Android is winning - if you're writing apps for China. Elsewhere, though…

If Android is winning (to quote Techcrunch, which managed to read the Gartner release on smartphone share and see the big 68% for the share of Android phones shipped), then howcome American developers still aren't battering down the doors to develop for it? Why aren't they all launching on Android first?

To quote the Techcrunch article,

There is no denying Android's dominance anymore. There is no way even the most rabid Apple fanboy can deny that iOS is in second place now. Android is winning… Developers cannot ignore Android. The old mantra of releasing on iOS and then eventually hitting Android needs to be rethought… Android the ecosystem still sucks, but Android the mobile platform is winning.

To which the simple answer is that those US developers would certainly be rethinking it if Android were really running away with the smartphone business in their own backyard. But it's not quite that simple. Why? Because Gartner's headline figure is for worldwide share. In the US - where many of those developers are - different measures pertain. And in Europe, where many more developers are, different measures pertain again.

What those differences add up to is a world where while it's absolutely true that Android is the best-selling mobile platform, for US and European mobile developers - in fact for any who aren't developing specifically for China - the whole topic is more nuanced.

Let's revisit the Gartner numbers: those said that Android had a 64.1% market share worldwide, and Apple's iOS 18.8%. Other platforms - Symbian, RIM's BBOS, Windows Phone - were all under 10%, and totalled 17.1% (that is, less together than iOS). Android is winning, right?

The China syndrome

But the world is not a level playing field, and it's not split up into evenly-sized areas. Digging a bit deeper into those Gartner numbers (as I was able to do), you learn that China is the world's largest market, and that it's there where Android really stakes its claim. Of the 153.7m smartphones shipped in the quarter, 39m (that's 25% of the total) went to China, and of those 31.2m (80%) were running Android. Apple managed just 4.7m (12%). Everyone else? Just 3.1m. Slim pickings.

Which is great - bringing the internet to people in rural China must be a good thing - but means that the story for Android, the iPhone and other platforms in the rest of the world is rather different.

The US, the world's second-largest smartphone market, took 23m devices (15% of the world total): there, Android was 58% (13.3m) and the iPhone was 36% (8.3m). Everyone else? Just 1.38m, which is astonishingly slim pickings in terms of new phone sales. Yet it fits with what we've seen over Nokia's Lumia sales, which were reckoned to measure in the hundreds of thousands in the US, despite an encouraging launch.

Here's how the picture looks if we split the world into three territories - China, the US, and everywhere else - and show the averages from the Gartner figures. Suddenly it's clear that Android is over-represented in China, and that it distorts the average. If you're a smartphone developer in China, then you'll want to be on Android. But you already knew that.

Yet even that doesn't tell the full story. "Market share" reflects sales, not installed base - and some (many?) smartphone sales are upgrades. What we really want to know is the installed base for each platform; if in the US all of those iPhone sales are going to existing customers, but all the Android sales are going to new customers, then that's significant - for developers, the new customers represent potential growth in sales, whereas someone who already has an iPhone and just gets another one won't have to re-download all their apps. (Of course then you run into the question of whether they give their old iPhone to someone who didn't have one before, who sets up a new account, and then does download new apps. But this is why we're looking at installed base.)

It turns out that the only set of data that gives us a picture of installed base anywhere comes from ComScore, whose monthly data sampling of thousands of users in the US has been giving us a picture of smartphone evolution there since 2009 or so.

And what does that tell us? That Android is indeed the dominant installed base in the US, with nearly 57m users; next is iOS, with just under 36m users. RIM has 11.8m; Microsoft, 4.2m, and Nokia's Symbian is still in the not-yet-cold-or-dead hands of just under a million people. In all, more than 110m people in the US has a smartphone - but that's out of a mobile-owning population of 234m. There's a huge amount of room for those four platforms (and any new ones) to expand into; more than half of users there don't yet have a smartphone.

New users, or just new phones?

So in the US, by the ComScore numbers, Android has 51.6% of the installed base of smartphones in the US; iOS 32.4%; RIM 10.7% (but falling); Microsoft 3.8%.

The shocking thing, really, is that the ComScore numbers suggest that the net increase in smartphone ownership over the past three months was less than 4m - yet in that time Gartner is saying 23m handsets were shipped (of which most, one would think, were sold). That's a lot of trading up by existing users.

We can even break it down further: 13.3m Android devices were shipped to the US; the Android installed base increased by 2.65m (on my interpretation of the figures ComScore releases).

For iPhones, it's 8.3m shipped, while the installed base increased by just over 3m. On that data, Apple would seem to be getting less churn, and growing its users more as a proportion of shipments than Android.

The intriguing thing with Android's installed base share is that it is bumping along at about 51%; the distortions in the US market (which make iPhones on contracts cost almost the same as much cheaper Android phones) mean that the iPhone gets a disproportionate share. Every month, about another million featurephone users shift to a smartphone; and it seems that Android captures about half of them, though sometimes the iPhone does. (It's hard to be sure, though, because there are also defections among platforms, with people leaving RIM and Symbian in particular.) What's clear is that feature phone users are, slowly but surely, upgrading to smartphones.

Everyone's a winner

The conclusion? Android is indeed winning - in the US as well as China. But the single "worldwide" figure masks far more complexity in the market than you might think, and the fact that Android's "market share" (the sales snapshot) is high and growing actually masks a situation in the US - where, like it or not, many of the headline apps appear - that favours the iPhone.

In Europe (for which figures weren't available), Android likely has a similar advantage - but remember the distortion of the Chinese market on the "average". For the rest of the world, Android is closer to that 56% figure of the US, and other platforms stronger than iOS.

Even then, though, the story isn't so simple, because of the twin hassles for Android developers of different devices, and of OS fragmentation, and different monetisation.

From the Techcrunch comments, one finds this:

Meanwhile in the discussion on Branch (side note: it's already useful) about Android fragmentation, there's a comment from Matt Brezina - co-founder and chief executive of Sincerely, and previously the founder of Outlook inbox organiser Xobni:

I'm on the front lines - building both iPhone & Android apps. I won't comment on the revenue disparity between the two - that is for a different branch. But I will say we release lower quality, less tested apps on Android because we can never test all devices - we test live with our customers. It sucks. Luckily the Android customers know to submit their android device & OS whenever they send us bug reports.

Fragmentation in my experience has produced less, and lower quality apps.

Of course fragmentation is often in the eye of the beholder; for someone using a device, it doesn't matter what the latest and greatest version of Android is (and, as is always pointed out, developers can target multiple versions of Android inside the development framework). All that matters is what they have.

And while we're on that topic, the latest monthly figures from Google Play are in, and show that Android 4.0 is rising rapidly, and has now passed "Froyo" (2.2, released in June 2010). It's still the year of Gingerbread, but all the signs are that Android 4.x is going to rule 2013.