January 11, 2012
Android as we know it will die in the next two years and what it means for you

I used to think
that, as with Linux and web services in the early part of last decade,
Android was going to be the mortar for the Internet of post PC devices—
an essential ingredient to put stuff together. And as bonus, unlike
Linux which puttered away quietly in the background doing the heavy
lifting for services like Amazon and Google, Android was largely
user-facing and would also therefore benefit from massive platform scale
(and the resulting de-facto standard it would create) the way no piece
of software since Microsoft Windows had. To to see the early onslaught of CES announcements, one would think so.

What all of the talk of Android momentum and inevitability
obscures though is that the dream of a common Android that developers
can write/deploy apps to and users can become familiar with is burning.
More specifically, three events in 2011 burned it and we're now holding
on to a charred corpse that is quite different: an Android so splintered
that it will make the glass on your Galaxy Nexus S2 Prime Pie dropped
on concrete look like an ice skating rink.

The three events: 1. Google buying Motorola and alienating all of the
tier one handset makers (none of which to this day have the spine to
state it publicly but all of which have now come up with their "plan
B"), 2. Microsoft extracting licensing fees from these same handset
makers in the form of IP indemnification and 3. Amazon shipping a wildly
successful, yet unidentifiable, version of an old Android build over
the holiday... and making it a wild success. Of the the three, #1 was
completely avoidable but the other two may just have been the name of
the game when there is so much at stake in the fight of who paints the
interface for the next generation of computing.

The result of this elephant dance? Well it depends on who you are:

Web heads: All of the HTML5 folks should be ecstatic
as it means that despite the laggy performance of mobile Webkit based
"applications," we are going to see a resurgence in startups who target
the emerging Android splinters with interfaces which leave the heavy
lifting on the deployment side to the the web (see the bit about how the
Kindle Fire blocked the Google Market
and vice versa for why) and on the runtime side to the mobile browser.
It won't be as nice and in the short-term and it will lack access to key
device sensors (though it may accelerate our getting those as API
extensions of the DOM) but it is just not feasible to support iOS,
Googlorola Android, HTC Sensedroid, Amazon Fire Droid, etc. if you are a
startup. Big win for this emerging standard.

Users: Remember the olden days when the carriers
were in charge and you got whatever they were serving for dinner? Well
we aren't ever going back to that but I can't help remember a
conversation I had with the head of product for a US carrier last year
at Mobile World Congress where he told me that their ideal world was
"5-10 platforms with 10-20% each." Why? Because in that mess someone has
to help the user figure it all out and they are back to being in a pole
position. I'm not sure they'll pull it off but device OS fragmentation
definitely gives them another at-bat and if there is one thing these
guys have proven it is that preloads work magic to overcome totally
busted user experiences.

Let's not forget of course that as users you'll have to deal with the
aforementioned jankiness of HTML5 applications for a few revs. Trust me
though, short-term pain, long-term benefit.*

Entrepreneurs: last year my advice
was, build iOS and mobile web app and wait until you've got a million
downloads before targeting Android. I see almost no one pursuing that
approach these days, so I'll revise it a little: build an iOS app and a
mobile web app and then go hunting for dollars/help to develop for the
splinters of Android, opting to build yourself only the most generic
bits of app code that you will for sure be able to reuse. If you want to
get on a market where no one will pay you either in $s or in in-kind
promotion, go super lean and build all of your interface in Mobile
Webkit (through something like Phonegap) until you've got a feel for
whether the particular splinter presents a juicy vein of user adoption.

It not a particularly well-kept secret that when WebOS was in its
death spiral, HP would happily pay developers to port any application
which had shown traction to their platform. To my knowledge the Android
tier one handset guys have not done this yet, but given a little time it
may become a reality. There will still be all sorts of headaches
involved, and you might be better off taking the love from Microsoft,
but in a world of several warring Androids, you are the scarce
commodity. Though the more popular splinters such as Amazon's will
likely never have to pay for developers, especially given the fact that
with only one Christmas under their belt, they are already outperforming
the standard Google Market in terms of downloads for some app
categories, the rest will, probably in inverse proportion to how
valuable they will be to getting you users. And in the meanwhile
consider them non-dilutive equity financing sources.

It's going to be a really interesting year for mobile. Having tackled
Android, I'll do my thoughts on iOS next (and it's not coming out all
roses there either).

Finally, one last comment on the
promise of Android to be the "people's operating system—" helping to
bring general purpose computers to those who couldn't afford it
otherwise. I think this is a noble and incredibly worthwhile goal.
However, in the face of all of the IP issues that are facing and will
continue to face the Android splinters, I am not sure the cycles are not
better spent supporting some of the emerging Ubuntu efforts around
mobile phones and tablets. While they too may be unable to escape the
aggressive IP volleys of a monster industry afraid for its survival, and
while there are tons of usability issues to solve, it is a viable
option, especially if Shuttleworth ships his tabletized version of the OS.

Some interesting links for background reading:

Trouble with Android - on the pain of supporting screen size OS version differences.
Android is Clopen - brilliant piece on how Android is more closed than open.
Slowing Android Shipments - The Asymco magic on the platform's growth

Antonio Rodriguez is a General Partner at Matrix Partners.  You can find this blog post, as well as additional content on his blog called The Grain.  You can also follow Antonio (@antrod) on Twitter by clicking here.