fits the audience. If it gets on Techmeme or in the early days, Hunter Walk might tweet a link to something and it would just pop onto Techmeme every so often. I’ve tried to avoid feeling like I should just write that type of stuff because I’m interested in other things, but certainly that gets the widest audience.
For me, I enjoy putting ideas out there because I got more emails on this than anything I’ve ever written. Some of them had interesting feedback. I like conversation as a way to challenge ideas and a way to mold them. Putting it out there for a wide audience gets me a broader set of people to talk about it with.
Was this a reflection of what’s in your head? I’ve got to confess, I wish you had done this earlier in your life, this kind of blogging, because I didn’t notice it when I first met you. When I met you, your job was … I think you were doing product at Hulu. The reason I knew about you is because you were the guy … The person who had done comms had left. You were the guy whose job duty now included responding to Peter Kafka’s annoying emails and phone calls. You were part-time comms guy.
Right. I could have been a bot, actually, because I mostly said that we couldn’t comment on anything.
Yeah, you were very good at that. I will confess that I’ve known many people who were in comms before who were very good at their jobs, but I generally don’t assume that people in comms are writing super-deep dives into product stuff as well.
Sure. You know the game.
You’re doing that on the side.
You know, when you’re actually at a company in comms, you really can’t be that interesting. It’s not your job to be interesting. That’s a different game that comms people have to play with the media.
Yeah, I just feel dumb. I didn’t realize that there was a Eugene Wei resource available to me. I just didn’t know how to access it.
It’s sparse. It’s certainly not like Ben Thompson or anybody like that that’s writing still regularly. Every now and then.
Let’s talk about the post that popped for you.
I had to Google up what an asymptote is. Am I pronouncing it correctly?
How did you learn what an asymptote was?
Anyone who studied certain parts of math probably back in the day are familiar with that.
It’s something you picked up in high school or at Stanford?
Probably. Somewhere in math.
Okay. The invisible asymptote. An asymptote is? For the handful of listeners who don’t know what an asymptote is.
Yeah. When people talk about something like converges on an asymptote, it’s like if you look at a curve or something like that and it starts to flatten out over time. The asymptote is that ceiling.
For dummies, this would be an invisible ceiling. The idea here is every company has an invisible ceiling. It’s invisible. They don’t know what it is, but their growth is going to stop. They’re going to hit it. Figuring out what that ceiling is allows you to then keep growing.
Yeah. I think there’s been so much good writing done about startups. Trying to find product market fit. That’s a different set of problems, but there’s a lot of stuff. Chris Dickson has written about the idea maze and Andreessen about how important it is to find …
You’re trying to get out of the gate.
People are familiar with that idea. How do you get going? How do you find a thing people want?
Now you’ve been growing. Your growth is going to stop. How are you going to deal with that?
Yeah. I think there was just less writing maybe about that. On the traditional S-curve, it’s the top where it starts to flatten out. I’ve been at big companies and I’ve also looked at the companies in tech and long been fascinated by that, which I just think is somewhat understudied. Part of maybe why that post resonated was I think there are just a lot of companies — even a startup that’s found initial product market fit — that quickly hits some flattening or some plateau.
What I wanted to do is try to just talk about some different companies, talk about some theories on why they might be hitting that.
Right. I’ve always thought about this as a problem with management. The company that breaks out with 10 or 100 people, the problem usually is that the people who run that company with 100 people aren’t equipped to run it as it hits 1,000. I think that is true.
Yeah, that’s definitely true.
People don’t talk that much about that either. You’re saying, “Look, it’s the actual … I want to view it from the product lens. What is the issue with the product that is going to cause that problem?” With Amazon, you say the thing we figured out early on was people didn’t want to pay for shipping. We spent a lot of time trying to figure that problem out and we did. That helped allow us to keep going.
Yeah. I think at Amazon … Well, both things you said are right. One is that there’s very clearly a set of internal cultural, organizational, structural things that many companies run into at certain sizes. Those are probably fairly fixed on employee number and size. Actually at Amazon, I felt like there was a lot of study done on that as well.
Separate from that, there are some things just with your product and service and interaction with the market that can create a natural ceiling. That is an area that’s I think often more difficult to understand because it can be multidimensional. There could be very many different things that cause that.
Also, it seems that people who run a company, who start a company, might be more willing to accept in a way … Maybe I’m just making this up on the spot, that there’s an issue with management. Maybe it’s even an issue with the founder/CEO and they’re not equipped for that. They might feel accepting of that criticism because they can get their head around it as opposed to the thing you built has a natural limit and isn’t going to move beyond this unless we change it in a significant way.
Right. When [Marc] Andreessen writes about, “Hey, product market fit is the only thing that really matters in the beginning for a startup,” that’s very clarifying because you’re like, “Hey, look. We’re just going to keep pivoting and trying different things until you get some traction.” You’re accepting of that. Once you have some adoption, though, I think it’s very, very easy to fall into the trap of being blind to some of what’s holding you back. One of the things I talk about there is, sometimes your earliest customers are the ones that will blind you to that.
Right. Let’s go into the initial companies you’re looking at. That’s the one you bring up at the beginning when you’re talking about Twitter. Twitter’s fundamental issue is the thing that made Twitter successful, basically. Am I summing that up correctly?
Yeah. I think Twitter is one of those cases where every little change they’ve made to the product, the heaviest users have really protested.
There’s people who’ve loved Twitter for a long time. You and I are in that group.
I don’t care what happens to the product. The point is it really works for them, and that any attempt to expand it into a bigger audience is not … runs the risk of upsetting those people and/or those people are too loud in Twitter’s head when they try to make those changes.
Yeah. There are certain products for which I think later users and later adopters actually do want the same thing as the early adopters and there are certain products for whom that’s not the case. Actually, my theory on Twitter is that that is the case. They’re one of the latter type of companies. Your most vocal and heaviest users may be the ones that prevent you from making the types of changes you need to to reach a broader audience, whether that’s on the same product itself or on derivative products that better serve another audience. That’s another matter entirely, but that’s definitely a trap.
Quoting you, “falling into the trap of thinking other users will be like you is especially pernicious because the people building the product are usually among that early-adopter cohort.” It reads better than it sounds coming out of my mouth. Do you think that there could have been or is still a way for Twitter to innovate its way into a bigger audience or do you think this is just a finite audience? It’s around 300 million people worldwide. Everyone has now seen Twitter and they get what it is. They either love it or they don’t want it.
Yeah. In the current form, Twitter, the app that people have on their phones and that one that we’re all familiar with, I certainly think there’s some asymptote to how large that audience can be. But if you look at the raw material of Twitter, how many people are on it, commenting on different things, posting links, posting media, commenting on other things that are happening in the world, from that raw material, there are actually probably other things you could build. Other products or services that would be more broadly appealing, appeal to different customer segments. I think that actual data and the structure of that data is actually very fascinating, sort of in the way that Google looked at web links way back in the day and said, “You know what? There’s something we can derive from this structure of the web itself, which is different, but very useful.”
Just so I’m clear, do you think the idea here is Twitter could build new products that would appeal to other people who aren’t using Twitter or that Twitter could make new products that would appeal to the existing users and keep them around longer, keep them more engaged, give people a way to monetize that audience?
Yeah. I meant more the former, though I certainly think the latter also holds true in that there are probably things you can do on Twitter now to increase the usage among the current cohort of users. That is work they’re continuing to focus on and I think it’s important, but I’m more interested in the first case just because the expansion of the total addressable market for Twitter has long been something that people have puzzled over.
The existing Twitter managers, some of whom listen to this podcast, called you up and said, “All right, smart guy. Let us know what you think these mystery products are.”
Not really, though. I know people there. There are a lot of smart people inside Twitter, working hard on this type of problem. So yeah, fingers crossed.
We went through Twitter. Let’s talk about Snapchat. Snapchat seems to me like the same issue, but do you have a different take on it?
A bit. Snapchat is fascinating to me in the way that a lot of social networks are fascinating to me because I think there are certain … What might you call them? Truths about people as they grow up and what they want from the world. I actually think there’s one way of analyzing social networks, which is just on the utility they provide. Messaging apps are probably the cleanest or purest example. We use those to communicate with people that we know. There’s utility in that.
One traditional school of thought in social networks is that in the beginning when there’s nobody on the social network, you have to provide some single-player utility. Then once a lot of people are on there, if you’ve structured it properly, it becomes a multiplayer game.
Single-player utility meaning you can go use it. You don’t need to know anyone on the service. You can still get value.
Right, even if there’s nobody else on the service.
You can go play Fortnite without having any other friends that play Fortnite.
Right. Some games are like that. That’s one way of analyzing social networks. I actually think that that way of thinking about them is actually pretty easy for most people to grok. It’s pretty clear if something is giving you a single-player utility. Most people can see that.
But another way to think about social networks is in terms of how you accumulate social capital. I’ve found that a good way to think about early social networks and whether they’re getting traction and whether they’ll stick around is whether you as a person can go on there and accumulate some level of social capital. When I think about Snapchat versus Facebook and Instagram and all of that, that’s an alternate way …
What does social capital look like in a social network? Are you talking about “Likes” and that sort of scoreboard counting that we’re not supposed to care about, but a lot of people care about because we’re petty people?
Right. Humans want a lot of things. You may want financial capital. You may want money. Social capital is a little different. Social currency. “Likes” is one way of measuring that.
Some kind of recognition that other people can identify.
Yeah. Followers, your place in a social hierarchy and how you measure that. Social networks are all pretty good about coming up with the basic ways for you to accumulate some of that. “Likes” and followers are a very common way. On Twitter, you can get retweets. Anything like that starts to give you a sense that, “Oh, you know what? I have an audience that have a following. This is great.”
With Snapchat, that doesn’t really exist. Right?
Well, what’s interesting about Snapchat was I think if you look at young people on Facebook, their parents were joining. A lot of people who are in their teens and things like that are still trying to figure out who they are. They’re trying on identities. In the old world, before social media, you could kind of do that without having some permanent record that would follow you around in life. They were a double whammy for Facebook at the time. One is that your parents were there, monitoring everything that you …
It’s uncool and there’s a permanent record.
I get the value of that for Snapchat.
Flip that around. Your parents literally can’t figure out how to use the app and there’s no record. What is the limit then for them?
I think one of the challenges for Snapchat is there’s been a lot of discussion about them broadening out to a wider audience. There are a couple things. One is this sort of ephemeral content bit. I actually don’t know that older people need that as much. I think as we grow older, we actually place more value on nostalgia.
Yeah. I love the Facebook feature where they show me pictures of my kids that I posted from nine years ago. That’s my Facebook group.
Yeah. When I grew up, my dad would film everything we did. He was one of those early camcorder adopters. At every event we’re like, “Dad, put the camera away. Why?” Years later, after my mom had passed away, especially, we went back to all those old video tapes and transferred them to DVD. Now you watch them. You can sit around, watching them for hours. I was like, “I wish I had the camera on even more.” You can go back and just live the moment. The time value of content has a weird curved shape to it. I think for Snapchat, that’s one challenge. Look, I don’t know that ephemerality is a key appealing …
Again, it seems like the Twitter problem, the thing that you built that lots of people in the beginning responded to, has a ceiling. There’s only certain people who want that version of the product.
Yeah. The second thing is that if you look at Snapchat’s interface, which is famously opaque to older people … Don’t know how to deal with it. It is sort of really prioritized for a generation that’s super visually literate. I mean, it still opens to the camera.
When you watch young people versus old people using Snapchat, my observation is that they use it very, very differently. Older people I know on there are usually posting a lot of stories, using it kind of like a broadcast medium, like you’d use Instagram or anything like that. Whereas young people mostly message each other with it, and mostly do it in pictures and videos.
Even though content is ephemeral on Snapchat, I think they’ve done a very smart thing in understanding how young people want to accumulate social capital. In the earliest days, I don’t know if you remember, there was like, kind of, the best friend leaderboard on Snapchat.
It may predate even when you used it. But based on your interactions …
I’ll confess to be an almost-never user of Snapchat.
Only in an obligatory, professional way.
Confirms my thesis. You would get a ranking of who a person’s best friends were based on the volume of Snapchat interactions. That’s a form of social capital. It’s like the structure of your social graph is being made more visible. When they took that feature away, the users of Snapchat protested heavily. What they replaced it with is something super interesting, streaks. So now, every day that you and a friend snap each other, your streak is incremented by one. If you forget to do it one day, and you both don’t respond, then the streak is broken.
Is that public?
You can see it on your own friend list. You can see that on Snapchat. If you watch young people, like my nephew, for example. I look on his phone, he’s got dozens of streaks going.
Right. But that’s an internal measure, it’s not a public…
Right, you can’t see …
It’s not a bragging point. Right? It’s a nudge.
What matters, the social capital here that matters is between you and your friends, right? And I think that is smart. The problem with the best friend leaderboard is you could only have one best friend.
I’m still trying to get you to offer advice for Evan Spiegel. What is the thing that he doesn’t get about his product growth limit?
Well, I think, actually …
His user growth limit.
I’m not sure that Snapchat should focus on trying to get old people to switch over to using the product. I would be much more focused on the next generation of young people and whether they will be also using Snapchat, or whether they will find some other network of their own. Because, to some extent, I think the old people are … They don’t need a lot of those things, and it’s a distraction to the company to try to rejigger the entire product interface to attract those folks.
Let’s move to Instagram and Facebook, because they’re both owned by the same company.
Some similarities, both huge.
Instagram’s growing insanely fast. Facebook is still really growing, given that it’s at two billion. It should’ve topped out a long time ago.
There’s a lot of issues, specifically for Facebook. But it seems like they kind of have a handle on product.
Yeah. Facebook, kind of like Amazon, I think, is a company where you have to slice it and look at it in different markets. So Facebook in different countries really serves different purposes. In some countries, for example, you travel through Asia …
It’s the internet.
Yeah, it is the internet. It’s weird. When I’m here in the states, I might look up a restaurant on Yelp or OpenTable or Google Maps. But when I’m abroad they don’t have something that’s quite like that. You just go to Facebook and you find the page for a business, and that is their presence on the internet. I think it’s hard to analyze Facebook as just one monolithic thing. It’s a series of individual social graphs, and then there’s kind of like the country-level social graphs that fit together.
I think they are, probably in many markets, still growing. Because they are kind of like AOL was back in the day, sort of your one way to get on the internet. The U.S. market, because it’s more mature for Facebook, is where I spent some of my time in my piece thinking. Because I think that’s an interesting subset of Facebook. That’s the one where the graph has gotten so large that I think there may be some negative economies of scale.
Meaning the very size of the network sort of works against them. So every additional person on Facebook makes the service a little bit less useful, or less usable, to you.
If I’m sharing with my friends, and my friends are on Facebook, why do I care whether Facebook is adding users or not?
Because I think there’s an effect by which, if that audience of people, in your Facebook friends graph, gets large enough, it starts to include all sorts of people.
Right. That’s different. That’s whether I’m adding people to my Facebook friend group, versus the total size of the Facebook.
Right. What matters for you is really who’s in your friend graph, because that’s who you’re publishing to. I notice this for myself. I started adding people I’d met in a business setting, coworkers. You have family, you have friends. Humans, by nature, are very good at maintaining different identities for different groups. But Facebook, by being the largest social network in the history of the world, has basically munched them all together. Now, I know with the algorithmic feed, if I publish something, I’m not sure who’s going to see it. But, potentially, all those people are going to see it. So, I think most humans aren’t sort of like, by nature, good at figuring how to broadcast to everybody they know.
And whether they want to. Right?
Right, right. That makes it hard.
You have a line here … Basically you say that, “The longer you’ve been on, the less likely you are to post.” There’s a hill. Right?
Then there’s an aside, “Not everyone is like this. There are some psychopaths who are comfortable sharing their thoughts, no matter the size of the audience.” I mean, I do notice this when I go on Facebook. I’m not saying my friends are psychopaths. But there is a consistent group of people that are posting a lot. I don’t know how many Facebook friends I have, but I assume it’s hundreds, and I’m only seeing, probably, a very small slice of them on a recurring basis. I guess what I’m describing is what you’re describing. There’s less utility there for me.
Right. There’s a form of, I think, adverse selection, where you get to a point to where the oversharers start to take over, and that’s all that you can interact with in the feed. That makes the other people even more reluctant to share.
Leaving aside the fake news problem — and I guess you can argue that it’s related, and the regulation problem, which is also related — you’re saying that Facebook’s wall is Facebook’s size. Right? They’ve grown so big that the more they grow, the more they’re depressing use of other existing users.
Yeah. I think the way to maybe address that or think about it is think about how to break up that massive single surface area into smaller surfaces, where we are more accustomed as humans.
I never heard anyone using “surface area” as a description until we had the Facebook execs on a Code Media for the first time. And people said, “Oh, no, no, no. All the product people now use surface area.” When did that become a thing?
You know, I don’t know. We were talking about asymptotes earlier. I just think there’s something … You have a lot of engineers and people who like using math terms. If I use “asymptote,” it sounds more scientific.
“Surface area” sounds great. It sounds super cool, like you’re going to the moon. So break that up. Instagram, again, doesn’t seem to have any problems at all now. As someone who uses it a bit, I’m seeing a lot more ads there.
I can sort of tolerate it. It’s a free service. There’s cool stuff. It seems like they’ve got a handle on it.
Yeah. I think there’s something about the visual medium, first of all, that … I note this, is that Facebook has all these media types. They have status updates, they have some videos, they have photos. They have a lot of different formats. But Instagram was sort of visual from the beginning. I think that was a huge advantage for them. It kept some of the more overt toxicity out of things.
It’s harder to upset people with a single image than it would be with …
Yeah, it’s a little bit harder to troll visually. I mean, it’s certainly possible.
But I just think just sort of the offensive status update from your uncle who thinks Obama was a Muslim or something, I usually get less of that. I think that visual mediums, in general, are just more appealing than texts, overall. Maybe you know this more than anyone. The average person still watches like four to six hours of TV a day. Maybe not all the super-productive professionals that you and I know.
I think they’re still cramming in that TV time.
Yeah, maybe they are. That’s probably the most addictive medium we’ve ever created.
What is the Instagram limit that they’re not aware of?
Well, that one’s a tough one. I’m not completely sure …
You think they’re good. They’re excluded from this.
Well, certainly they’re still growing, and they have a lot of headroom for growth still. I do think that over time they could probably spend more time on more overt narrative and narrative structures. Because it’s kind of like an unstructured bunch of images and videos right now. You start to see some people doing that with stories, where they’re doing like linked storytelling, in a way. But the app isn’t really optimized for that.
At some point I just think, with all things — and this goes for Facebook as well — one of the things is that the News Feed structure, the vertical scrolling interface itself, is good for some things and not other things. Right? It’s always tempting because you’re one thumb swipe away from something new. So sitting on one piece of content for a long time, spending time with that, it’s going to be hard.
But it seems like, if I’m Instagram, what I would say is, “That’s great. I know that, and other people can create other products. I have a version of a product that people love, and I don’t want to try to create other verticals and other story types because this thing works really well.”
Right. I think that. But if you look at how flexible Instagram’s been across their history, and just sort of changing things up, I imagine they do have … I have this theory that all tech companies eventually will want all things, when it comes to attention. So I think they’re a company that would be willing to say, “Hey, you know what, we should build a separate app that’s more optimized for whatever.” Messaging. More optimized for long-form narratives and things like that. I think everybody still looks at that four to six hours that people watch TV, if you’re in tech, and think, “Wow, I would love me something of that time.”
Yeah, let’s talk about that in a minute. Let’s take a quick commercial/sponsor break. We’ll be right back with Eugene.
Back here with Eugene Wei, who just provided a nice segue to this part of the conversation. Want to talk about Amazon, which is where you started your career. We’re still talking about this viral essay you did, “The Invisible Asymptote.” This idea of where a company’s limits are.
You started out in that essay talking about Amazon, figuring out that they were going to bump into this limit where people wanted free shipping, and if you didn’t give them free shipping they were going to shop less. They solved that. Amazon’s cranking. A minute ago you said it seems like most tech companies sort of want to own some of that TV time. Is that sort of a crucial issue for Amazon, moving into that time we’re spending watching television, watching video? I can’t tell from the outside whether it is something that they are very passionate about and very serious about, or still haven’t fully committed to.
Well, I’m like you, on the outside there.
Yeah, that’s all right.
So not sure. But certainly …
But you did spend some time in this essay talking about where you think their limits are.
Which I guess is a different question.
Yeah. That’s a different question. I do, certainly, if you look at the hiring and spending at Amazon, it seems like they’re serious about that space. Unlike, maybe, some other companies, it’s not quite as clear to me, maybe, what their strategic approach to that space is. But, they’re an interesting company in that right now Prime Video is bundled into the Prime subscription, which make it a little different than the other competitors in the streaming video space.
At one end you’ve got Netflix, they’re a video company, right? That is what they do. They’re clearly going to spend as much as they can as long as the market allows them to keep spending as much as they can. The more content they build and create, the more users they can subscribe.
I think at the other end is companies like Facebook and Google, which are in the media business, but don’t want to say they’re in the media business. But they still really want to be a platform. And they’re willing to invest some amount of money in media, and they’re not quite sure.
In the middle, maybe, is Apple, which has avoided it but is now getting into it. For a long time, and the Netflix people would say this, “Well, our obvious competition is going to be Amazon. They’re going to be matching us. They’re going to be competing really hard with us.” But they’re not spending the same amount on content as Netflix is.
They certainly have the resources to do it. So when I see them spending billions, but not the number of billions that Netflix is spending, I’m trying to figure out what they’re thinking about.
Yeah. Well, if you look at it as an add-on to Prime right now … There’s one way you could look at all their spend, which is just as a customer acquisition vehicle for Prime.
Which is what Jeff Bezos says out loud, right?
More people watch “Transparent,” more people buy shoes from us, or whatever it is.
Right, exactly. That’s one way to look at it. I think they’re sort of spending at a level where … I would think that that wouldn’t be the only reason you would do it. At some point, every one of the streaming services has to think about retention. You’re trying to minimize churn. You’re continuing to spend on video at Netflix to keep people from churning out. So you have to analyze content both from a customer acquisition standpoint and from a retention standpoint.
But either way, that view is, we’re creating this media for you. We’re buying it or creating it, so you either come work with us or stay with us. Either way, it’s the same thing. But it’s to support our other business, which is a real business, and this is kind of our loss leader.
I think that is one of the challenges for Amazon is that because the video is part of Prime … I think one of the advantages for Netflix, for example, is that it’s so pure a strategy, like that piece in New York Magazine this week about Netflix makes clear. There’s some real advantages when all you are is a video company, and all your spending goes to support your single subscription service.
For example, in Prime we have video, you’ve got music, you’ve got … There’s tons of things in Prime that I don’t even know about, probably. It gets a little bit harder, probably, to measure what’s contributing what. How much is free shipping contributing to your Prime subscription versus video and all of that? Can their video spend actually be measured fairly, if it was just the standalone business? How can you assign that credit? I think that’s certainly one of the challenges.
You were at Amazon earlier, like we talked about. What year did you get there? ’97?
Yeah, I was there ’97 to ’04.
You said that early on they were in books, but you were right away talking about launches into music and video and other stuff. Is what they’re doing now what you imagined, or what you guys discussed them doing, in 1997?
It’s certainly much broader than I could’ve imagined. I think, when I got there, I right away was working with a bunch of MBAs, like Andrew Jassy and Jason Kilar, and all these folks on different business verticals. We were going to go into music and clearly into video and software and a bunch of things. But, the things like Web Services and even, I guess, the Prime Video, and a lot of this would’ve been hard to imagine.
So when you talked about video and music, the idea was, “Well, other people will create music and we’ll sell it to them. Other people will create TV shows, we’ll sell it to them.” Which they obviously do.
But this is something you hadn’t contemplated back then.
Yeah. I guess the only hint we would’ve had on that back in the day was just that when Jeff wanted our mission to be the world’s most customer-centric company. That’s such a broad, broad mission that it can encompass almost anything. Just like the fact that our name was Amazon. It wasn’t like OnlineBooks.com, or something like that. There was always that possibility that our business would be broader. You’d certainly be in long-term planning meetings where crazy ideas would be thrown around. But to see them successful on so many different fronts is probably something that I don’t know that any entrepreneur, no matter how large you think, can imagine happening.
You mentioned the folks you worked with, including Jason Kilar. But you didn’t call him Jason Kilar. Have I been pronouncing his name wrong for a decade?
Okay, good. So he was at Amazon, you were at Amazon. He went to run Hulu shortly after it launched, back when it was originally sort of conceived of as the big TV guys’ response to YouTube and Apple, and it was a place to go watch their TV shows the day after they aired. What did you learn … That’s a really interesting company. It’s moved around. The networks have tried to figure out if they’re going to sell it, if they’re going to keep it. It’s still sort of up in the air, as Disney and Fox and Comcast are all jockeying around.
But what did you learn, working at Hulu in the early days, that is worth thinking about as you’re watching HBO and Netflix and Disney/Comcast/Fox, some derivation of that, all fight for streaming video?
Yeah. This probably relates, also, to the streaming music space as well. In the early days of Hulu, a lot of what I heard from the media companies was, “Hey, piracy is becoming a problem for us, so we see sort of Hulu as addressing the piracy problem.”
Because we’ll give it to you for free. We’ll put ads there. Then you have no reason not to watch it from us, and we’ll still make some money.
Exactly, exactly. Sort of, you know, how the music industry, in the beginning, may have reacted to Napster and things like that. But, as you’ve seen with music, it took a while for a number of factors to happen. The iPhone got created. Everyone was connected all the time, so now you could listen to music anywhere. All these conditions come together. Then one day, boom, the streaming subscription service thing happens. I always had a broader thought early …
Almost a couple decades for that business to really work at scale.
Partly because the reasons you described, partly because the industry itself didn’t want to embrace that idea. They still wanted to sell CDs for a long time. Then they said, “All right, we know CDs are going away, but we’re going to sell you downloads. That’s still the business we know.” And the idea of selling you access was something … They were very worried about cannibalizing those existing businesses.
Exactly. I think that when there are a variety of factors that converge at the same time … If they don’t all converge at the same time, you just think, “Oh, okay, the time isn’t right for a certain business.” But then, one day, all the factors converge and it happens very quickly.
I think, from my early days at Hulu, I would say you can play defense in the media space for too long and think that you’re safe. Then by the time you realize you have to play offense, it’s often too late. Right? Your competitors have sprinted past you. You know, I look at it coming at HBO. If I were at HBO and I saw the first few Netflix original shows, I probably would’ve been very tempted to just laugh them off as pretenders. I don’t even remember what the first couple of …
The first one had the guy from Bruce Springsteen’s band.
Oh, right, from “The Sopranos.”
Yes. Set in Norway.
“Lillehammer,” the show.
It’s kind of a punchline now. But then maybe they did something else after that. And then they quickly said, “All right, we’re gonna do this for real.” We’re gonna make an HBO show with HBO talent. David Fincher, Kevin Spacey. Later regretted Kevin Spacey. But we’re going to spend $100,000,000. We’re going to tell everyone we’re spending $100,000,000. So right away, that was an HBO-level show.
Right, right. And if you look at HBO, I think, you know, if you scope your ambition to be kind of based on what the business has always been, it’s easy to scope it too small. So if I were HBO, I might’ve said, “Look, I want to win Sunday night. Because Sunday night is sort of prestige TV night now. I want to win all the Emmys and I want to be seen as that brand.” You know, HBO stands for that. Like, it’s not TV, it’s high quality.
But I think if you look at Netflix, they came along and they’re like, “Yeah, we’d like to win some Emmys. That would be great. It would be great to win Sunday night. But actually we’re doing something very different. We’re trying to win all the nights for all the people. And not just all of the nights, we’re trying to win the afternoons and anytime.”
Well, they started off very consciously sort of going after an HBO audience, going after prestige stuff.
Going after you and me as consumers and people who would write about this stuff in the New York Times as consumers. And then over the last few years they said, “Well, now we’re going to give everyone something to watch,” kids, people who like Ashton Kutcher, people who voted for Donald Trump. Just flood it.
Yeah. And what they have now is an economies-of-scale advantage, which I don’t think traditional media companies necessarily thought there’s an economies-of-scale advantage. Because in the old days, you sort of thought, okay, there’s only so many nights in the week and in the programming guide when adults are watching television, so we’re a finite inventory.
And Netflix came along and was like, “You know what? We can probably broaden that,” and if we get an economies-of-scale advantage, meaning they have more subscribers than anyone, now every additional dollar we spend is spread across the largest possible subscriber base. And now they’re the heavyweight in the room, right? They have 125 million subscribers.
A big concern at HBO is we don’t have enough resources. It used to be we don’t take Netflix seriously. And then you guys are giving Netflix too much credit and now we need more resources to compete with them. And that’s gonna be the issue for the AT&T deal. Will AT&T give them this?
At Amazon, I always knew from the very beginning our whole business was an economies-of-scale business. We had to get to scale and now Netflix is at that. So for them, once you have an economies-of-scale advantage over your competitors, the No. 1 thing you should do is actually just leverage your scale, right? So it makes sense for Netflix to keep spending to a degree that none of its competitors can because that’s its core asset. So I always likened it to it, you know, if you’re a sumo wrestler and you’re the heavyweight in the room, you don’t want to be trading punches with the karate fighter, you just wanna like sit on top of them.
The one thing Netflix and the other internet companies that are playing around in content haven’t done, Amazon is doing a little bit, is gone after the big-ticket sports stuff. Gone after an NFL night or season.
The stuff they’ve been doing is all playing at the margins, right? You can watch Thursday night NFL games on Amazon, but it’s really broadcast on CBS or NBC. They’re not spending real money on it. A lot of people, I think including the leagues, assume these guys are going to come and they’re gonna spend a ton of money and that’s what’s going to keep our rights going. Do you think that’s going to happen?
I think at some point. I think the prices are still a little high for most tech companies.
Even though they can easily afford it, right? They have more resources than the networks that are paying for them.
Right. I actually think … Oh, one thing on this front is that if you look at a lot of tech CEOs, many of them don’t watch much sports themselves.
Culturally. I remember when there was a discussion whether YouTube was going to go after these NFL rights and someone explained to me that their business/content guys had to explain to Sergey Brin and Larry Page basically what pro football was and how it worked and they were all confused by it and confused about how the rights were spread out and this was not that long ago.
Yeah. Yeah. I don’t know how many meetings I’ve been in at different tech companies where I’ll have to pause and say, “You know who LeBron James is, right?” Or, you know, Beyonce?
Did they not? Or?
I think the way …
But they also, they’re smart people, they understand that this is a popular thing in the end.
Yeah. Ultimately, I think the way to frame it for a lot of technologists, because there is that stereotype that tech people think of all content as a commodity, they’re just like dismissive of it.
There’s a lot of truth to it, right?
Yeah. Yeah. There’s a lot of truth to it. So the way I try to explain these things to some of the tech people to help them better appreciate the value is to say that content can be like a social network. You know, “Star Wars” is a social network in its structure, right? I was almost shocked that my nephews …
What does that mean?
What that means is that all the people who watch “Star Wars” and are fans of “Star Wars,” one of the reasons we watch is that we can talk about it with other people. It’s a shared narrative for us. And so then you …
Watching at home, but part of the value is sharing it with somebody in real life or on the internet.
Yeah. It’s cultural capital for conversation and for binding us together. And so a network effect is classically, you know, every additional person adds more value to that. So every additional person in the world that watches “Star Wars” makes “Star Wars” more valuable to me because that’s one more person that we can, you know, make a joke about Yoda or something and have that.
And so if you look at certain types of content, sports is one of them. They have that quality of being like a social network. Like if you broke down the attributes of the content and you looked at all the fans of the Philadelphia Eagles or all the fans of the NBA, it is like a shared religion in a way. And so the same way that Zuckerberg at Facebook probably looked at WhatsApp and looked at Instagram and said, “Wow, people are going to think I’m paying too much for this, but I understand the value of social networks, network effects by reality. And I think I’m underpaying for it.” I think a lot of content gets undervalued, in a way.
So this is what the, again, the sports owners think is going to happen, whether they think it’s because the Zuckerbergs of the world are super smart and see the value of the product, or they’re suckers. Either way, they think that they’re going to show up with a big checkbook.
And the reason I asked you this is one of the essays that really struck me that you wrote before, I can’t remember what it’s called, but you’re talking about Will Smith and the NFL and turtles in the Galapagos and ESPN. But it ends with you being very down on the value of professional sports as a media property, as someone who is a sports fan but saying, “Look, I think the value of this is really diminishing and regardless of whether the TV guys are paying too much, I just think this as a product, as an entertainment/cultural product, is going down because there’s better stuff available.”
Right, right. Um, I think one, the whole Galapagos analogy was just that the sports league had a monopoly for so long.
Like a lot of other media, newspapers, people who thought they were really successful and in fact they just didn’t have competition.
Right, right. I think they are slow-moving. They may not recognize how competitive the entertainment landscape is getting and I think they get locked into a format rigidity. There’s this concept of nominal rigidity where the price is what the price is because it’s always been that way and it gets locked in. And I think that is how sports leagues think about their product. They’re very used to broadcasting in a particular way to a particular audience.
And I think if you’re a tech company and you’re looking at sports rights, it’s a little bit of a game of, “Wow, they want so much for this, but I think the value is diminishing over time that I can probably afford to wait.”
And by the way, the reason it’s diminishing is in part because of me, because I’m providing people with other things they can do with their time. You’ve got this interesting riff in here, and again, it’s better to read it than to hear me talk about it, but you’re talking about measuring entertainment based on the brittleness of the narrative.
Right. Part of it is this thing is valuable to you if you just read a summary of it — which is the reason why I’ve stopped watching “Westworld” now, I realized enjoyed reading the recaps of it more, I kept falling asleep watching it — and/or if you just show me the highlights of it, is that as valuable?
Which is very interesting to me, because I think of — I’m just riffing now instead of you — but I think of House of Highlights, the Instagram NBA highlight thing, we’ve had the folks from SBNation — sorry, from Bleacher Report — on here before talking about that product. To me that’s a symbol of how smart the NBA is. They’ve allowed that product to thrive. It adds value. It makes me more interested in the NBA. But if I’m reading your essay, the fact that I enjoy House of Highlights so much should make the NBA worried, because the suggestion there is I’m going to be more happy watching House of Highlights that I would watching an NBA game.
Mm-hmm. Yeah. I think the NBA is smart in that, look, there are always going to be some people who watch the full games, and if I’m the NBA, I actually look at it as just multiple ways for people to enjoy my product. But that’s not typically a traditional media company way of thinking.
You know, if you look at the NFL, which is probably the polar opposite and being sort of the most locked into specific formats, not allowing your content to travel on social media just means that you’re not gonna reach certain people. If there’s anything you can learn from a company, like a media company like BuzzFeed, it’s that, hey, all this attention is spent on all these social networks. You kind of have to go where the fish are. I think sports leagues have to recognize that over time, the value of their content is actually slowly eroding and if they don’t get ahead of it, like I said, you know, you can play defense for the longest time. By the time you recognize you have to play offense, it’s probably too late.
We’re jumping around as I knew we would. Two more things I want to make sure we talked about. You were at Oculus for how long?
Almost two years, I guess.
Two years. And your job was to go, create, acquire content for those headsets. Anyone who’s listening to this has a notion that VR had a lot of hype.
People are now soured on that, at least temporarily. One of the arguments is just the equipment was too bulky and too expensive. And everyone who went to the Code Conference earlier this month got a $200 Oculus headset. I’ve got one sitting at home. It was pretty cool. It works, you don’t need a phone for it. At 200 bucks, that’s supposed to be the magic price point. So is this a problem where the product is now cheap enough, works well enough that VR is going to take off again? Or is there a content issue with VR?
I think it’s multiple things, which makes it tricky. I don’t know that it’s one thing. I certainly think the Oculus Go headset, by not requiring a phone and being portable and not having to tether it, solved a number of issues. Now the question is, you have all these factors that are holding VR back. Like it’s a multidimensional asymptote. It’s not one thing. That it’s going to take pushing a lot of these things uphill together before I think it becomes a mainstream …
What are, top of your head, the first couple problems you think about? If we solve price and portability.
Mm-hmm. Resolution of the screens, quality of this sort of like, I guess, storytelling and sort of broad content.
I actually think a lot of the games in VR are amazing already. If you’re a console game player, some of the stuff you can do with a nice VR headset is unbelievable. But for other people who aren’t heavy gamers, I still think there’s probably a gap where you may leave that headset on your coffee table for weeks on end.
It’s funny, we got it last week. We did a “Jurassic Park” thing, which I’d actually seen earlier. Kids were blown away by it. Put it on the coffee table or actually out of reach of the kids. And they haven’t asked to use it again. I haven’t thought to use it again. And then I got a press release saying that Fox Sports was going to make part of the World Cup available in VR. I thought, “That’s stupid.” Anyone who’s put this headset on would not want to watch a game in VR. Maybe two minutes of VR. And it seems to me that’s one of the issues is, there’s a lot of, like, a lack of creativity or just people haven’t figured out how to be creative enough to figure out what content works this way.
Right, right. I actually think there are individual pieces of content that are amazing in VR. But there are no economics right now, right? It’s all subsidized by Facebook and Google and these companies that try to keep this VR train moving forward. And that’s what’s sustaining it. But we don’t have these positive feedback loops yet where you see, wow, more people using it.
Which is unusual for a Google or a Facebook, right? To put a lot of resources into a thing that doesn’t show results at the beginning, it seems like.
Yeah, it is. But I also will say this about VR. I actually do believe it is going to be a transformative platform shift at some point. And so all tech companies who have experienced missing out on a platform shift are a little bit hesitant to miss out again.
So we don’t know what this is going to be, but we think it’s something big. There’s a version of this with scooters right now, right?
Maybe a little more crass, but we think there’s a thing we can afford to experiment on it. It’s not AdWords. It’s not a News Feed where we can see that it works right away. But we’ll keep going.
And then I want to bring this all the way back to Amazon. Ask you one more Amazon question. I mean, I can’t stress this enough, your insight into Amazon, a company that’s famously closed-mouth about what they do, gives very little information, I think is tremendously valuable. You should be, we should charge for this podcast. You should charge for your blog. But you had a great post. They all bleed together, sort of explaining that people still fundamentally don’t understand the company, especially when it comes to them losing money or having very slim margins. Can you tease that out for us? What’s the thing people don’t get about Amazon?
Yeah, and I don’t even know if this is a mainstream thing that people don’t get. Maybe it’s more of a media thing.
Media business-y thing.
Yeah. I still run into people who are tempted to lump Amazon in with like say MoviePass or different things where they’re like …
Yeah. These are just companies that just keep spending and spending.
Yeah. It’s like some sort of sham and you know, if you just look at the unit …
They’re not a real company because they can’t turn a profit and they’ve been at it for years.
Right, right. And that just speaks to maybe some of the financial opacity that’s inherent in looking at income statements and balance sheets and things like that. But you know, look, as a retail business it’s pretty simple to say if I buy something for five cents then I sell it for ten cents, there’s a profit in that. And if I decide, hey, I actually want to do that all over the world and all of that, I may invest ahead of when I realize those profits.
The unit economics of retail may not be like software-level margins, but Amazon makes a profit on most of its retail transactions. And so, when people see Amazon not making a profit in a quarter, they don’t realize that it’s a lot of just forward investments on Amazon’s part. And so it’s very different from a businesslike MoviePass, where I’m not totally certain about the economics, but it feels like …
And Amazon will say this. They’ll say, “We’re taking the money that we made selling you that pencil and we’re putting it into our next distribution center or R&D for our robots or whatever it’s going to be, or we’re going to build a new cloud business.”
They do say it. I think part of it is a natural skepticism that anyone who remembers the first set of dot-com companies where Kozmo.com really was just blowing its money every time you ordered a Ben and Jerry’s, right? They were losing money on every transaction.
And there’s that fear, or like MoviePass, right?
You go see a movie, MoviePass loses money. Seems like a terrible business.
Yeah. You know, like if you take a dollar and sell it for 75 cents, that’s one thing. But that’s not what Amazon’s in the business of doing. And so I think if you look at companies like Uber or if you’re going to look at the scooter companies, if you’re looking even at Netflix, when it finally reaches whatever level of subscribers it reaches, what the important thing is actually like at that level of business, how much do they have to spend and what are the unit economics of their business at that point? And I think that’s just the level of analysis that’s necessary to understand whether hey, this company’s losing money because it’s just a terrible business or they’re losing money because they understand that it’s winner-take-all and you have to get big as fast as possible.
So you went to Amazon basically out of college, right?
No, I had a job, well, I was set to go to law school. I deferred and then I went home and worked in consulting for like a year and a half.
So that’s sort of standard track, right? You go to a fancy school and you go into consulting. Do people still do that now?
Oh gosh, I don’t know. Maybe you really want to do funds or something.
Probably. But then you went to Amazon and I was listening to you on another podcast where you talked about writing a three-page cover letter to get a job you were not qualified for.
I didn’t get that job, by the way.
But you did get a job and we didn’t get to the film school stuff, but we’re at an hour, so we’re gonna end it here on this question. Since we know that college students do listen to this podcast, what advice do you have for someone who’s going to be leaving college in the next year or so and is looking at creating their career and whether they want to work in an Amazon or anything else? How would you advise them to go about thinking about that?
Yeah. I would say the biggest thing is that I, looking back, I had no idea of the number of jobs that were available to me when I was in college. I really did feel like if you came out, it was like law school, med school, you know, go into consulting. Go into finance. There weren’t very many things that were made available to the career center.
So my career, what I’ve always prioritized is, okay, just work on building your skills and knowledge in a lot of different areas. School probably didn’t prepare me to be a product manager at Amazon, but it gave me a framework where I was like, okay, I can learn new things on my own. If I’m disciplined enough, I’ll pick up things.
And so I come from an era where it’s probably still like this, where I was more of a self-made product manager. It wasn’t really even a job when I was at Amazon. And so I think your earnings and your job titles and things, they may lag your actual skills accumulation, but over the long haul they tend to converge.
It’s like a commencement address. Free, bundled at the end of this podcast. If you made it to the one hour and one minute mark, you got a free commencement address from Eugene Wei. Eugene, thank you for coming. I’ve wanted you to come for weeks and now you’re here.
Yes, I’m glad we finally had it did happen.
I’m sorry I can’t pay you for your time.
The La Croix’s enough.
All right, good deal.