Well known-VC/blogger Fred Wilson of Union Square Ventures has reiterated his call for companies to shift to an Android-first development stance over Apple iOS.
This was instigated by some new stats from Comscore on OS market share.
There was a predictable response from the pro-Apple camp. Frankly there are arguments to be made from either side, but most of the valid arguments get lost in the religious fervor that seems to pervade the debate.
But I found a post from Instapaper founder, Marco Arment, which I felt was instructive in contemplating what platform to prioritize.
Arment was nice enough to share statistics regarding the OS & device share of Instapaper iPhone users.
He started tracking this information so that he could better understand his customers better and know when he could start dropping support of older OS versions.
He wrote the post because felt it would be useful info for other developers… exactly the reason I’m sharing it (one thing I like about stats is that it’s pretty hard to spin them and therefore they help facilitate a more objective analysis).
OS fragmentation is important because it has a major impact in how long you need to support code and also what restrictions you have on bringing new features to market.
Happily mobile devices are far more current than desktops and laptops (says a guy who’s writing this on a PC that’s running Windows XP). But it’s not an insignificant issue.
Arment’s stats show that iPhone and iPad users are very good at upgrading their devices to the latest OS. While his stats are only for his users, they’re in-line with numbers from Bump, so I would say that they’re pretty reliable as a benchmark for you to use.
As you can see almost six out of ten are on the very latest version of iOS, more than 9 out of 10 are on 4.2 and virtually all are on iOS 4.0 or higher. The numbers are similar for the iPad.
Now contrast that with the latest official stats from Android (4/1/11) where 63% of users are on Froyo release (Android 2.2) while only 2.5% are on a version of Gingerbread (2.3) with only Motorola’s XOOM tablet on Honeycomb (3.0).
That leaves a full one-third of Android users on Android 2.1 and prior. Now while 2.1 and 2.2 don’t sound like a big difference version-number-wise, the functionality and performance of the two versions are very different indeed.
Froyo brought the much-heralded Adobe Flash compatibility (albeit most Froyo devices today don’t have CPUs to support hardware acceleration in 10.2).
Further many of the Android handsets are hampered from being upgraded to new OS versions by the handset makers who have implemented custom UI mechanisms.
So the lag for keeping the Android OS current will always be pretty significant as the only way many Android owners can upgrade their OS is to buy a new phone — meaning it could be a 1-2 year drag as many wait for the contracts to end.
So, just a little more food for thought as you prioritize your mobile development resources. Don’t just think about market share, think about the implications for your development team too.
Thanks for the trackback
Blah, blah, blah…what’s this got to do with real, practical ‘fragmentation’ concerns?
All of that stuff about Android OS versions is largely pointless info, given that 90=% of all phones are running one of the last three revs: v2.1-v2.3. Are there any significant differences between developing for Android v2.2 and v2.3? I can’t think of any. And except for a couple features related more to storage and UX, I can’t think of any potential problems with embracing v2.1 at the same time.
Well I normally don’t respond to people who don’t use a real name, so I will keep this short.
There are big differences from 2.1 to 2.2/2.3 related to performance (2x-5x performance speedup for CPU-heavy code over Android 2.1 — source Android Platform Highlights page) which impacts the kind of features you may choose to add, Exchange support (modestly important if your app targets business users) and other stuff. So perhaps your app is completely vanilla and this stuff doesn’t matter to you, but it might to other people. Google went out of their way to position Froyo (2.2) as a major improvement over Eclair (2.1), not sure why you don’t see it that way.
And yes there’s more to 2.2 v. 2.3 than you state.
Just to support my view, here something I saw in an GigaOM article today:
“On the fragmentation front, it’s a major concern of developers. A recent report from Robert W. Baird & Co. found that 55 percent of Android developers find OS fragmentation to be a meaningful or huge problem.”