Mobile social networks not in Java
According to Ajit Jaokar mobile social networks should not be built in Java. Changes to content need to be propagated immediately. Users should not be paying for downloads. So just XHTML/CSS and the likes should be used...
Crosslinked from Ajit Jaokars' Open Gardens blog (June 13, 2008).
Why should mobile social networks not be built in Java (or other downloadable platforms)?
Yesterday at the Mobile Web summit, it was great to hear Mark Curtis and his views on how they built up Flirtomatic. Mark is a good friend and also one of my authors. Mark mentioned something subtle .. Which I think many people(including me) had missed for some time ..
He said .. They started with Java for their Mobile social network, then just before their big launch they abandoned Java .. A costly decision but one he has never regretted it.
Instead they exclusively adopted the Mobile Web (XHTML and above) and have stuck with it ever since with great commercial success
Why should mobile social networks not be built in Java (or other downloadable platforms)?
They key to flirtomatic’s success is its extremely fast feature turnaround, adaptation to customer needs, introduction of new features, responses to customer needs etc.
For example – he mentioned ‘roses with glitter’ which they introduced immediately since people seemed to like them, A ‘ring’ for February 29 (the only day on which women can propose to men it seems) etc. Many of these features arose from customer feedback and were implemented immediately. And why not implement in Java?
Because – changes to the service are not propagated immediately. It is a mess – and the user pays for the downloads (airtime) – apart from the time needed to test the service in the first place on a wide variety of handsets.
With the Mobile Web, this is not a problem. They can change the service every hour and the changes are reflected immediately with no cost (and pain) to the users.
This is not a fluke ..
ALL the big social networks are based on the Mobile Web - and none on Java - which is counterintuitive since developers may want to design cool, sexy features and a great user interface - but they dont make business sense in light of the above.
The big players are peperonity , mig33 , itsmy , flirtomatic, mocospace are all above (or in the ballpark) of million plus profiles.
Java has never been to get to such numbers and with good reason as we see from Flirtomatic's experience as above.
Ajit Jaokar
mig 33's reaction
I think this is dependent on markets and product. At mig33 (with 13 million users) we haven't had a big issue with users wanting to download and install a client application. I can't be categoric, but as far as I'm aware most of the current mobile social successes are apps, not web sites.
A mobile web experience on a mobile is going to have less functionality (dynamic updates, phone integration, interface capability) and be significantly slower than a native app. However, there is certainly a barrier in getting someone to download that app in the first place.
Ultimately, it really depends on how important those features are in creating value for your users versus the barriers that creates.
For mig33 we are primarily a J2ME app, but support WAP, mobile web and desktop web, so users can start simply and then upgrade.
As an aside, markets can create some unique barriers, such as in the US with carrier restrictions and a relatively less sophisticated (mobile) audience.