Why aren't smartphone app developers going where the sales are?

When you mash up the latest smartphone data from Gartner with the latest app development interest from Appcelerator, you get a picture that doesn't make much sense

This week, Gartner produced the quarterly smartphone sales figures, which were duly sliced and and diced by Charles Arthur: these figures show that Android rules the world, with over half the share of units shipped during the three months from July to September.

But a survey of developers by Appcelerator, prepared around the same time as a Gartner numbers shows a disconnect: developers are targeting platforms with minimal market share and not going all out for Android.

What's going on? Have we all gone mad? Why's the market telling us one thing, yet we're doing another?

The data

The article contains a chart showing the change in percentage of actual sales over time. You can see the massive green blob representing Android's takeover, smushing Symbian, RIM and the others out of relevance.

In the Appcelerator survey, we see this chart that shows iOS development is very interesting to developers, Android development is the next most interesting, HTML5 mobile web cross-platform development next, and then Windows Phone 7 makes a decent showing with 38%. And then come the others.

(One issue with this comparison is that Gartner's data does not represent people who own two devices. In the end-user, consumer space this number is not likely to be relevant and hence won't be under-reported. In the developer space, most developers are likely to be thinking about multiple devices ? ie a developer who's interested in building for Android may also be interested in cross-platform HTML5-based development. As we go though, we should be able to render this distinction moot.)

What hasn't been widely reported from the Appcelerator survey is this chart that shows the change in developer interest over time:

(You can find the chart on page 8 of the PDF report.)

Over the period from Q1 2010 to Q3 2011, Gartner reports that iOS's position in the smartphone market has held steady at around 15%. Likewise, on the interest scale, Appcelerator's data shows developer interest has been such that about 90% of developers targeting mobile have been interested in iOS. So one reading would be that 90% of developers want to invest in 15% of the market.

Looking at WP7 next, this starts off in the Appcelerator data with niche interest in January 2010 [miraculously, since it wasn't publicly known of; possibly this is the aggregate of Windows Mobile and Windows Phone - Tech. Ed] but then climbs quickly to the low-to-mid 30%s in March 2010 and holds firm to the present. Taking a rough average, one-third or mobile developers fancy their hand with WP7.

Finally, we have Android. If we ignore the first three months and go from March 2010, interest runs at around 82.5% for the whole period. But in the same period, the Gartner shipment data goes from around 10% to over 50%.

To sum up:

? 90% of developers like iOS, even though it's only got 15% of sales
? 33% of developers like WP7, even though it's only for 1.7% of sales
? 85% of developers like Android, which once had 10% but now has 50% of sales.

What's happening here is that attitudes are remaining the same over a roughly two-year period, in that iOS and Android are cool and WP7 might be interesting, even thought the market is now flipped on its head and Android owns it.

But should we instead be running in terror from building anything that's not Android, deleting Xcode and installing Eclipse?

Disconnect

Developers are, at the heart of it, engineers, and as such we look for patterns. We look to build controllable machines where known input plus known quantity equals known output. In the fluffy world of customer behaviour, the only thing that we have to go on is history; and history, in our case, suggests that if the availability of Windows apps killed the Mac in the PC era (and by "killed" I mean gained 95%+ market share), then surely we as a community can do the same thing and control the market? If we build it, the users will come.

Except for that's not happening. The Windows Phone Marketplace now has 35,000 apps, and 1.7% of sales. The Android Market has around 360,000 apps and 52% of sales. The Apple App Store has around 500,000 apps and 15% of sales.

Even a "quality" argument doesn't stack up here. The Apple App Store has much better apps than Android (take a look at this article for more thoughts on this topic). So we can't say that the smaller number of Android apps has driven platform adoption by giving users on that platform better quality.

Clearly then, it doesn't matter what we as developers do. The market's going to do what it fancies with or without us.

Reacting

What I think these numbers show us is that as developers we've got to shift out of this position of assuming we're important in all this. We don't need to build up app store numbers for the good of the vendors. We need to be watching the market and understanding where the customers are taking us. This is a good thing ? it's a sign of maturity in our industry. After all, no other mature market behaves like this. Most people listen to their customers rather than blithely continuing with old thinking as the world changes underneath our feet.

Matthew Baxter-Reynolds is an independent software development consultant, trainer and author based in the UK. His favourite way to communicate with like-minded technical people is Twitter: @mbrit.