by David Strom
I have been using Pandora's online stream music service off and on for
several years. What got me more interested lately was it being one of
the many services on my Roku video streaming box, which my wife and I
use mostly for watching movies from Netflix's "watch instantly" queue.
As I investigated the service more, I came to understand exactly the
challenge of what it takes to be truly multi-platform in the current
era. It isn't just about having both Web and mobile phone versions of
your service, but how you have to go deep into a lot of different
devices to appeal to your customers.
The cool thing about Pandora isn't that you can create your own custom
radio station that will try to find music based on a particular artist
or genre. But that once you set up your account on one platform, you
can access it in your car, in your home, and on the road in between.
All with the same collection of stations and music. As you spend more
time with the service, it tries to figure out your likes and dislikes.
Let's look at all the various places you can get your Pandora fix as
an example of how hard it is to become this ubiquitous. First is the
Web browser: you have to work in a bunch of them properly, so there is
the usual testing in IE, Firefox, Chrome, Opera and Safari. Add Mac,
Windows and Linux versions of each browser, and that's 15 regression
tests right off the bat.
But we have just gotten started. Add in the
newer brower versions, like IE8, the fact that Linux isn't a single
OS, and 64 bit Windows. Then stir in support for both Flash and HTML
v5, and you can easily get more than 200 different environments if you
want to support a wider base. Pandora, by the way, doesn't officially
support much beyond Flash on Firefox, IE, and Safari on Mac and
Windows.
Then we have separate apps for each of the five mobile phone platforms
(Blackberry, iPhone, Android, Palm Pre, and Windows Mobile) and four
cellular providers because their phones work differently on each
network. Never mind that each phone's ecosystem has different rules on
how an app can get posted for download and get itself updated. There
are at least twenty different tests there. The phone apps have to be
designed to work with the limited screen real estate available on each
phone, and yet still connect to your account in a way that you can
recognize without a lot of user training.
Some of the phones have different screen and control button configurations, so just supporting
the Blackberry line, for example, isn't so simple. You also need to
get the development environment for the phone (typically these run on
PCs with simulators that show you what your phone user will end us
seeing) and probably a bunch of phones to test out too.
But wait, there is more. How about Facebook, My Space, and other
social networks? Don't you want to integrate with them and leverage
them to make your app viral? More code to write, more interfaces to
learn, more tests to run to make sure you new versions don't break
these links.
Then there is support for the home-based entertainment systems. While
each of these have some embedded Web browser in them (like the Roku or
the Samsung BluRay DVD players), you still have to test to make sure
that the pages load properly and the music keeps on playing and your
fancy navigation controls operate as intended.
There are more than a dozen different devices, including the Ford Sync in-car service that
will be available later this year, to test out. The trouble here is
that these devices typically have older and less capable browsers that
don't get updated, unlike the PC world where users are trying out new
versions.
As you can see, it is easy to lose count of how many different
platforms you want your app to run on. And then if you have to make
choices and limit yourself, how do you do the triage? Do you drop
Andoid in favor of Roku? Bring up the new Ford Sync API and leave the
Pre to wither away? The user populations of each of these communities
is constantly changing, as sales wax and wane.
It is enough to make many of us long for the simple days of the 1990s,
when we just had to worry about Mac vs. Windows support.
I got the idea to look at Pandora from an article in today's NY Times.
And while the service can wreck havoc on corporate networks (lots of
folks start the audio stream and then walk away from their PCs), I
think they are doing exactly the right kind of things when it comes to
managing their multi-platform strategy.
Comment and Contribute
