Handset Market Share Data (why it matters)
With Mobile World Congress and CTIA now behind us, it's difficult to not get hot bothered and drooly about what new handsets are going to be on the market. Don't get me wrong, hot, new, fully functional and super sexy are all good things to have in a handset, especially in these otherwise gloomy times. However, as a developer it's important to remember to not get too carried away in the excitement, and here's why. The handset sales figures for November and December in the US were recently released, and a look at them reveals something telling for how one might (or might not) target content development.
Topping the list, was the bread and butter enterprise device, the Blackberry Curve. No big surprise there.This was followed by everyone's fav toy, the iPhone, whose numbers are clearly limited only by the fact that it's shipping on a single network only. Will be interesting to see what happens in July of this year when the 2 year limit falls off of the 1st gen device... The new BlackBerry touch screen device, the Storm, shipping only on Verizon, comes in at 3, followed by an LG touch device, the Dare, and closing out the top 5 is another BlackBerry device, the Bold (also bound to AT&T only). The back half of the top ten is made up of three hybrid touch/qwerty devices, the LG Voyager, the HTC Pro, (in December the HTC fell off the list in favor of the Samsung 'Rant') and the LG Env2, and Samsung closes out the top 10 with a pair of touch devices, the Behold and the Instinct.
So what? Indeed. So What. Let's take a closer look at what this very limited data set means. First of all, there are no raw numbers or even percentage ranks here, so we've got no idea of how wide the gaps are between the top ranked and lowest rank devices, or now the rest are spread across the middle. Unfortunately, that puts a bit of a damper on making real data based decisions, but I'm going to choose to overlook that statistical inconvenience for the time being and just dive in full steam.
First of all, of the top 5 devices, 3 are RIM Blackberry devices, two of which are bound to specific operators. Again, this pretty much makes sense. Despite RIM's recent advertising push to position them in a more consumer centric light, they are still pretty much an enterprise device provider. Despite the fact the other offerings out there, they are still way, way ahead.
Next thing worth noting, more than half of the top 10 devices are not smart phone/open OS devices. All of those LG and Samsung devices are running closed environments and browsers.
Four of the top 10 are bound to Verizon Wireless only. Makes sense as Verizon is still the largest operator here in the States. Clearly not having the iPhone is not holding them back too much here. While all Verizon devices used to come forth with some flavor of the Openwave 6.x browsers, this is no longer the case. One comes with the Teleca Obigo browser, and the others fail to identify themselves in any coherent way.
Thee Samsung devices also make the top 10, sporting different browsers from NetFront and AU. The number here seems on the low side given that in the figures for Q4 2008 show that Samsung was the leading supplier here in the States in the previous quarter.
Something also worth noting in the breakdown of the data from the AdMob report from March is the distribution of screen sizes. Looking at slide 7 on the report shows that the breakdown is nearly even across the Small, Medium and Large screen size profiles defined by the MMA. (Note that the MMA is a tad remiss in actually defining what 'small', 'medium', and 'large' are, but they do provide guidelines on how to size banner images for those particular dimensions in their guidelines).
The other question that all of this leaves gnawing at my brain is that handset shipment and volume over time are one thing, but what about actual usage on the mobile web? Well, here's the thing. if you look at the monthly numbers that were showing up in the AvianResearch data that was being published by RCR (may they rest in peace), you might be led to believe that all the usage out there is going to be coming from devices with these nice big, bright displays. However, when you look at the AdMob data from actual usage about display size, clearly there are quite a number of people out there with moderately sized devices still using them for web browsing on a regular basis.
Ok, so back to the point, and some simple design suggestions for you today.
- Please, please, please don't give up on device detection
- Design your applications/sites to either work w/o scripting, or degrade gracefully in the absence of it
- Serve 2-3 different sizes of your images based on device detection. Either pre-size at creation time, or leverage something like ImageMagick
At some point web (and application) development for mobile is going to be just like building/developing for desktop, at least that's how it should be, but I think that we've still got a way to go. Until then, I'll keep sharing what I know and what I find out...