The Happy Medium: Building a Smartphone App Store that Works

State of Play for Apple

It seems that the Apple App Store has hit its first real hitch, with the self-interested rejection of applications that potentially compete with Apple's own solutions.

This has, unsurprisingly caused a stormy response. The problem is a combination of factors unique to the Apple App Store:

1. The App Store is the only channel to market for iPhone apps.
2. The App Store reserves the right to refuse applications.
3. There are no guidelines (at all, let alone binding ones) about what sort of apps will be refused.
4. Refusal occurs at the point of publishing (ie. after development).
5. There is no appeal process if refused.
6. The App Store is run by the platform owner (ie. there is a conflict of interest).

This combination of factors creates an extremely unstable situation for developers, because they can invest heavily in creating an app only to have it denied access to the only channel to the end user, for no good reason, with no way to appeal.

Prior to this uproar, the risk of refusal was viewed by developers as relatively low, and the possible return as quite high. Therefore the trade off was considered worthwhile. Now, however, the estimation of risk has rocketed, and the potential return hasn't improved proportionately, so many are reconsidering their investments. This is simply good business sense.

State of Play for Symbian

On the Symbian side, we have quite a different situation.

1. On Symbian phones there are no barriers to applications for many types of applications (but not all types).
2. For more sensitive applications, Symbian Signing is the only barrier to applications.
3. Symbian Signing is run by the OS vendor, who is not the platform vendor (ie. Symbian doesn't sell phones, or even make them -- Symbian has no vested interest, and there is no conflict of interest). Symbian Signing has strict, public processes (not perfect, but at least they're there, and they're improving).
4. Applications are sold via one of four broad channels:
1. self-distribution (via own website, etc.)
2. online distributors (currently limited to two major ones: Handango and Motricity, although Nokia uses a third for their shop)
3. operator shops
4. bundling on device (eg. Quickoffice is bundled on most Nokia and SE phones)

So, the Symbian world has low risk of an application being actively blocked from reaching the user, but as you can see, it also has a low chance of an application making its way into the user's consciousness.

Why do I say an application struggles to come to a user's awareness? Well, look at the distribution channels the Symbian ecosystem offers. Apart from bundling on the device (which is reserved for a very few applications), the channels are scattered and inneffectual. Here in Australia I have seen Apple's TV ad (during prime time), promoting the app store. When will I see such an ad from Handango or Motricity? How come I've never seen Nokia or Vodafone advertising their app store? (In fact, I saw a TV ad for a new S60 device last night, and it merely advertised one feature. Given that the device was the 6210 Navigator, I'll let you guess which one. There was no indication whatsoever that this device was extendable in any way.)

For more commentary on this issue, see All About Symbian here and here.

Solutions

Both of these situations need solutions, and quite urgently.

I think Apple needs to separate the App Store out as a separate company, so that there is no conflict of interest. It needs to publish clear guidelines and follow them strictly.

Symbian needs the Symbian Foundation to step up to the plate and create an app store with all the good features of Apple's (easy to use, well advertised, cheap for developers, source of all quality apps). The Foundation should do this rather than Nokia to avoid conflicts of interest and to allow the store to function for the whole ecosystem (not just the Nokia bit). This could also be a good source of funding for the Foundation, if it's properly managed (Apple claims that their 20% cut is merely covering their costs -- if so, all I can say is that their costs are ridiculously high).

One last point: Apple are evaluating apps on the basis of the impact they have on users or companies. Symbian Signed evaluates apps on the basis of the impact they have on the device or network. I think the former is really impossible to evaluate, and Apple is foolish to even make the attempt. Therefore the proposed Symbian Foundation App Store should merely evaluate the latter (impact on the device/network). Since this is already done by Symbian Signed, we already have a mechanism for determining whether to allow apps on the store or not, and we can simply focus on improving this.

Addendum:

23/09/08: This excellent post from John Puterbaugh (via this week's Carnival of the Mobilists) gives a much broader overview of the distribution system for mobile applications. However, all this new information simply reinforced my ideas shared above -- I still believe that Symbian (and Apple) can improve their ecosystems via the changes I have suggested.