This week we pulled the data for app installs for a randomly-selected set of 5,000 anonymized users of AppESP. These users had installed a total of 246,458 apps on their Android phones, or 49.3 apps per user. We wanted to understand how many distinct apps were installed by these Android users, as well as the distribution of number of installations per app. The results were pretty surprising.
Ever since Apple launched the iTunes App Store for iPhone back in 2008, smartphone platform vendors, media companies, bloggers, and geeks have been talking back and forth about the number of apps available and whether quantity or quality is most important. Apple has famously taken an approach that tries to tightly control what shows up in their App Store, while Google’s Android Market has been a bit more of a wild west, anything goes environment. Microsoft’s Windows Marketplace for Mobile also applies significant editorial control, as does Amazon’s Android Appstore.
With hundreds of thousands of apps available, Apple has — until very recently — ruled the market in sheer quantity and has had a credible claim regarding quality as well. Within the past couple of weeks, published reports have shown the Android market catching up to Apple in the number of free apps, and forecast to pass it overall in the very near future. Meanwhile, Microsoft and RIM can already boast of 10’s of thousands of apps, with strong, continued growth.
On the other hand, some observers have begun questioning whether these huge numbers of apps even matter. If the average Android user installs about 35 apps and the average iPhone user about 48, do we really need 300,000 or more apps on a single platform?
At AppStoreHQ, we are in the business of mobile app discovery; that is, we help users find high-quality, personally relevant apps for their smartphones. We do this through our proprietary AppRank technology, which combines media analysis, social media tracking, and collaborative filtering to deliver a ranking score for each app. For the Android platform, our AppESP app uses AppRank along with the set of apps that users have installed on their devices to provide personalized app search and recommendations.
The 80-20 Rule?
The Pareto Principle, often called the 80-20 rule, says that, for many events, 80% of the effects come from 20% of the causes. For example, 20% of your customers may generate 80% of your sales (or refunds!) In the case of mobile apps, there are a small set of applications that make up a very large fraction of all the installs.
The data from these 5,000 users shows that, in fact, it only takes 10.5% of the distinct apps to make up 80% of the total app installations. So, at first glance, it seems that perhaps the nay-sayers are correct and that there’s really no need to have even the 18,000 apps from the Windows Phone platform, much less the hundreds of thousands available for iOS and Android.
The picture is not as simple as it seems…
We begin to see that users have pretty distinct tastes when we look at the percent of users who have specific apps installed. Not surprisingly, the vast majority of Android users have the Gmail and Google Maps apps installed (in fact, it would be interesting to know who these users are who don’t!) And, it turns out that Facebook is pretty popular as well; installed on 86% as many phones as Gmail. Amazon’s MP3 player comes in at 49%, the official Twitter client at 44% and Shazam at 42%. But, only 19 apps are installed on 30% or more as many phones as Gmail.
The bottom line
We randomly selected a sample of 5,000 Android smartphone owners who have installed a total of more than 246,000 apps on their phones. The big question is, how many different apps did they install? Do ecosystems really need so many apps? Do users care about this much variety? Well, these users actually installed 20,100 distinct apps! Seen another way, among these 5,000 users, on average, they each had over 4 apps of which they were they only user. Out of the 20,100 apps, over 19,000 were installed by fewer than 250 users (5% of the sampled users).
There are a couple of interesting implications of this surprisingly high number. First, there actually is a need for a wide variety of apps in a well-developed mobile platform ecosystem. You can’t just rely on having a few dozen of the top apps and call it good. Given the tens of millions of smartphone owners, it’s clear that a large catalog of apps is a competitive differentiator and that it is more than just marketing spin that gets Apple, Google, Microsoft, RIM, and others talking about how many apps there are.
Second, with so many mobile apps being used, and such distinct tastes among device users, app discovery becomes critical. Just looking at Top 25 lists, or even Top 200 lists, will not expose users to the apps they want to maximize the enjoyment and usefulness of their phones. Social discovery, powerful analytics, and human curation will all be needed to solve this problem in a scalable way.
AppStoreHQ — and our Android app, AppESP — are designed to solve this discovery problem by providing search and personalized recommendations of mobile apps based on a combination of online media analysis, collaborative filtering, and social graph traffic. We’d love it if you’d check them out.