So the decision to accept a new job at AT&T Interactive as an Architect of Lead Generation means I’m moving away from mobile after three years of doing mobile development or mobile-related development. It doesn’t mean I’m giving up writing software for the iPhone or for Android or Windows Mobile: I have an idea I’m tinkering with in my spare time.
But it does mean I won’t be doing it professionally.
The decision to move away from Mobile development when it is glowing red hot is a deliberate one. I believe mobile development is currently a bubble: I do not believe–with perhaps some unforeseen exception–that some mobile startup will ever grow to become the next Google, Microsoft or Amazon. At best I believe you can make a nice living at it–but you’ll never reach a billion dollar market-cap, unless we have a replay of the Dot.Com bubble of the late 90’s.
The rush into Mobile has been driven by the observation that the iPhone has finally created an environment that is relatively “open” (in the sense that the development tools are free and not the thousands of dollars Microsoft and Nokia used to charge–buying a basic Mac Mini capable of running Xcode is still cheaper than Visual Studio Professional for Windows Mobile.) and shipping an application is relatively “easy” (in the sense that it only costs $99 for a year to obtain a hardware signing key, unlike Microsoft and Nokia who used to charge more). It also has created a mobile computing environment that is powerful enough to run sizable mobile applications, with an always-on Internet connection which allows creating internet-enabled mobile applications a no-brainer. Say what you will about Android and the improvements for Windows Mobile 6.5 and Nokia–but Apple showed the way.
And in the process Apple has made a ton of money.
The believe is that by extension there is a ton of money to be made in mobile: if Apple made a lot of money, so can other software developers. But what they don’t realize is that Apple is making money hand over fist by creating an extremely well-designed product (that, at the time, was a revolution in design) that disrupted the traditional relationship between MNOs (Mobile Network Operators) and headset manufacturer. Never before has a headset manufacturer demanded as much control as Apple has–Apple even went so far as to control the pricing structure, options and the initial sign-up relationship with AT&T’s customers.
Riding Apple’s coattails may make you rich–but it probably won’t. And thinking that Apple is succeeding because they just happened to be the right company in the right place at the right time denies Apple’s contributions to the state of the art: nothing prevented Nokia or Microsoft or Sony or Motorola from doing what Apple did, starting the revolution without Apple.
Which means the analysts claiming the Mobile Revolution is finally here–and Apple just happened by coincidence to stumble, like some drunken idiot, onto the center stage just as the elevator started to rise–those analysts are complete fucking idiots. It’s not a Mobile Revolution–though it is a disruption in Mobile. No, it’s a consumer design revolution. Other companies still haven’t figured that out.
Further, and more importantly, I believe development in mobile is–career wise–a dead end. Basically, any mobile company doing repeat and sustained business will take one of two shapes: a company which repeatedly puts out new products (a game company), or a company providing mobile-based information. The former will probably consist of a handful of two- or three-man teams. The latter, at best, will consist of a two to three person client team and a two to three dozen server team.
The smallness of the teams comes from the reality that there is just so much code that you can put on a mobile device. The mobile software I worked on for Geodelic, for example, was perhaps 15,000 lines total–the server side, on the other hand, winds up being megabytes and megabytes of stuff in the source control worked on by a dozen or so great engineers. And ours was a “fat” client: I suspect programs like Where or the MasterCard Priceless Picks clients weigh in at under 8,000 lines total. Even something relatively complicated as the YPmobile app can’t weigh in at more than 12,000 to 15,000 lines.
Mobile information companies like Where are like iceburgs: the 10% that is visible is kept afloat by the 90% you don’t see. Companies as ambitious as Geodelic–which are seeking to integrate all sorts of stuff into their data stream–are even more so: I suspect once Geodelic has built out the system they are seeking to build, by line count the Geodelic Mobile Client will probably be less than 0.25% of the total lines of code that comprise the overall system.
Such companies are extremely ambitious, and perhaps they may grow–not to Google size, but at least to a respectable middle-sized company. But that means that while the server team may serve to occupy two or three floors of a major office complex, the client side of the company will still comfortably fit into a small 8′ by 10′ corner office in that complex.
I love doing mobile work. I intend to keep doing mobile work in my spare time. But my ambitions are greater than being the lead guy of a three person team when I’m in my 50’s. And I just don’t see that happening in mobile.