Utilizing App Stores as part of an ‘in-the-large’ methodology requires researchers to have a good understanding of the effects the platform has in the overall experimental process if they are to utilize it effectively. This paper presents an empirical study of effects of the operation an App Store has on an App lifecycle through the design, implementation and distribution of three games on the WidSets platform which arguably pioneered many of the features now seen as conventional for an App Store. Although these games achieved in excess of 1.5 million users it was evident through their App lifecycle that very large numbers of downloads are required to attract even a small number of active users and suggests such Apps need to be developed using more commercial practices than would be necessary for traditional lab testing. Further, the evidence shows that ‘value added’ features such as chat increase not only the popularity of an App but also increase the likelihood of continued use and provide a means of direct interaction with users.
TopIntroduction
Adopting the methodology of ‘in the large’ advocated by a number of researchers in the field of ubiquitous computing (McMillan et al., 2010; Rodgers et al., 2007) is particularly attractive to researchers creating applications, now simply referred to as ‘Apps’, on commercial mobile phones as it offers the potential of accessing a large audience of users from a broad demographic compared to lab based studies. However, if adopting this methodology it is important that researchers are fully aware of the effects that the distribution platform or ‘App Store’ has on the users’ consumption of applications.
Initially ‘in the large’ distribution was limited, due to restricted availability through operator portals or aggregators using systems that had their roots in the ring-tone and wallpaper markets built around Short Message Service (SMS) payments and Wireless Application Portal (WAP) push (Garner et al., 2006). As these channels generally required a commercial working relationship with the companies operating portals they were effectively closed to mobile researchers and many smaller mobile development companies. This often led to many researchers having to make applications available through their own websites and hoping for a ‘viral’ uptake to occur. Things improved somewhat with the emergence aggregator sites such as GetJar™ (initially targeted at Java Micro Edition (J2ME) applications but now one of the few stores offering Apps for the majority of platforms) which allowed free Apps to be shared easily by researchers and small developers. The main difficulty for researchers with such services in the early days was the collation of data relating to application as it was often marred by others users of the site ‘high-jacking’ an application rising in popularity by re-posting it on the store in an attempt to drive traffic to some particular web service (Chehimi et al., 2008).
When the Nokia WidSets™ platform (Figure 1) appeared in 2006 it was the first adopting what would now be considered as ‘the App Store approach’ made famous by Apple™ for its iPhone™. These characteristics being the simple search, installation, and rating of applications for the users and the ability for developers to easily push new versions of the applications to those users.
Figure 1. Top: WidSets dashboard with minimized widgets. Bottom: maximized widget views
One of the strong features of this service was the IT was largely device agnostic running in J2ME meaning that it was available across a wide range of handsets not just those which would be considered as ‘Smart phones’. To put this in perspective in August 2010 Oracle suggested that 2.1 billion of the approximate 4 billion phones in the world are JME enabled. The consumer base of a particular App Store or device is an important factor for researchers to consider when considering its use for in-the-large research. For example, whilst some have suggested that Apples iPhone and its App store enables access to a wide audience (Morrison et al. 2010, Miluzzo et al. 2010) the AdMobs May 2010 report on mobile metrics highlight that iPhones represent less the 2% or the world phone market. Further from the breakdown of percentage of these phones within a particular country, shown in Figure 2, it is clear that it is a device primarily used affluent developed countries. This distribution will clearly have an effect on the adoption of particular applications and services.
Figure 2. iPhone ownership worldwide (AdMob, 2010)
In the following sections we present an empirical analysis of the WidSets platform operation through the creation and distribution of three games that achieved in excess of 1.5 million users.