a16z Podcast - a16z Podcast: The Topic That's Lasted the Entire History of Computing -- Bundling and Unbundling

Episode Date: August 15, 2014

Jim Barksdale in the run-up to the Netscape IPO told potential investors that you can make money in software in two ways: bundling and unbundling. Benedict Evans and Steven Sinofsky revisit that thesis in the context of a mobile app world -- how Facebook for example, is unbundling itself, while at the same time Baidu is bundling everything together as fast as it can. How and why Barksdale's thesis is very much alive and well in the mobile world. All that, and the proper use of "fissiparousness" in a sentence.

Welcome to the Andrews and Horowitz podcast. I'm Benedict Heavens. I'm here talking today with Steven Sinovsky about the topic that's lasted the entire history of the competing industry, which is bundling and unbundling. I wrote a blog post on Sunday that started out by quoting was it Jim Barksdale, who supposedly during the Netscape IPA presentations, told people that there are only two ways to make money in software bundling and unbundling. But of course it goes back a lot longer than that. And the thing that I was sort of poking away at is that when you look at what's going on in smartphones today, on the one hand, you have this trend towards unbundling features from inside apps and making stand-lane apps. I think Fred Wilson called this a

constellation of apps. And so you have Facebook pulling stuff out of Facebook. You have four-square splitting its app into, you have lots of other people trying to do single-purpose apps. And the motivation for that is a limited screen real estate, which changes the discovery dynamic. And B, that any app is just two taps away. So it's much easier to switch between different apps. And so you see WhatsApp and Instagram unbundling Facebook, and you see Facebook unbundling itself. But you also fundamentally, I think, see apps unbundling the web, because everything used to be just in that one icon, and now it's pulled out into lots of icons. And, you know, Steve, as you were saying, this is kind of a story that goes back to office and it goes back

to mainframes and everything else. The whole thing just goes, I mean, you know, in a sense, it's just, it's about innovation. I mean, you know, you, you have an idea to, to complete some task. It could be a consumer task, a business task, and then you build it. And, and then what you find is that your idea is sort of, you know, related to other ideas. And so you start to add features. And then one day you have like someone else's whole vertical integrated into your product. But it's never as good as the separate one. But it might be good enough to keep your customers from jumping and using another product yeah and so you you and then one day everybody's telling you your product is too big and bloated and complicated so then

you think okay well it's kind of time to split us up into different modules and you know also you know you're competitively you were frozen in time and also somebody else might pull your feature out into another model in fact that's what that what often happens like the feature you did someone says well they're not going far enough because of course to you it was sort of a checkbox the the best An example of this is historically it was always adding email. And so people would add just a little bit of email. And now it's Messenger.

For a while in the middle it was always photos. And so you can look at basically every app that included photos, and they all do a little bit of auto brightness and contrast, a little bit of red eye, and now, of course, everybody does a few filters. But they're never as much as what you could get in a full-fledged photo app. So then those are going to keep existing. and then you're either going to completely bloat your messenger-ish kind of application with more photo editing,

or you're going to lose to the photo editing people, and you're sort of in this quagmire of discoverability, if you will. Yeah, so the thing that struck me was, and I was looking particularly at what Bidu has been doing with maps in China, which is we've got this done out, we've got this narrative that says there's these endless sort of fissipariousness of applications, that things endlessly get kind of split apart and split apart and split apart. like, kind of Trotsky is sex, you know.

Wow, there's a lot of metaphor in that whole sentence there. The point is that all these things kind of split apart and spit apart and spit apart. And then every now and then there's this kind of, it's almost like a kind of gravitational thing that things get kind of sucked back in again, like, you know, the cycle of the universe. And, but the thing that you see in China is that you have these apps that bundled lots of stuff inside. So just as we feel like there's this totally inevitable, totally inherent in the platform, and this is just the way it's going to be.

And then you look at China, and it completely undermines all of those assumptions, because what you have is this very strong portal models, and particularly on mobile, you have, for example, by due building a Maps app, in which you can book restaurants, and you can book cinema tickets, and you can not just book cinema tickets, but choose the seat in the cinema, and you can not just book a restaurant, but, you know, order home delivery from within the app. And all of those things kind of work, it seems to me, because they flow out of the context. That is to say, you know, if you had an app, as it were, might be trying. to do everything, that would be the web and you'd need Google in order to find things. Without Google, you just kind of get lost in the sea of features. But if you have an app that says, okay, it's about location context. We don't need to have 45 buttons, one for restaurants, one for cinemas, one for hotels, one for taxis. We can just wait for somebody to do a search for restaurants. Right.

And then when they do a search for a restaurant, all the restaurant icons appear. And when you tap on the restaurant icon, well, what would make sense to appear there? A booking option. Right. And so it feels like there's this sort of the scope somewhere in the middle for integrations that focus on particular sets of use cases, whether it's location, which is obviously, you know, kind of sort of fairly logical one, or alternatively social. But particularly when we kind of, you know, the difference, I think, between one of the difference is between doing this in the sort of the package software world and doing it in the internet world is that the big challenge for everyone is how you get, content. You know, if you're a web publisher, you're not really sitting and thinking about HTML versus apps, or should it be in Facebook, or should it be in this kind of app, or should

it be in that kind of app. What you really care about is how you get traffic and what different things you need to do in order to get that. And the interesting kind of pendulum that swings back and forth is, okay, do I get traffic by having my own app and having a link from that from within the Facebook app, or from having my own website and getting a link to that from in Facebook? And do I expect the traffic to that website to come from Google search on the smartphone? Or do I have an app? If I have an app, how do I get people to install that app? Are they going to find it inside the app store? If so, what would they have searched for? Or maybe I can get people to install the app by an app within Facebook. So there's just this

kind of enormous mesh of what your options are and where your traffic comes from. Whereas on the web, it was basically, you did a website and then you were Google. You're kind of done. Right. But what's fascinating about this is that it actually opens up a really a really strong front for product managers to think about in terms of design, because you're really dealing with a very clear measure, which is usage of your service. So if you stick to that, it's a really good thing. Like, in fact, like, I could, I would argue that, like, all the places that are downloading the number, you know, measuring the number of downloads on their app and things like that are, that's like a secondary measure. The primary measure is engagement with the

service. And in fact, you could go too far and focus on just getting the downloads, and then you'll be just victims of, you know, download shrapnel where, like, you just got a lot of downloads and no one is using your product for a whole variety of reasons. And, but what you have to keep in mind is that that you can continue to play out that metric as a designer, and then you could start to sort of force awkward behavior. Like, you can drive people to try to do things in the app that are sort of for engagement as opposed to solving their problem. And I think in particular because of the way that the apps have evolved in China, they're sort of not at the monetization phase just yet. They're really at just trying to drive engagement, and they're

solving the problems that people have. And I think we have a history of websites that start off by engagement, and then they try to over-monetize that engagement, and then the design choices sort of lose out to the business case, and it's sort of a thing to really watch. And where that will manifest itself in apps is if, you know, you just get prompted to often to download a companion app or you will review you an app well like the the feature breakout just seems kind of artificial like it's not really driven by screen real estate or capabilities or something it seems like you're just being launched into another place because they really wanted to break up the app into more atomic units like if your

favorite social network all of a sudden breaks out photos as a separate app because they want to be more like Instagram which seems broken out the problem is is that there all these things that they're going to end up duplicating. And then you're just, you're not even going to be sure where to start. And I think for me, the whole thing that's fascinated about China is they know where to start. They're starting with the place that I has my credit card and my profile, and I just flow from there and solving problems. Yeah, you're kind of, this is the, you're kind of turning the,

um, you're turning the aggregation upside down in a sense. Because on, you know, in the US, every separate app has a map built in. and if it's a location thing, or it has messaging building, if it's a sort of a social thing. So, you know, Yelp has maps in it, and TripAdvisor has maps in it, and, you know, Eater has maps in it, and all these different things will have maps in it. And Uber has maps in it and Lyfts have maps in it. Whereas in China, the Maps app has taxis and restaurants. Right, right.

And the messaging app has taxis and restaurants in it. And indeed, if you go into Baidu and you want a taxi, it will show you all the cameras. in your city from the five different taxi companies that are in that city, and then it will bid for all of them, and whoever gives the best response gets the thing first. And then you also, of course, have a tapped call button so you can send a voice message to the driver to tell him where he is from inside the taxi module within the Maps app. So this is like the old, this is basically this very old story in philosophy or astrology, astronomy, that a sort of a Hindu guru or Eastern guru is saying that the world is flat and it sits on the back end of a giant turtle.

And somebody says, well, what does a turtle sit on? And he says, another turtle. Exactly. Exactly. And you feel you're looking at some of these things, it's like, okay, how many different layers of aggregation have we got within this thing? Okay, so I've got Siri or I've got Google now or I've got Windows Live tiles. And then I've got the OS layer of the all the different, the grid of icons, and then I've got widgets, and then I've got maps, and I've got the social application. And then within the maps application, I've got these different aggregators of restaurants or taxis that are all integrated into that. And it's like, okay, I've got, how many different ways do I have to achieve the same thing? You could get, you could get all deeply philosophical, but there's a more practical, they're not

a more practical, but there's a practical way that this is really going to play out, which is I always like to fall back on the classic psychology paper, the classic, the magical number seven plus or minus two. For those outside the U.S., there's a lot of history and how that is tied to why phone numbers in the United States are seven digits for example and of course now you don't have to remember phone numbers at all so sort of but or websites yeah but the idea is is that you know for better or worse like most everybody is going to just use about seven apps on their phone yep and and it the same holds for PCs I mean I remember you know we we used we did what the

Windows 7 launch we talked about that there were over 800,000 apps that got used just during the beta process. Of course, no one person used 800,000 and there were several million people in the beta, and it turns out that most people use, you know, a single digit number of apps. And you get out to all this complicated stuff, you know, like is Facebook one app or two apps or three or four apps right now? Well, it really is the Facebook app is the one that most people will launch. A very large number of people might launch the Instagram app as a standalone one. But for the most part, it's the thing that the seven plus or minus two is the action on the home screen that drives that. And that's a good thing because it means that, you know, an app, apps like the Maps in China or Commerce in China,

you think about launching that one app. And it's so long as there good scenarios and user flow, it works fine. And it doesn't really matter how many other apps you use. It doesn't matter how many things are integrated into it as long as they play naturally out of the play. Exactly. And the problem is really if you start to litter the home screen or think that the person is just going to magically remember 10 different things for you, it's really not going to play out to your advantage at all. So this was, I mean, I think, you know, there's a story that in the Wall Street Journal, I think there was some survey that most people have got, use an average of 25 apps ever. And again, probably seven or eight at a time. And I wonder how much that aligns to how many websites people use. So it's probably almost exactly. the same. And then what you have on the back end of that. So we had this data from Flurry and Comscore earlier this year that, you know, as it might be 80% of time spent on a smartphone

is within an app. And the remaining 20% is within the web. And so say that 80% is within those two dozen apps. Then you go to the web and you say, okay, probably 80% of the time was within one to two dozen websites and the rest of the web. And so what happens is that Google picks up that that 20%, which is the other 790,000, 995 apps, 980 apps or websites or whatever it is. So Google picks up that vast long tail of, and it probably, in fact, on the desktop internet, Google also picks up the 20 things that you use every day because that's how you get to them. You just go into Google and you search for Facebook or you go into Google, of course you're going to Google and you search for Google as well.

Right. So you've got this kind of question of, okay, how many different services are you using them, are you using? How often does that, does a mix within that changed? How do you become one of those or how does one of those fall out? And what is the other, as it might be, 500 things that you might use in the course of a month or two months, and how do you get to those? I mean, the quote that I used in my blog post is this famous line from Lenin that the only question that matters in politics is who, whom, who has the power and over whom do they exert it? Right. And you could say the only thing that really matters in the internet for any kind of publisher or developer is, okay, who has the traffic, who has the distribution, who has the users, and who do they give it to?

And on the desktop internet, Google had them all and Facebook had them all. And on the mobile internet, it's just kind of a lot more complex and kind of unsettled and confused. Well, it's a little bit like that. I mean, I think that, you know, if you just take the assertion that, like, it's going to be a single digit number of things, Then it's just, it's a matter of counting and deciding how to divide up the pie chart in terms of, I'd like to think of it in terms of time more than anything, because look, you're going to use, you know, probably the messaging app, you're going to use the mail app. You know, a few of us might still use the phone app, like, for making voice calls. Yeah, Grandpa. Then there's a browser, which you're going to end up going to from any of the other apps that you use.

Now, here's an interesting one. If you only use apps where the browser is embedded, how does the OS or measurement systems even count that as a number? app or not an app like if you you know you're in Twitter as an app yeah and then you look click on a link and you don't click on open in Safari you click on the the link and it just opens it up in place you know you know a little bit of a weird penalty box but which you know you have to deal with well it's as much the death it's not so much the death of the web it's the death of the URL right right and and so to me like you could get all tangled up about like what it is

your measuring but just keep in mind that for the most part people are gonna tap on something in single-digit ways. And then it's really important to remember that the tail is essentially infinitely long, like mathematically speaking, that there are very few apps in either of the major app stores that have never been used. And even the ones that have used by a very small number of people, there are still people who use them a lot. Like I was talking to somebody who built an app for calculating some specific kind of time

in photography. There's probably a couple of people that use that for their workflow. And that's what makes the App Store so valuable and why you shouldn't sort of lose faith if you're not the app used by, you know, 30 hours a week by somebody. Because even if it's seven plus or minus two, there's a hundred percent certainty that you could be in a conference room with ten people, and each of you will have one app that the other nine people don't use. So this is, this reminds me, I mean, the thing that I, that society have, I was going to do. to say unkind, perhaps just a realistic thing to say, which is that when developers say I made this amazing app and no one confided on the app store, that's a bit like saying, well, I wrote this book and it's not on the homepage of Amazon.

They've got a million scoos on the home page, in the app stores, and Amazon sells, whatever it is, two or three million books. And, you know, you reach a point where the internet is everything. Anybody on earth can make an app, anybody on earth can make a web page, anybody on earth can write a book. And so that discovery problem is just sort of inherent in the sort of, system and that fact that most people are only going to use 5, 2, 7, 10, 15, 20 is also just inherent in human nature.

I mean, this is the point that our colleague Scott Cooper wrote this blog post about how venture capital is changing, that on the one hand, the cost of creating anything has collapsed by probably two orders of magnitude in the last 15 years. But on the other hand, the cost of you going out and getting five or six billion consumers hasn't suddenly collapsed in the same way. You know, yeah, your marginal cost has collapsed, but if you want to go out and get a billion people to use your product you either need to get spectacularly lucky or you need to spend a lot of money or just realize that you can build a business out

of having a half a million people use you a lot yes and and that's part of why like this you know I got to be on the the top of the you know the productivity category in the app store something is sort of a bit of a crazy dream now the flip side of that is it's also the same problem if you're already established and you have a bunch of customers there's this gravity that pulls all of your new R&D to be included in the app that you already have because you don't want to have to go out and distract your marketing and sales efforts from the big thing that you're

trying to get everybody to use which is why like in a sense in a way to wrap up like for me at least you know is the bundling unbundling you know is it a pendulum well it's not really a pendulum there are just two approaches and really they're both good unless one of them isn't good for you well that's the thing there's kind of a trade-off between application discovery and feature discovery that if you've got a simple app, if you've got an application with all of your stuff in it, it's probably easier for you to get people to install that app, but then they're never going to find all of your features.

Right, exactly. Whereas if you break stuff out into multiple apps, then the feature set is really easy to discover, but you've got to get people, to persuade people to install those apps, which is why the sort of, you know, that's the kind of the Facebook approach versus the BIDU approach, you know, if you stuff everything into the maths application and you do it right, then you don't have a problem getting people to install Bidu maps, but then you've got a real challenge around how do we make... get people to know what the hell is they can do within the app.

Exactly. Whereas you pull people out, then you don't have a problem of understanding what to do in the app, but you've got a problem getting people to install the thing in the first place. And when we post this one, just wrap up, I have a post that I did that talked about how to do this from a product management perspective and give some pros and cons of the different approaches for discoverability and usability. So thanks a lot. Great.

Thank you.