Posts Tagged ‘mobile applications’

While the developed world is rapidly moving over to smartphones, the preponderance of the wireless users are still using basic devices. Sebastian Anthony, in this post, asks how so many people – 4.8 billion, by his count – could be ignored when it comes to providing them with exciting applications.

I don’t know Sebastian, but it seems he’s asking the same questions many new or unfamiliar with the wireless business ask when they encounter this field. This has grown even more true of commentators – professional and amateur – who know only the success of the iPhone and wonder why that success can’t be applied to other devices.

There are, in my view, two reasons why the explosion of applications – available at a full range of prices – happened only on the iPhone (and, more recently and to a lesser extent, on Google’s Android platform): the wealth of the underlying market and the inherent compatibility of all other devices.

Sebastian, here’s something that’s hard to swallow: there might be 4.8 billion mobile phones in the world, but only a few hundred million of those devices are owned by people who have the disposable income to purchase (what amount to) frivolous applications. By frivolous I mean applications that are not life-saving, do not substantially reduce personal or business costs, and do not result in significantly lower expenditure of natural resources. Modern applications for the iPhone and Android are, largely, for convenience. And the members of world society who are willing and able to pay for convenience have, relative to the owners of those other 4.8 billion devices, wealth. Their wealth makes it profitable to develop applications.

The bigger problem, though, is device incompatibility. The marketplace for feature phones (which, in fact, have very few features) is vastly different from the marketplace for smartphones. Feature phones dominate the world market in terms of volume – those 4.8 billion devices are mostly feature phones – but are themselves dominated by the demands of the world’s mobile carriers. Let me tell you about how that market works.

Feature phones run one of many real time operating systems (RTOSs) all closely guarded by, and usually proprietary to, the device manufacturers. Until Nokia’s Series 30 and Series 40, numerous variants of each RTOS where created as new “features” were developed. I’m talking about “features” like one-tap-to-voice-mail, or special-ringtone-for-mom. Nokia tried to standardize that a bit, given the breadth and depth of its worldwide coverage but, largely, the operators rebelled. They wanted each phone to be slightly different from its logical twin on another network in order to differentiate in the marketplace. Further, each feature phone is tediously vetted by each mobile carrier (at great expense to the application developer) and carefully packaged with “features” that lock the customer into other carrier services (voicemail, ringtones, payment, etc). Sony Ericsson’s C902, for example, might have different features in the UK market on the O2 network than it does on the Telefonica network in Spain.

The software industry behind the very popular Java programming language tried to change that. Put one language runtime on all proprietary RTOSes and developers will be able to write one application that will run on top of all of them without change, they said. But the software industry did not anticipate how deeply ingrained the differentiation model was to the carriers’ way of doing business. Each carrier mandated different release levels of the Java runtime, and different suites of Java’s “optional packages”. Developers were back to square one – building for over 100 variants of Java in any typical release cycle. It was simply unprofitable to continue, and firms that tried to make it easier also shut down. Java’s promise of write once, run anywhere became a myth in mobile, recognized as early as 2002 – years before iPhone. But until iPhone, feature phones – even with very little true software innovation – remained the core of the business for both device manufacturers and wireless carriers. And it’s still the success model the carrier’s know best.

In late 2006, Apple, with its huge modern reputation staked on the iPod and iTunes Music Store, leveraged its strength to completely reverse this model. Instead of cow-towing to carrier demands and building tens of incompatible devices, Apple went to market with a single carrier and a single device. Luckily, it happened to be a compelling device – the likes of which the industry had never seen. Customers doubled their expenditure for mobility overnight and every carrier on the planet wanted to take part, no matter Apple’s demands. And when, later, Apple developed a software model that was controlled by Apple rather than by the wireless carrier’s – a model where any developer could create an application for the phone without needing to have it vetted by the carrier – developers jumped at the chance. Even more important, over time Apple has maintained a high degree of software compatibility across its devices. An application that runs on an early iPhone still runs on the newest iPhone – whether it’s running in the US on AT&T or in Luxembourg on Tango. Simply put, this model makes software profitable. Google has successfully used this same strategy for it’s Android platform which has the advantage of leveraging the big device vendors and their creativity and experience in building hardware. But Android devices are expensive smartphones, too.

There is some hope. In modern times we have GetJar, a big-tent app store which is now vending applications at a rate of 80 million units a month to a vast gallery of mobile devices. The jar in the name comes from Java’s software packaging model – witness to the fact that Java is barely hanging on as the only cross-device compatibility model. GetJar’s modest success shows that developers are keen on the low-cost, high-volume model. And it also shows that consumers are willing to trust a previously-unknown entity (GetJar) much more than they are willing to trust their wireless carrier when it comes to buying applications. Let’s hope so – at 80 million/month it’ll take only 600 months for GetJar to reach every phone in the world!