Mobile Browsers vs Mobile Apps, Fragmentation
Last week, Deutsche Bank analysts Jonathon Goldberg and Brian Modoff (excellent analysts, BTW) published an insightful report called “The Big Idea – HTML 5” exploring fragmentation of mobile platforms and the role of the mobile browsers. The thrust of the report was (simplifying greatly): there is a lot of fragmentation and anarchy in the mobile platform and that was not likely to change quickly; with HTML 5 being the first platform to fix this. And earlier today, the good folks at IMS Research chimed in, saying that fragmentation in the Android OS world causes extra hassle for developers, and that if Google doesn’t manage this carefully, it creates an opportunity for LiMo and Symbian.
There is definitely a big issue here. The DB guys hit the nail on the head with their implicit understanding that what is required is a handset “platform”; whether it is a mobile OS or a Browser doesn’t really matter all that much to the developer, as long as the developer can do what they need to get done. I must say that the IMS folks lose some street cred by referencing LiMo (LiMo? I can’t think of any serious handset efforts on that one) and Symbian (Symbian? Hello? Even Nokia has given up on that OS). My work with startup companies confirms that it’s taking more and more development time/people to deal with all the Android variants out there. For those coming with a web developer background, this is hugely annoying. For the minority coming from a traditional mobile application development background, they’re still giddy about the fact that their development team size is single digits rather than triple digits.
I think there is a key observation for entrepreneurs grappling with this fragmentation issue, and that is around what sort of experience you are trying to create with your mobile offering. Of course, there are some applications that require interacting with a specific part of the handset hardware, and there are some that require extracting every ounce of performance possible out of the hardware. For the remaining majority of apps, the idea of trying to match the richness of the current desktop experience (and then complaining about the state of the mobile browsers out there) is probably a mistake. If you’re just trying to port something over from the PC to the mobile, end users will have an expectation of how it is supposed to work, and the fragmentation (Mobile OS platform and/or Mobile Browser platform) issues are going to be really frustrating. But, is “porting something over from the PC to mobile” really going to be where great companies are built in the mobile world? Maybe a few decent companies, but I would argue that the big home runs will come from companies that build their offering from the ground up to take advantage of what is truly mobile. Sometimes that will involve things that can only be done with an app running in the OS, but most of the time, the lowest common denominator smart phone mobile browser is good enough.
The second reason I am significantly biased toward Mobile Site approach has to do with virality. Many of the truly explosive growth offerings in the Internet world have had a viral customer acquisition mechanism tightly integrated into their primary use case. Obviously, this doesn’t apply to all offerings, and trying to “add virality” as an afterthought is highly unlikely to work. But, if your offering has viral customer acquisition as part of its primary use case, then a key part of making that viral mechanism work well is for a large fraction of the people who receive that viral call to action, to actually be able to act on it! If you sent an email or sms to 1000 people in my address book, what fraction would have an iPhone or an Android device? Maybe 20%? So, 80% of any viral customer acquisition messages coming from me would be wasted if the App only worked on iPhone and Android. What if the we ask the same question instead about what fraction had at least a lowest common denominator mobile browser (even BlackBerry’s awful browser counts)? Maybe 90%? Now, only 10% of those viral invites would be wasted. The Site-based approach is likely to be 4 or 5X “more viral” than an App-based approach.
For these two reasons, my view is that it is more likely that the first true mobile Killer App will be Site (Browser)-based rather than an App running in the OS.