00:00hello and welcome to the a 16z podcast
00:04this is Benedict Evans I'm here today
00:06with Steve Sandusky we're going to spend
00:08a little bit more time thinking about
00:10what Apple announced yesterday at WWDC
00:13in the context of some of the challenges
00:15that our portfolio companies face around
00:17development and particularly
00:18cross-platform development so Steve I
00:21was at WWDC yesterday I kind of
00:24experienced a moment where I four
00:25started talking about a new programming
00:26language and half of them went
00:27completely quiet and the other half went
00:29wow that's really cool and I was in the
00:31house it was going completely quiet
00:32having no idea what these people are
00:33talking about and but coming away from
00:38one of the things that you and I were
00:40talking about is that Apple is doing
00:42quite a lot of stuff in one way and
00:44Google is doing quite a lot of stuff in
00:46another that doesn't really have an
00:47analogue on the other platform so if you
00:49were to look at you know Windows and Mac
00:51you have a file picker you have Windows
00:53you have scroll you have you have a
00:56bunch of tools that basically have
00:58direct analogs on the other platform
00:59where it says Apple does more on the one
01:02hand more and more stuff it's about the
01:03hardware and Google does more and more
01:05stuff that's about the cloud and Apple
01:07starts doing its own SDKs that address
01:09the harder in particular ways and does
01:11ibeacon and so on it feels like you
01:14might be building an app that takes
01:16advantage of stuff that does that Apple
01:18gives you that actually does exist on
01:20Android or vice versa so you can't
01:21really do an app that does the same
01:22thing on those what do you think about
01:24that well I mean I think that first good
01:27morning everybody but I think that
01:29that's that's actually sort of this core
01:31this core challenge you know at the
01:33heart the goal of developing a platform
01:35you know has the side effect whether you
01:38think it's intended or otherwise of
01:40making cross-platform development harder
01:42I mean the developers don't come into
01:44work every day thinking how can we
01:46advantage our platform or not they
01:48actually come to work to develop new
01:50scenarios new capabilities and what's
01:53fascinating if you just look at the iOS
01:55and Android with you you know which is
01:56only two of the many platforms including
01:58web or Windows Phone or Chrome that that
02:01people are targeting you know they just
02:03have very different perspectives on on
02:05the architecture or on where to solve
02:07things you have Google very very focused
02:09on cloud services and
02:10because of the nature of hardware you
02:12know they focus on that seam between
02:14hardware and software and so that's just
02:17gonna lead to a whole bunch of different
02:19features and so then you sit down to
02:21write an app and and you're you're sort
02:23of like right away asking yourself these
02:25questions about what to use where now
02:28historically it's really super
02:30interesting because cross-platform sort
02:31of follows a very tried-and-true well
02:34understood pattern which is early in the
02:36evolution of sort of what I would call
02:38like a meta platform like the new thing
02:41the new thing whether it's the mainframe
02:43or the wearable or the mobile phone or
02:46the PC they're they're all the platforms
02:48that are competing to win tend to look
02:52kind of the same so early in the PC era
02:55most of the PC ish platforms this is
02:57does I don't mean PC like Doss I mean PC
02:59like thing with a microprocessor yeah
03:01they all had a box in the keyboard and a
03:02command line and a command line and the
03:05the api's were about drawing on a screen
03:08and black and ye and reading writing
03:10into a disk and so they all have those
03:13and all the vendors are trying to make
03:15those things and then one day they all
03:17sort of make them and then they start
03:19going well we have to do stuff that we
03:21want to do and then they look to
03:22differentiate their products and then
03:24they look to you know what's their view
03:26of the future and then pretty quickly
03:28they all just start diverging and and so
03:31then everybody was really comfortable
03:32like if you wanted to write a thing that
03:34worked across all the early pcs
03:35abstracting out like basic graphics
03:37they're only like eight api's to begin
03:39with and then file systems there were
03:41another eight and so it was pretty easy
03:42to come up with your own sort of
03:44abstraction and then all of a sudden
03:46like the graphic system turns into like
03:49a bitmap raster graphic system and then
03:51this whole platform thing starts all
03:53over again every basic okay if we just
03:54use set pixel then we can be
03:56cross-platform across all the the raster
03:59systems and then all of a sudden
04:01everybody starts doing their own visuals
04:02their own controls their own metaphors
04:05for for windowing and then that kind of
04:07breaks down and and then like really big
04:10abstract concepts happen like networking
04:12or cross application sharing and and you
04:15know that that all of a sudden it breaks
04:16down and we saw the same thing with the
04:18web you know like the web had hopes of
04:21being cross-platform for
04:22everybody and you looked at HTML and you
04:24said oh cool I have to just write a
04:25rendering engine for these 40 HTML tags
04:28and that worked and then all of a sudden
04:30people started adding other tags I got
04:32harder and harder and harder and then
04:33there were compatibility tests and then
04:34the committee got together and said
04:35we're gonna do html5 but then right when
04:38that happened the vendors started sort
04:40of saying yeah we have other ideas like
04:42Google had a very different idea because
04:44they were trying to build your entire
04:45computing experience into the browser
04:47not just the browser part of it and and
04:50so this pattern just keeps happening I
04:53mean I think you can see that that
04:54inflection point very clearly now with
04:56with with smartphones because it feels
04:58like what happened is the original
05:00vision of 7 years ago has got built out
05:02and all the stuff that kind of should
05:03have been in that vision like you know
05:04notifications and multitasking and so on
05:06has got done and it's all there and it's
05:08all been kind of iterated through three
05:10or four versions and now you're kind of
05:12getting to the point we think yes okay
05:14so we've got that platform and we've
05:15done like the stuff that should have
05:17been there but what do we actually think
05:19about what this should look like in it
05:21in another five years time I think the
05:23thing that was very striking about WWDC
05:25is this his whole mean that Apple
05:27doesn't get the cloud and all the stuff
05:29that they were doing was about the cloud
05:30you know all the new features were about
05:32the cloud but it was the cloud as a sort
05:33of invisible bit of plumbing that drives
05:36a rich native app where it's of course
05:38for Google it's exactly the other way
05:40around and say that the thing I said on
05:41Twitter was it for Google all hardware
05:44is just kind of dumb glass that links
05:46back to their cloud whereas for Apple
05:47the cloud is kind of dumb storage that
05:49supports their platform but the
05:51consequence of that if you are very very
05:52different philosophies about what it is
05:53that they're not supposed to be doing or
05:55how it's supposed to go about connecting
05:56to things or using resources right
05:58absolutely and then of course people
05:59come along and they they think like you
06:01know we talked to a portfolio companies
06:03they think about where they are today
06:04and then they look at all the work
06:06that's being done across the to and they
06:07go ok well we can find the 90% that is
06:10ok but it's actually the last 10% that
06:13makes the difference in a quality app
06:15experience on those platforms or a
06:17quality you know back-end and cloud
06:19experience and so it's it's actually not
06:21so straightforward to even carve out
06:23like what's common anymore because
06:25what's common is essentially what the
06:27market is already defined as
06:28commoditized you know like your basics
06:30of making calls and sending messages or
06:33of an app or all sort of commoditized
06:35and so as a as a start-up or as a new
06:38app what you have to do is you have to
06:39really really focus on what your
06:41differentiation is gonna be I mean like
06:43I take something like really as as it
06:45could be either as thin or less relevant
06:48as you want it to be or it could be a
06:49huge thing like homekit so it you know
06:51first there's nothing like it on the
06:53Android platform so right away if you
06:55make a device like light switches or
06:57whatever and you're looking you know you
06:58probably already think I'm gonna build
07:00an app and it's gonna have a big you
07:01know on/off toggle and I'm gonna have
07:03rooms and a layout of my house and all
07:05this other stuff and then you know that
07:07looks great on Android and it looks
07:08great on iOS right this minute but then
07:11on iOS you can say to Siri hey I'm going
07:13right or you just there's a whole app
07:17infrastructure for doing that which then
07:19takes over some of your UI so if you're
07:21an established player you're just gonna
07:23stiff-arm that whole thing and hope it
07:24goes away but while you're doing that
07:27all of these new players are thinking
07:29whoo this is such a relief because now
07:31we don't have to hire a whole software
07:32team we could go and make Apple
07:34certified devices we can build on the
07:36retail presence that Apple has to drive
07:38the devices will be in their stores will
07:40be online will be in their commercials
07:42like there are actual tangible benefits
07:45to building on the platform like to you
07:47as a company and so you're sort of
07:50squeezed if you think you're just going
07:52to ignore it especially if Apple goes
07:53and continues to double down and does
07:55more work and makes the experience
07:57better integrates it with Siri
07:58integrates it at the Mac and does like a
08:00bigger cooler dashboard who knows what
08:02they're gonna do yeah but but and that's
08:05just one of the things I mean you
08:06mentioned photos so now there's a way to
08:08integrate with photos in an even deeper
08:10way so you know normal developers go oh
08:13no big deal I'm just gonna build my own
08:14photo picker I'll ask for permissions to
08:16the library and I'll do my own
08:17thumbnails and enumerate the photos
08:19well except now your app doesn't look
08:22like Android or iOS and you could maybe
08:24try to copy them a little but they're
08:27moving targets so I remember using Lotus
08:29Notes in a miserable period of my life
08:32and the I mean this was a period when it
08:37was kind of a poster boy for kind of
08:38questionable you are choices and one of
08:40the big drivers was you had the same
08:42Lotus Notes UI on every device you might
08:45own yeah and the result was
08:47didn't look like that they didn't look
08:48Viton anything but be just because the
08:52developer knows that it looks the same
08:54on every device the users don't know it
08:55looks the same on every device yeah well
08:57that's the whole thing I mean
08:58cross-platform has always been this like
09:00dream of programmers an empty promise
09:02from our industry you know in a
09:04nightmare for customers because even
09:06even today like little things that you
09:08noticed like you know the fact that the
09:10on iOS the the really marquee social
09:13apps are all just better than they are
09:15on Android things like copy paste of
09:17images and sharing and integration with
09:19contacts all are much smoother and more
09:21seamless even though Android is more
09:24extensible and open by some accounts and
09:27it's because that's where the focus has
09:28been on building a really great job and
09:31you know if you look back historically
09:32like Lotus Notes was just one example I
09:35actually my own first
09:36my very first coding work at Microsoft
09:38back in the 80s basically was building
09:41cross-platform and and you know people
09:44have actually recently told me that some
09:45of that code still lives in the iPad OS
09:48iOS apps for from Microsoft because it's
09:51this Mac and why did we do that library
09:53well because at the time going back to
09:55your original thesis you know what what
09:58differentiated Windows from the Mac was
10:00kind of not very much like Windows were
10:02the sort of the same and and what was
10:04weird is the feedback from Mac customers
10:07back during like that that sort of mid
10:1090s era was that the the Microsoft Mac
10:12apps were starting to diverge too much
10:14from the the Mac platform in fact even
10:17very very early on the Microsoft Mac
10:19programs all had keyboard shortcuts back
10:22when Mac software was not supposed to
10:24have keyboard shortcuts that was a very
10:26big you know quasi-religious debate
10:28about about right-click and keyboard
10:31shortcuts but Microsoft not only thought
10:33they were just a better idea we we
10:35wanted to have them across all of the
10:36platforms consistently because that was
10:38the value proposition do it much like
10:40your Lotus Notes one but then one day
10:42you realize that cross-platform is a
10:44developer thing because no one person is
10:48yeah I mean it's a little bit different
10:49with the mobile devices because you are
10:51much more likely to go from an Android
10:53to an iPhone to an Android but only
10:55every two years well not every not every
11:00and that still is likely to be an edge
11:02case although actually Apple was talking
11:03yesterday about all the people moving
11:05which which really gets to the heart and
11:07soul of your decision framework from a
11:09business perspective as a as a as a
11:12start-up or building a new app you know
11:13cuz up until like just the past six
11:16months or so I think a lot of and you
11:18know your expert on this but I think a
11:19lot of developers had a tendency to be
11:21developers and look at these sort of
11:22broad aggregate numbers and have big
11:25debates over aggregate share or run Rea
11:28or or even fragmentation when in fact
11:30you have to sort of drill down when
11:32you're making a business choice yeah so
11:34I mean Apple guilty of this is anyone
11:38else I mean they put up a chart
11:39yesterday where they said well eighty
11:41percent of forty fifty percent of max of
11:44running um the latest version of Mac OS
11:46and a much smaller proportion of them
11:48PCs are running Windows 8 we think you
11:50okay what's the absolute number of PCs
11:52that are running yeah Windows 8 is I bet
11:54it's a lot a lot more than the 80
11:55million Mac's that are out there um so
11:57you have to kind of sit and think about
11:59your target market I've been this you
12:00know there's one can have a whole other
12:01conversation about this but it's one
12:03thing to say that there are you know as
12:05it might be a billion Androids Google
12:09Android's as opposed to Chinese androids
12:10there's as it might be a billion
12:12androids in use at the moment and
12:13there's maybe 400 million iPhones 500
12:15million iPhones something like that but
12:17then San Francisco is 60% iPhone in the
12:19USA is 50/50 iPhone out Japan is over
12:2350% iPhone now and the rate of that
12:25change and that's swimming Apple is
12:27gaining sharing this yes and then when
12:29you think okay Apple of only selling
12:31$600 devices and Android is selling at a
12:34range of prices so what is the
12:37distribution within those smartphone
12:39owners of people who really really care
12:41you know the people you bought galaxy s4
12:43care about out but most of the androids
12:45being right not those devices they're
12:47the devices you get when you go into the
12:48store and don't care so you know if you
12:50go to an if you go to India or Indonesia
12:52or Vietnam or China you will get a load
12:55of really aggressive highly engaged
12:56users on Android you go to San Francisco
12:58you won't you go to the USA you're
13:00probably going to find that two-thirds
13:01of the aggressive engaged users or I
13:04think so those aggregate market share
13:06numbers tell you they tell you
13:08relatively little because what they're
13:10really just telling you is look
13:11everybody on earth is going to have
13:12smartphone and some of these people are
13:15going to be more worth more than others
13:17and some of them going to be in
13:18different places so when you make those
13:19platform decisions yes there is that API
13:23but it's how does that API relate to the
13:26customers that you want to use yeah
13:28how do those development choices relate
13:29to the addressable market I mean like
13:31even if you're doing you know something
13:33something you know high-end retail
13:35oriented you know if you're doing an app
13:36for a high-end retail brand it's sort of
13:40intimate in the US market it's sort of a
13:41crazy idea I think you need the Android
13:43one first or best when you you know you
13:45already know the demographics of the iOS
13:47customer but conversely you're doing a
13:49communication app that you're likely to
13:51release in in Western Europe or China
13:54first you sort of be crazy to do it on
13:56an iOS first yeah so I have I have a
13:58slide with a kind of a pyramid and an
14:00inverted pyramid and the pyramid is you
14:02know you've got a small portion of the
14:04population that thinks wow I can't wait
14:06to see what new your iPod Beacon apps
14:08are coming out you know not wait to see
14:10what new intense and what cool new stuff
14:12is going to be done on Android you've
14:14got a big chunk of people in the middle
14:16who don't really care and then you've
14:17got another big chunk of people at the
14:18bottom and then when you look at the the
14:21new technologies the really new
14:24cutting-edge stuff applies to a
14:26relatively small portion of the base so
14:29you've got to kind of make that decision
14:30well you know if you are Tiffani's an
14:33ibeacon is really exciting if you're
14:34McDonald's maybe not and that applies to
14:38cross all of this stuff so you know the
14:40cutting edge you know sense if you're
14:42going for the lowest common denominator
14:43of customers the lowest common
14:46denominator of development tools and of
14:48api's and of technologies is probably
14:49going to be the best as well if you're
14:51going for those specific segments then
14:53focusing in on one platform or one
14:54platforms tool is gonna work better
14:56yeah I definitely think that you know
14:58just sort of to wrap up that like you
15:00know certainly if I were thinking about
15:01this now I'd be very focused on figuring
15:03out who our customers are as the driver
15:05for this as opposed to sort of aggregate
15:07market data or geographic data that
15:10could really confuse you into things but
15:12I would also say like we're only at the
15:14very beginning this this really is
15:15almost at a tipping point of these
15:17platforms and so it's very likely that
15:19things are gonna get more difficult and
15:22probably even more quickly like we don't
15:25even know you know androids up next yeah
15:27we have no we have no idea what I know
15:29it's gonna be in five years time and
15:30yeah you know am I feeling yesterday was
15:32this is almost like iOS too
15:34you know there's just a such a big
15:35change yeah lots of it and you know and
15:37this is one where I kind of want to just
15:39you know this is not a new pattern
15:40you know developers love patterns and
15:42this is not a new pattern we have seen
15:44this movie before so just encourage
15:45everybody to really think hard about you
15:48know this least common denominator or
15:50trying to thread the needle because
15:52it's put you at a disadvantage over time
15:55to your domain-specific competitors and
15:57so if you've got the opportunity to
15:59either focus on one platform or treat it
16:02almost like you're building two products
16:03and really really focus your energy on
16:05being great on both and I think that's
16:08just going to be the sustainable
16:09competitive advantage for any company
16:10building mobile apps right now yep so
16:15Steve thank you very much that was
16:16really interesting and we're gonna carry
16:18on trying to work out what we think
16:19about this sure so and also follow this
16:21one on Twitter because I bet there are
16:22some strong opinions and we'll also post
16:24a link to a pretty long blog post I had
16:26a while back on cross-platform
16:28development yeah thanks a lot great