Tuesday, 24 January 2012

Mobile App OS choices

When deciding what OS to go out on there on are a number of factors to take into account, unfortunately, as with most things apps, all most people is see is numbers. Numbers are important, if you want distribution, but it is also important to look at the product, who uses it and its fit.

Bloomberg, for example, an app I have helped port to different OSs, is a great fit for Blackberry, Nokia and iPhone as the key business phones in the key Bloomberg markets: finance. While it may be good to be everywhere, you have to ask the cost of that and the incremental gains. Moreover is managing all these versions - they may be within budget to build, but as new services are rolled out, managing a roadmap across lots of mobile OS and their subsets (BB6, BB7, ipad, 3G and 4G screen sizes, etc) can become costly and difficult to manage.

Spotify is on Meego and the Nokia N9, as well as iOS and Android, but not on windows market; why - well the N9 has hit a sweetspot and will ship around the world to the kind of audience who will pay for a premium service, and the are relatively few apps on the device. Sonos is just on iOS and Android, probably based on the premise that while its users have other devices, if you have a premium music device, you will have either a tablet with the aforementioned OSs or a dedicated controller.

Social networks want to be pervasive, however with each OS version comes fragmentation, then there is the issue of prioritising the versions when you update an API and risk offending part of your base: the only option is to update all at the same time or plan very, very carefully and risk backlash! An developer from a social network recently told me that if they did it all again they would do probably just one or two native apps (no prize for guessing which ones) and the rest in HTML5 - he then corrected and said - no, I would do them all in HTML5 and have done: after all you get the most important APIs: location, audio and video, working out of the box, across all devices, in one go. How important is this, well very;
  1. one it means you can stay on or ahead of the curve - rolling out new services on mobile almost or even in parallel
  2. it means you avoid embarrassing issues like the iOS facebook app that recently had everybody in London posting from Walton Abbey for a while
  3. moreover, fragmentation is minimal: another developer recently expressed that his users have upto four different experiences of their product: one on a tablet at home, another on a PC at work, and then another one or two on their work and personal mobiles
A weather app, like accuweather for example will have different priorities, from bums on seats to strategic deals, which will see them having everything from bespoke versions just for the Nokia N97 and Voda 360 homescreens (yep I did that, it seems to be everywhere now, from samsung in store ads to tablet online promos - which is great) to then being on maybe android more than iOS now (how many weather apps must there be on the App store!).

So, in short there are many things to consider:
  1. What APIs if any you need and which support them and which handsets. Apple is still all inclusive (all devices have a compass, gps, etc, etc) however other devices / OS have all sorts of fragmentation 
  2. how easy is it to integrate the service you want on that platform and which platform(s) Android is not too far from Java J2ME, you may want to focus on this from the beginning
  3. How much is your app, how important are ads vs. revenue, freemium support
  4. do you have the need for fancy functionality such as in app subscription, upgrade, etc.. your mileage WILL vary by OS
  5. How important is design and brand. Apple are very aware of the style, UX, fonts, etc and encourage homogeneity with the OS, a clean experience, etc. This means not only is the experience generally more "pleasing" for want of a better word, the apps around it are generally of a high, homogeneous quality as well: no negative association. Android are now starting to do this as well, as per this new announcement of a few days ago of a style guide. Believe it or not, Vodafone were doing this in 2009 with Vodafone 360 and Nokia did it with their own beta apps but that is it.
  6. Your target audience: How many people use and will pay for your service: you may get 500,000 downloads for free on one platform vs 50,000 paid for on another... quality is generally better than quantity.. (unless you are targeted on quantity, of course!)
  7. The submission, update and approval process
  8. The stats, settlement, promotion, how easy will you be to find on the store, etc
  9. what phones your analysts have (if you are quoted :))
  10. can you do it in HTML5, get the usage and adoption stats, build on what you know, not what you think... so much time, so many wasted features have been built over the years on assumption: just get it out there!
Hope this helps...

No comments:

Post a comment