More apps is not better

As a standards researcher (and pundit and consultant), one of the myths I’ve been trying to fight is that more applications are necessarily better, i.e. a source of competitive advantage. It’s quite clear — as I observed almost a decade ago — that this one of several over-generalizations from the VHS vs. Beta format war of the 1980s. This morning I found another industry — smartphones — where the evidence is compelling.

The influence of the VHS vs. Beta war was powerful, because it happened at a time when contested consumer electronics format battles were rare. Atari and a handful of firms were producing consoles. All of the 8-bit home computers had pretty much died, supplanted by the 16-bit IBM PC. Due to fortunate timing (and a dramatic outcome), it became the exemplar of what a standards contest is about. However, it was not a typical standards contest. In particular, we have different types of 3rd party “software” use patterns. A movie (other than a kid’s video pacifier) you consume once or twice, and thus novelty and variety are key. It doesn’t generalize to other entertainment, because even music is something that you buy once and may use dozens or hundreds of times.

At the other extreme, PC software is something you get one of and tend to use over and over again. Yes , you want a program to serve your particular (perhaps niche) need, but you don't need five or a dozen. One spreadsheet, presentation and e-mail program are quite enough, even if I have 3 browsers installed on my laptop. (Videogames are somewhere in between, but that’s a whole ’nother post).

In my dissertation and subsequent paper, I documented a dramatic example in Apple’s recovery in the first few years of the Jobs II era (1997-present). The Macintosh market share was essentially the same as at the bottom, but if it had good implementations of the key applications — office, a browser, an e-mail client, a music player — a large number of users would say that‘s “good enough”.

In economics, the Betamax model became accepted dogma through the argument of “positive network effects.” I argued that Apple’s recovery demonstrated that PC users didn’t seek the maximum number of applications but “satisficed” if there was enough variety to cover their needs.

For a talk I wrote in the past 12 hours (and gave this morning at the Symbian Partner Event in San Francisco) I did some research on the relationship between app quantity and Q3 global smartphone platform market share.

Vendor	Platform	Est. Share	Apps
Nokia etc.	Symbian S60	≈ 40%	> 12,000
Apple	iPhone	17%	10,000
RIM	BlackBerry	15%	n.r.
HTC, Palm, Motorola …	Windows	≈ 12%	18,000
Palm	Palm	≈ 2%	29,000

Even if the data is crude, the results are dramatic. (Disclaimers: Q3 data is estimated from the Canalys press release, which doesn’t provide an accurate measure of Windows or Palm OS market share or non-Nokia S60 devices. Q3 data for Apple is distorted by iPhone 3G launch. Measures of the number of applications are from 2008, but reported by various sources that may not be strictly comparable).

Looking at the data, the biggest driver of the number of apps (stock) is the age of a platform, even if the recent trend (flow) reflects the demand perceived by ISVs, i.e. current market share and momentum. The dramatic example of the latter is the iPhone, where the app store went from 100 applications in July to 10,000 earlier this week. Still, the iPhone — with rapidly rising share — has far fewer apps than the less popular, older and stagnant (or declining) Windows and Palm platforms.

Thus it’s clear that a huge supply of apps don’t cause success. A certain number might be a prerequisite, but beyond a certain number, more is not better. In fact, too many applications can create fragmentation and confusion among users until demand coalesces around 1-3 apps in a given category that gain enough scale to keep going.

Why didn’t we see this before? Distribution and disintermediation. When you had to get store shelf space, that limits the carrying capacity of the ecosystem, with the entry barrier being firms needing enough scale to win shelf space at dealers (or even catalog stores). In a virtual, post-Amazon electronic goods world, no such natural limit exists, so the hobbyist-programmer in PJs can post an app to the Apple or Google or Nokia app store.

Still, the fragmentation should now discourage (but not eliminate) further entry. New entrants will need to provide a compelling implementation or a find a significant unserved niche, or they will fade into the clutter.

Additionally, differences in the rate of application library growth will not significantly impact the market share of the top 2 or 3 platforms. As with videogames, some apps will come out on a particular platform first, but if they succeed on one, they will be quickly ported to the other major platforms. APIs and tools may slow the availability of apps for some platforms, but if there’s enough demand, eventually they’ll be ported.