Measuring usage concentration of smartphone applications: Selective repertoire in a marketplace of choices

Abstract

Using the tracking data of 1,645 smartphone users from Nielsen KoreanClick’s panel, this study examines the degree of concentration in smartphone application use in South Korea. The findings from this study are consistent with what we have learned from previous research of usage concentration and selective repertoire in a multichannel and multimedia environment. Overall, the levels of concentration in app usage are high, particularly in the communication and social media categories. Empirical evidence also suggests the existence of network externality in mobile app communication.

Keywords

Concentration network externality Pareto principle smartphone application

Introduction

Mobile phones are arguably one of the fastest growing and most impactful consumer devices introduced in the last decade. Recent advances in mobile broadband technology and applications have further changed the way many people see and use their mobile phones. Beyond regular phone calls and short message service (SMS) exchanges, this ICT device, in the form of the so-called smartphone, is now also a personal multimedia device.

According to comScore (2014), a leading digital measurement company, 163.2 million people in the US owned smartphones (i.e., 68% mobile market penetration) during the 3 months ending in February 2014, up 7% since November 2013. The rapid diffusion of smartphones is even more noticeable in a country such as South Korea, with an established telecommunications infrastructure and consumer demand. The number of smartphone users in Korea topped 38 million as of February 2014 (Ministry of Science, ICT and Future Planning, 2014), accounting for more than 75% of the population.

Corresponding with the accelerating rate of smartphone adoption, new smartphone applications are introduced daily. In 2013, more than one million apps were available in the Apple App Store, with more than 60 billion apps downloaded (Graham, 2013). At about the same time, Google also announced that its Play Store had more than one million apps available for download, with the number of downloaded apps exceeding 50 billion (Victor, 2013).

Faced with so many options, how do media consumers utilize their phones in terms of mobile apps? It is important to note that the number of apps a mobile device owner has downloaded does not necessarily reflect app usage. Furthermore, with an increasingly large array of apps and ever-more complex functionalities, the use of smartphone apps may differ on the basis of functional categories. Certain app categories may receive more traction than others. What do people use among the diverse apps of different formats and content? Can certain behavioral patterns shed light on how today’s mobile communicators utilize these apps to complement their communication needs? What social and market implications might be drawn as the mobile telecom industry develops?

This study examines the patterns of smartphone app use in the context of a country with advanced telecommunications infrastructure and strong demand for and usage of information and communication technology products, namely, South Korea. Although smartphone owners have numerous choices, their limited time and personal needs are likely to lead to certain app selection behavior. Such usage patterns have significant implications on how the mobile industry might evolve as it matures. Analyzing smartphone users’ app use behavior may elucidate consumers’ choices in a multichoice, mobile media environment.

Literature review and hypotheses

There have been various attempts to understand audience behavior. Many of recent studies have adapted Giddens’s (1984) theory of structuration as a way to understand how individuals operate within the media environment (e.g., Taneja, Webster, Malthouse, & Ksiazek, 2012; Webster, 2011; Webster & Ksiazek, 2012). This work sees agents (media users) drawing on the structural resources of the media to achieve their own ends. These resources include the available technologies, programs, and services. As agents use media, they reproduce and alter the structural features of the environment. In this view, agency and structure are mutually constituted, and affect users’ behavior (Taneja et al., 2012). Based on this notion, this study reviews critical approaches on media choice such as uses and gratification, habit, and repertoire to predict smartphone users’ app usage pattern.

Uses and gratifications (U&G) and habit

This study operates under certain assumptions rooted in the U&G and media habit formation traditions. The U&G construct focuses on the uses people have for media in relation to the gratification they can achieve through such use. It relies on the assumption that users are no longer passive but, rather, active audiences that use media content to create meaningful experience (Baran & Davis, 2011). This idea theoretically supports the notion that mobile users seek different forms of gratification through the active selection of specific applications. Furthermore, the U&G approach argues that the level of reward or gratification a user receives from a given medium or message, in addition to the amount of effort required to attain that gratification, directly affects the decision of which media platform to use for which purpose. The proposition supports the notion that a mobile user tends to use preinstalled apps for similar gratification due to greater convenience, less effort, and minimal search costs.

This study also subscribes to the media habit acquisition construct. While traditional media research places the U&G- and media-habit-related concepts at odds with each other, recent developments have demonstrated that habits can be reconciled with the U&G constructs, provided there is a separation between habit acquisition and habit activation. Specifically, during habit acquisition, habit strength and contextual cues have less of an impact on media consumption decisions, since a conscious evaluation of the possible gratifications is required to determine whether the media consumption is fulfilling its intended uses. However, once habit acquisition has been established, habit strength and contextual cues can overpower conscious intentions and dictate future consumption patterns (LaRose, 2010).

Combining the U&G and media habit assertions in media selection and usage behavior, one can conclude that even though smartphone owners have numerous app choices, limited time, personal needs, and habit formation will lead to certain app selection behaviors. Specifically, it is plausible that mobile users will not evaluate all of the options to determine the best applications for their needs. Instead, some level of conscious evaluation will occur prior to habit activation in selecting certain types of apps to meet certain needs and there is likely to be some concentration of consumption within certain app/needs categories because of habitual tendencies.

Usage concentration in media channel

Since the development of multichannel environments, scholars have been examining the concept of channel repertories. Heeter, D’Alessio, Greenberg, and McVoy (1983, p. 133) coined the term channel repertoire to describe “the set of channels watched regularly by an individual or household.” They found that although the cable subscribers were offered 34 channels, the average home watched fewer than 10 channels a week. These repertoires were conceptualized as a mechanism that viewers use to cope with an increasingly abundant and complex media environment (Heeter, 1985). Other scholars reiterated that although a large number of channels are available in the U.S. marketplace, only a small number account for the lion’s share of viewing (Ferguson & Perse, 1993; Neuendorf, Atkin, & Jeffres, 2001; Webster, 2005; Webster & Ksiazek, 2012). Subsequent studies have confirmed the existence of channel repertoires in China (Yuan & Webster, 2006) and Korea (Y. Choi & Chang, 1998). The notion that a few television channels tend to dominate a viewer’s time in a multichannel environment seems to hold true over time and across countries.

Exploring the functional similarities between television and the Internet, Ferguson and Perse (2000) extended the television repertoire concept to the context of website use. They suggested that web audiences use only a small fraction of the content among the numerous sites available to them. Webster and Lin (2002) also concluded that Internet audiences are highly concentrated among the most popular websites, conforming to Pareto’s law, or the so-called 80–20 rule, which states that 20% of a brand’s customers account for 80% of its sales volume (Anschuetz, 1997). Yim (2003) further confirmed a tendency toward web use repertoire formation based on the website traffic data from the usage diaries of sampled college students. Using the web traffic data of 1,669 samples from Internet Matrix, which installed log-tracking files on users’ computers, Hwang (2002) also suggested the existence of a web repertoire by identifying a high concentration of usage of the top 10 websites. Specifically, users spent 72% of their time browsing the top 10 sites out of 100 sites.

The concept of a concentrated media use has been expanded to the newly emerged media such as user generated content service (UGC). By analyzing the content on YouTube and Daum, the most popular UGC service in Korea, Cha, Kwak, Rodriguez, Ahn, and Moon (2007) concluded that the consumption of UGC exhibits a power law distribution with truncated tails. Specifically, 10% of the top videos accounted for nearly 80% of views on both YouTube and Daum. The skewness of user interests across UGC is even greater than the general 80–20 rule suggests. A few years later, as UGC became even more popular, Bent (2010) confirmed the findings of Cha et al. (2007), concluding that 20% of the YouTube videos still generated 89% of the views. Interestingly, not only the number of views but also the active life span of a video (Cha et al., 2007; Cheng, Dale, & Liu, 2008) and the number of comments posted to a video (Bent, 2010) on YouTube comply with the Pareto principle.

With the growth in multiplatform usage, scholars have recently begun to examine cross-platform repertoire of media and content type beyond the repertoire of single mediums (Hasebrink & Domeyer, 2012; Hasebrink & Popp, 2006; Taneja et al., 2012; van Rees & van Eijck, 2003; Yuan, 2011). Findings showed that users’ interests, demographics, and availability (as the rhythms of daily lives) affect the formation of their repertoires across media platform and content. For example, through the recorded observation of locations, activities, and media consumption at a regular interval of a representative U.S. adult sample, researchers have identified four distinct repertoires, TV viewing, computing for work, media on mobile, and media online, out of 59 combinations of platforms and contents (Taneja et al., 2012).

Thus, reviewed literature on U&G, media habit assumptions, and the repertoire indicates a tendency of concentration of consumer media usage. It is our proposition that mobile app usage may be concentrated and follow the pattern of concentration observed in other media usage.

H1: The usage of smartphone apps will be highly concentrated and similar to that of other media repertoire.

Network externalities and smartphone usage

Smartphones, by nature, are communication devices that can benefit from the notion of network externality. Positive network externality is created when a node added to a network increases the value of the entire network. Indirect network externality creates a two-way contingency between the demand for one product and the demand for another (Gupta, Jain, & Sawhney, 1999). In the context of mobile apps, because a smartphone is a networked device, it is logical that the more users utilize a certain mobile function or app, the more utility or value the app can provide to its networked users. The notion of network externality is even more evident when the critical mass of users allows the app to fulfill certain social, belonging needs as opposed to just utility needs (Lin & Atkin, 2002). In addition, an indirect network externality may exist, since smartphones network capabilities can create a demand for relevant apps in each functional area.

Following the logic of network externality, communication apps, as characterized by interactions between users, will show the highest concentrated usage level. Social media apps such as Facebook or Twitter will also show very high concentrations owing to the network externality effects. It should be noted, however, that the role of network externalities in driving smartphone app usage would hold true for certain apps, but not all apps. For example, utilities apps such as QR code reader or alarm clock might be consumed rather individualized and independent of network effects. In case of game category, except for networked games, most such apps are consumed independently and thus, it is likely to be less affected by network externalities.

In fact, in their study of the specific U&G sought via smartphone usage, Bødker, Gimpel, and Hedman (2009) emphasized the functional and social gratification from communication and social interactions and the formation of relationships. Other empirical studies have confirmed that communications-related U&G are the most important function for smartphone users. A study conducted with a sample of Korean college students found that chatting via instant messaging was the most common activity performed on a smartphone, while using social networking sites such as Twitter and Facebook was the second most common activity (Park & Lee, 2012). A Microsoft study found that communications-related apps were the most popular among both high school students and knowledge workers (Falaki et al., 2010). A study of smartphone usage at the University of Colorado also found that 93% of the students use their smartphones regularly for email and text messaging communications apps (Dean, 2010).

It is plausible that mobile app usage would be more concentrated in certain functional categories because of their inherent network externality value. Thus, we expect that apps in communication and social media genre will show higher level of concentration than apps in other genres.

H2: App use in communication and social media genre will be more concentrated than that of other genres.

Additionally, to understand the characteristics of the most popular apps in each genre, the top apps, as measured by total time spent, in each genre were analyzed to answer the following research question.

RQ1: What are the characteristics of the top apps in each genre?

Methods

This study utilizes a completely behavioral data set that measures the smartphone usage of a national consumer panel. Nielsen KoreanClick has a panel of 1,645 users out of 10,962,062 Android OS smartphone users aged 7 to 65 in South Korea. With permission, Nielsen KoreanClick installed “sTrack” software, which records smartphone users’ application traffic data in the 1,645 panel members’ smartphones. The panel consists of 62.8% of males, 38.2% of females. The age divisions are as follows: 10s = 7.6%; 20s = 41.1%; 30s = 32.2%; 40s = 14.4%; and over 50 = 4.9%, proportionally representing the Android OS smartphone users by gender and age in Korea at the time of the data tracking. The collected data for this study includes the 6,545 mobile apps rank ordered by the entire 1,645 panel users during 1 month in November 2011.

The total amount of time spent (TTS) on each app is used as the metric to assess the level of usage for each smartphone app. Such a measure more accurately reflects user engagement and usage levels than a simple count of usage incidents. The tracking software accounts for TTS following certain conditions. Specifically, the amount of time spent in an app is measured by the duration of time when an app is activated on the foreground of a smartphone. When the activated app goes to the background, it is automatically excluded from accounting for the time spent. Furthermore, the calculation of time spent is halted when an app is off-screen.

As indicated earlier, there are millions of mobile apps available for download by the consumers. The types of mobile apps range from navigation, productivity, commerce, education, to communications, media, and more. Because the emphasis of this study is to explore the role of smartphones as a media and communication device, rather than a strictly utilitarian, non-content-oriented productivity tool, we also examined in detail the apps that perform media- and communication-related functions. In particular, the following five app categories were identified and all apps in each category were analyzed. The categories were (a) communication, including mobile messaging, chat, VoIP, and email apps; (b) social media, including social network sites, microblogs, and community-related apps; (c) news, including domestic and international news, broadcasting news, and news portal apps; (d) entertainment, including music, video, photo, animation, and cartoon apps; and (e) games, including puzzle, board game, role playing game, simulation game, and action game apps.

Measurement of concentration

Multiple methods were adopted to ensure reliability in measuring the degree of concentration in application use. Traditional measurements used in media studies were made, such as the concentration ratios (CR), Herfindahl–Hirschman Index (HHI), and shares under the Pareto principle shown by the Lorenz curve and Gini coefficients. The following provides details about each concentration measurement.

CR4 and CR8

A concentration ratio is a measure of the total output produced by a given number of firms in an industry. The most common concentration ratios are CR4 and CR8, the market shares of the four largest and eight largest firms, respectively. Concentration ratios range from 0 to 100.

The HHI

The HHI, an alternative measure of CR, is calculated by summing the squares of the percentage market shares held by the respective firms. For example, an industry consisting of two firms with market shares of 60% and 40% has an HHI of 60² + 40², or 6,000. Unlike CR, the HHI is influenced by both the number of firms in the market and the differences in their relative sizes.

Lorenz curve

Lorenz (1905) curve is a graph that shows the concentration of ownership of economic quantities such as wealth and income. It is formed by plotting the cumulative distribution of the amounts of the variable in question against the cumulative frequency distribution of the individuals possessing these amounts. Thus a perfectly equal distribution is visualized as a straight line at a 45-degree angle.

Gini coefficient

The Gini coefficient was developed to measure the degree of concentration (inequality) of a variable in a distribution. The Gini coefficient ranges between 0, where there is no concentration (perfect equality), and 1 (perfect inequality). The following is the formula for calculating the Gini coefficient in this study:

G = \frac{1}{2 n u} \sum_{k = 1}^{n} | y k - u |

where n = number of apps, u = time spent (average of all apps), yk = time spent (each app), and k = 1, 2, 3…n.

Results

Concentration of smartphone app use (H1)

A panel of 1,645 Android smartphone users used 6,546 apps during the period examined. Among all apps analyzed, the top five apps account for 32.1% of total time spent. Previous channel repertoire studies have indicated that the top 10 to 15 channels account for the lion’s share of viewers. Thus, the concentration measures of the top 10 to 15 apps are also measured in this study. The top 10 apps account for 44% and the top 15 apps 54.1% of total time spent by the smartphone users. The top 20% of apps accounts for 84.1% of the total amount of time devoted by users, exceeding the previously stated 80–20 rule of thumb. Considering the fact that more than 650,000 apps were available to Android market users at the time of the study, the concentration toward a small number of apps is evident.

A Lorenz curve is used to examine the usage data graphically. If the total time spent on each app were equally long, the Lorenz curve would be a straight line at a 45-degree angle, which is known as the equality line. Evidence of concentration is depicted by the extent to which the curve bows down from the equality line. As Figure 1 shows, the curve ascends slowly across the majority of smaller apps but then rises sharply as the effect of the most popular app is added. This is a typical Pareto-like distribution.

Figure 1.

Usage concentration of applications.

The degree of concentration (inequality) can also be calculated as a Gini coefficient (Yim, 2003). The Gini coefficient of all apps used in terms of total time spent is 0.74, which indicates a severe level of unequal distribution. Concentration in the use of apps is also very high according to other concentration measures, such as the CR and the HHI. The severely unequal distribution of apps used by smartphone users is verified by a CR4 of 37.1, a CR8 of 39.2, and an HHI of 589.8. Studies on audience fragmentation note that labels of “modest” or “high” are relative and subjective, as long as one is not planning on hauling someone into court for violating antitrust laws (e-mail correspondence with James Webster and Thomas Ksiazek on April 20, 2012; see also Webster & Ksiazek, 2012).

Overall, as we expected in the H1, smartphone app usage is highly concentrated and the usage pattern is similar to the repertoire phenomenon shown in other media consumption.

Differences in concentration levels by app category (H2)

This study also explores the specifics of the apps that perform communication- and media-related functions. In particular, H2 assesses the differences in usage concentration for these apps. As described in the Methods section, apps related to communication and media functions are classified into five categories: communications, social media, news, entertainment, and games. The degree of user concentration in each category is then analyzed with various measurements. Table 1 compares the usage concentrations across categories for each measure. The Pareto principle (or 80–20 rule) varies by app category. Communication apps have the most skewed concentration levels out of the five different categories, since the top 20% of apps generate 97.7% of the total time spent in the communication category. Other categories of apps also have a skewed distribution. The top 20% of apps account for 94.8% of total time spent in the social media category, followed by news (92.8%) and entertainment (81.5%). The distribution of usage concentration in every category is higher than the Pareto principle (20% of apps account for 80% of total time spent), except for game apps (52.2%).

Table 1.

Comparison of various concentration measures by category.

	Share of top 20% apps (%)	CR4 (%)	CR8 (%)	HHI	Gini coefficient
ALL	84.1	37.1	39.2	589.8	0.74
Communications	97.7	74.2	90.1	2147.4	0.83
Social media	94.8	74.0	84.4	2015.9	0.79
News	92.8	39.2	64.6	656.1	0.74
Entertainment	81.5	47.0	62.6	1098.8	0.64
Game	52.2	28.0	37.2	547.2	0.35

Other concentration measures, such as CR4, CR8, and HHI, also revealed similar concentration patterns that are consistent with the Pareto principle. Communication apps show the highest concentration by the CR measure: the top four apps account for 74.2% (CR4) and the top eight apps 90.1% (CR8) of total time spent by users. The HHI is calculated as 2,147, indicating extreme concentration. Social media apps also show significant concentration according to both CR and HHI measures, with CR4 reaching 73.9%, CR8 84.4%, and HHI 2,016. Entertainment apps also show high levels of concentration, with 47.0% for CR4, 62.6% for CR8, and 1,099 for HHI. News and entertainment apps show similar levels of concentration. On the other hand, game apps were less concentrated than the other categories, with a CR4 of 28.0%, a CR8 of 37.2%, and the lowest HHI of the five types of applications.

The Lorenz curves of both communication and social media are far from the equality (45-degree) line, while the Lorenz curve of the game app category is closer to the equality line. Figure 2 shows the Lorenz curves of usage concentration by app category. The closer the Lorenz curve to a 45-degree line, the less concentrated the total time spent. The Lorenz curves of Figure 2 show that game apps are the least concentrated, while communication apps are the most highly concentrated.

Figure 2.

Lorenz curve of usage concentration by app category.

To further interpret the graphical concentration of the Lorenz curve, the Gini coefficient is calculated. The Gini coefficient of communications apps, 0.83, shows the highest degree of concentration, followed by social media (0.79), news (0.74), and entertainment (0.64). Game app is the least concentrated category. In general, the degree of concentration as indicated by Gini coefficients is high for all types of apps except games. However, even the Gini coefficient for games, 0.35, indicates inequality in user concentration.

A high degree of concentration in app use is confirmed by various types of measures in this study. Table 2 presents the relationship between the various concentration measures. Although the concentration measures adopted in this study have been frequently used in media studies, this study conducts multiple methods to confirm the reliability of each measure. Pearson’s correlations among those measures clearly show that most of these measures are significantly correlated (see Table 2).

Table 2.

Pearson correlations among concentration measures.

	1	2	3	4	5
1. Pareto (top 20%)	–	.985***	.730	.779*	.621
2. Gini coefficient		–	.712	.708	.600
3. CR4			–	.935***	.987***
4. CR8				–	.916**
5. HHI					–

p < .05; **p < .01; ***p < .001.

To assess the degree of difference in usage levels by app categories, the Kruskal–Wallis Test is conducted. The results of the analysis indicate a significant difference in the category medians (χ²= 179.6; df = 4, p < .001). Because the overall test is significant, pairwise comparisons among the five groups should be completed. Thus, Kolmogorov–Smirnov tests are conducted to determine whether the observed Lorenz curves for each pair are statistically different. Table 3 presents the results of the Kolmogorov–Smirnov tests between 10 pairs of app categories. As seen, the observed differences in the Lorenz curves are statistically significant, except for the pair between news and communication. It should be noted, however, that even the least differentiated pair is marginally significant at the p < .10 level. In other words, all pairs of apps are statistically different, as is the post hoc ANOVA test. The most significant difference lies between the game and communication app categories. The game app category is less concentrated than the other categories of apps, while the communication category is more concentrated.

Table 3.

Kolmogolov–Smirnov (K–S) Two Sample Test between categories.

Comparison group	Coefficient D	K–S Z	p-value
News vs. entertainment	0.440	3.111	.000
News vs. game	0.600	4.243	.000
News vs. communication	0.190	1.344	.054
News vs. social media	0.200	1.414	.037
Entertainment vs. game	0.390	2.758	.000
Entertainment vs. communication	0.580	4.101	.000
Entertainment vs. social media	0.440	3.111	.000
Game vs. communication	0.730	5.162	.000
Game vs. social media	0.650	4.596	.000
Communication vs. social media	0.390	2.758	.000

Overall, as expected by H2, among the five app categories, communication apps, characterized by direct interactions between users, have the highest concentrated usage level, followed by social media apps such as Facebook. The result indicates the existence of a network externality effects in the use of smartphone apps. On the other hand, the usage concentration level in the game app category is lower than that for the other categories, as the majority of the game apps were not networked at the time of the study.

Characteristics of the top apps in each communication/media category (RQ1)

To understand the characteristics of the top apps in each category, whether the app is free or paid, who developed it, and why it is appealing to smartphone users were investigated. Share of top apps in terms of total time spent within each category was also reviewed.

Of the communication apps, KakaoTalk (a free Korean brand messenger app similar to WhatsApp) commands 39.1% of total time spent during the sample period. KakaoTalk supports 12 languages and requires no registration or login process. Users can also make free voice calls to friends while chatting with them. KakaoTalk was released on March 18, 2010, by Kakao Inc., based in Seoul, Korea. KakaoTalk has risen to become Korea’s most popular instant mobile messaging service, with over 46 million users registered on May 2012.

Among the social media apps, Facebook is the leading app, accounting for 38.6% of total time spent by users. Note that Korean online users are known for opting for local services over global ones. Even the strongest global search engine, Google, has failed to dominate the Korean market, ranking fourth behind three local search services. Concurrent with the recent rapid adoption of smartphones, however, Korean users are now starting to leave Cyworld, one of the first social networking sites (SNSs) originating in Korea, and joining the global Facebook network (J. Choi, Jung, & Lee, 2013). With the popularity of Facebook and convenient attribute of mobile service, Facebook app beat out other local SNS apps in Korea. Right behind Facebook, Cyworld accounts for 18.2% of total time spent in the social media app category. Interestingly, Twitter, another global SNS service, ranks third in total time spent in the communication category. The power of network externality seems to extend beyond borders.

In the game category, Rule the Sky is the top app, commanding 14.9% of total time spent in game app use. Rule the Sky is a Korean-developed mobile device game similar to FarmVille, the leading global SNS game. Among the entertainment apps, Music Player is the leading app, accounting for 26.1% of total time spent. Music Player is a preinstalled music downloader app that allows its users to search for, listen to, and download MP3 music from the Internet. From the perspective of entertainment content consumption, smartphones appear to be replacing rather than supplementing the role of MP3 players.

The number one news app, Damoa, which means to aggregate in Korean, aggregates more than 20 newspapers’ articles and editorials. It also provides news articles aggregated in portal sites. Damoa accounts for 12.5% of total time spent in the news app category, ahead of the Chosun Daily News app, Korea’s top newspaper in offline circulation. It seems that, in a mobile, smartphone setting, aggregated news content may be a more accepted format for news consumption. Interestingly, Korean search engines and portal sites, such as Naver and Daum, which aggregate various news sources from the web, have already overtaken traditional newspaper and broadcasting news sites on the web. In other words, what attracts news consumers seems to be the convenience and availability of choices, not the brand power of traditional news organizations.

Overall, the share of the top apps in each category is relatively high, from 12.5% in news to almost 40% in communication. Except for Facebook in the social media category, the top apps in most categories were introduced by Korean developers and were not part of the branded apps from established offline media companies.

Conclusion and discussion

Using the actual log tracking data of 1,645 smartphone users, this study examines the pattern of smartphone app use in South Korea. The findings reveal interesting mobile device usage behavior in a multichannel, multimedia environment. Specifically, the empirical data showed that mobile app usage is similar to the usage of established multichannel media in terms of concentration and selective repertoire. There also seems to exist an effect of network externality for mobile apps. Furthermore, communication- and social-media-related apps show an especially high level of concentration. With the increasing number of apps available and growing demand for mobile apps, it is easy to conclude that the market of mobile apps is competitive or has a low entry barrier. Our study here suggests the contrary, as mobile app users actually spent over 80% of their time on the top 20% of the apps. Particularly, for communication and media genre, the total amount of time devoted to top 20% apps exceeded 90%. There is indeed a high level of usage concentration for smartphone applications, regardless of the abundant choices in the marketplace.

Theoretical implications

Theoretically, this study adapted Giddens’s (1984) theory of structuration, uses and gratification, habit, network externalities, and repertoire as a way to understand how individuals operate their smartphone apps usage within the media environment. By examining usage concentration levels in smartphone apps, this study submits that, similar to many traditional media, a certain level of repertoire also exists for mobile apps usage. Although no individual-level usage measures were used in the study, the actual aggregated behavioral data paint a clear pattern of app repertoire and concentration. It should be noted, however, that it is likely that there is a significant difference between the act of mobile app habit acquisition and activation as suggested by LaRose (2010), considering the extent of apps available. As consumers actively evaluate the gratifications and value provided by each app, there would be a higher level of diversity in app usage. Once app habit acquisition has been established, the habit strength and contextual cues lead to the concentrated level of mobile app usage as shown here.

Along the same line of media habit formation, since smartphone is deeply embedded in users’ daily lives, they rely on habit and iteration in creating their app repertoires, reflecting the agentic dimension of routinized action suggested by structuration theory (Giddens, 1984). In other words, in spite of abundant app choices within the same genre, habit formation will lead to certain app selection behaviors. After some level of conscious evaluation (e.g., the habit acquisition stage), habit activates in selecting certain apps and there is likely to be some concentration of consumption within certain categories because of habitual tendencies.

The empirical findings in this study also collaborate on the existence of a certain network externality effect for smartphone apps, especially in the communication and social media categories. The opportunity to interact with more users via mobile apps increases the value of such apps. The network effect becomes even more valuable and influential when we take the proposition of media habit acquisition into consideration, because the networked word-of-mouth can exponentially increase popular apps’ exposure opportunities.

Practical implications

The findings from this study correspond with what we have learned from previous studies of usage concentration in the multichannel and multimedia environment. Such concentration can accelerate the richer-gets-richer phenomenon, as shown in many previous media markets. However, there might be differences between the mobile app market and other media product markets because of the degree of consumer sovereignty involved (Waldfogel, 2005). For the app market, the supply of product is largely based on consumer desires and needs as they are acquired directly by the users and priced individually. On the other hand, multichannel media services are often packaged and funded partially by advertisers, thus involving a lower degree of consumer sovereignty. It is plausible that the negative social impact of concentration in mobile apps use might be minimized due to consumers’ power to choose. Furthermore, the degree of consumer sovereignty in the apps market might aid the financial feasibility of less popular mobile apps under the long-tail economic principle.

It might be fruitful to differentiate the phases between mobile app habit acquisition and activation when considering the practical implications of apps concentration. For instance, there is likely an optimal pricing strategy of mobile apps based on the type of functions it provides and the nature of its demand. Since there is more concentration at the habit activation stage, it would be best for mainstream, high-profile apps to offer subscription rates and niche, newer apps to price by each download.

There are other interesting practical implications regarding the concentration of communication, social media, and news app usage. First, as mentioned earlier, the effect of network externalities in some of these apps amplifies to the marketing utility of a successful app. Second, mobile apps that are more in demand might possess different characteristics from their offline counterparts. For example, aggregated news content, not branded apps of existing media organizations, was preferred by the users of the mobile app platform. That is, convenience and availability of choices are more important than the brand power of traditional news organizations. In general, the results of this study suggest that smartphone usage is heavily skewed toward a few popular apps and user behaviors have similarities between the smartphone platform and existing media platforms. However, content preferences may differ between this platform and traditional media platforms.

Limitations and suggestions

This study is significant in that it does not rely on user recall to assess levels of media usage. Nonetheless, it has its limitations. First, the Pareto distribution is used to perform a static analysis. The number of smartphone users is increasing rapidly and the popularity of specific applications is short lived. Usage concentration must therefore be measured over time. Second, only aggregated total time spent data are examined, so conclusions about individual users’ app repertoires are implied instead of observed. Third, the factor of culture might confound the results of this study. Since South Korea is considered a society with a collectivism orientation, countries with different cultural characteristics might not yield the same usage pattern in smartphone apps. Although network externality is a general phenomenon, admitting a threat to external validity of this study, the driving force of network effects in app usage needs cautious interpretation. Future studies can compare the role of network externality in smartphone app usage among different cultures. Lastly, although this study focuses on the phenomenon of the Pareto principle, it would be interesting to investigate long-tail phenomena in mobile application use. It is expected that the open production and distribution system in the mobile application market will encourage the continued growth of niche apps. Examination of the characteristics of apps positioned in the long tail may reveal greater insight into the changing behaviors of digital media consumers.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Author biographies

Jaemin Jung (PhD, University of Florida) is an associate professor in the Graduate School of Information & Media Management at the KAIST. His research interests include strategic management of media business, media users’ behavior, and media policy issues.

Youngju Kim (PhD, Sogang University) is a research fellow at the Korea Press Foundation. Her research interests include change of media users’ behaviors and business strategies of media companies in the smart media ecosystem. She also has been studying media policy.

Sylvia Chan-Olmsted (PhD, Michigan State University) is a professor in the College of Journalism and Communications at the University of Florida. Her research interests include brand management in media industries, audience adoption of mobile media, and strategic competition between traditional and emerging media.

References

Anschuetz

(1997). Profiting from the 80–20 rule of thumb. Journal of Advertising Research, 37, 51–56.

Baran

S. J.

Davis

D. K.

(2011). Mass communication theory: Foundations, ferment, and future. Boston, MA: Wadsworth.

Bent

(2010, January). The Pareto principle on YouTube. 12th Student Conference on IT. Enschede, the Netherlands.

Bødker

Gimpel

Hedman

(2009). The user experience of smartphones: A consumption values approach. Proceedings of the Global Mobility Roundtable Conference. Cairo, Egypt.

Cha

Kwak

Rodriguez

Ahn

Y.-Y.

Moon

(2007). I tube, you tube, everybody tubes: Analyzing the world’s largest user generated content video system. Proceedings of the 7th ACM SIGCOMM Conference on Internet Measurement. San Diego, CA.

Cheng

Dale

Liu

(2008). Statistics and social network of YouTube videos. In Proceedings of the 16th International Workshop on Quality of Service (pp. 229–238). Enschede, the Netherlands.

Choi

Jung

Lee

(2013). What causes South Korean users to switch from a local to a global social network site? The cultural, social, economic, and motivational factors of Facebook’s globalization. Computers in Human Behavior, 29(6), 2665–2673.

Choi

Chang

(1998). Channel repertoire of cable television viewers. Korean Journal of Broadcasting, 10(4), 323–353.

ComScore. (2014). U.S. smartphone subscriber market share. Retrieved from https://www.comscore.com/Insights/Press_Releases/2014/4/comScore_Reports_February_2014_US_Smartphone_Subscriber_Market_Share

10.

Dean

(2012). Smartphone user survey: A glimpse into the mobile lives of college students. Retrieved from http://testkitchen.colorado.edu/reports/smartphone/smartphone-survey/

11.

Falaki

Mahajan

Kandula

Lymberopoulos

Govindan

Deborah

(2010). Diversity in smartphone usage. Proceedings of Mobisys. San Francisco, CA.

12.

Ferguson

D. A.

Perse

E. M.

(1993). Media and audience influences on channel repertoire. Journal of Broadcasting and Electronic Media, 37(1), 31–47.

13.

Ferguson

D. A.

Perse

E. M.

(2000). The World Wide Web as a functional alternative to television. Journal of Broadcasting and Electronic Media, 44(2), 155–174. doi:10.1207/s15506878jobem4402_1

14.

Giddens

(1984). The constitution of society: Outline of the theory of structuration. Berkeley: University of California Press.

15.

Graham

(2013). Apple announces 1 million apps in the App Store, more than 1 billion songs played on iTunes radio. Retrieved from http://www.theverge.com/2013/10/22/4866302/apple-announces-1-million-apps-in-the-app-store

16.

Gupta

Jain

D. C.

Sawhney

M. S.

(1999). Modeling the evolution of markets with indirect network externalities: An application to digital television. Marketing Science, 18(3), 396–416.

17.

Hasebrink

Domeyer

(2012). Media repertoires as patterns of behaviour and as meaningful practices: A multimethod approach to media use in converging media environments. Participations: Journal of Audience & Reception Studies, 9(2), 757–779.

18.

Hasebrink

Popp

(2006). Media repertoires as a result of selective media use: A conceptual approach to the analysis of patterns of exposure. Communications: The European Journal of Communication Research, 31(3), 369–387.

19.

Heeter

(1985). Program selection with abundance of choice: A process model. Human Communication Research, 12, 125–152.

20.

Heeter

D’Alessio

Greenberg

McVoy

(1983, May). Cableviewing. Paper presented to the annual meeting of the International Communication Association (ICA), Dallas, TX.

21.

Hwang

(2002). Web repertoire and concentration rate: Analyzing web traffic data. Korean Journal of Broadcasting Research, 54, 371–398.

22.

LaRose

(2010). The problem of media habits. Communication Theory, 20(2), 194–222.

23.

Lin

C. A.

Atkin

D. J.

(2002). Communication technology and society: Audience adoption and uses. Cresskill, NJ: Hampton Press.

24.

Lorenz

M. O.

(1905). Methods of measuring the concentration of wealth. American Statistical Association, 9, 209–219.

25.

Ministry of Science, ICT and Future Planning. (2014). Statistics of smartphone subscribers. Retrieved from http://www.msip.go.kr/www/brd/m_220/view.do?seq=424&srchFr=&srchTo=&srchWord=&srchTp=&multi_itm_seq=0&itm_seq_1=0&itm_seq_2=0&company_cd=&company_nm=&page=1

26.

Neuendorf

Atkin

Jeffres

(2001). Reconceptualizing channel repertoires in the urban cable environment. Journal of Broadcasting & Electronic Media, 45, 464–482.

27.

Park

Lee

(2012). Social implications of smartphone use: Korean college students’ smartphone use and psychological well-being. Cyberpsychology, Behavior, and Social Networking, 15 (9), 491–497.

28.

Taneja

Webster

J. G.

Malthouse

E. C.

Ksiazek

T. B.

(2012). Media consumption across platforms: Identifying user-defined repertoires. New Media & Society, 16(4), 951–968.

29.

Van Rees

van Eijck

(2003). Media repertoires of selective audiences: The impact of status, gender, and age on media use. Poetics, 31(5–6), 465–490.

30.

Victor

(2013). Android’s Google Play beats App Store with over 1 million apps, now officially largest. Retrieved from http://www.phonearena.com/news/Androids-Google-Play-beats-App-Store-with-over-1-million-apps-now-officially-largest_id45680

31.

Waldfogel

(2005). Does consumer irrationality trump consumer sovereignty? Review of Economics and Statistics, 87(4), 691–696.

32.

Webster

J. G.

(2005). Beneath the veneer of fragmentation: Television audience polarization in a multichannel world. Journal of Communication, 55(2), 366–382.

33.

Webster

J. G.

(2011). The duality of media: A structurational theory of public attention. Communication Theory, 21(1), 44–67.

34.

Webster

J. G.

Ksiazek

T. B.

(2012). The dynamics of audience fragmentation: Public attention in an age of digital media. Journal of Communication, 62, 39–56.

35.

Webster

J. G.

Lin

(2002). The Internet audience: Web use as mass behavior. Journal of Broadcasting and Electronic Media, 46(1), 1–12.

36.

Yim

(2003). Audience concentration in the media: Cross-media comparisons and the introduction of the uncertainty measure. Communication Monograph, 70(2), 114–128.

37.

Yuan

(2011). News consumption across multiple media platforms. Information, Communication & Society, 14(7), 998–1016.

38.

Yuan

Webster

J. G.

(2006). Channel repertoires: Using peoplemeter data in Beijing. Journal of Broadcasting & Electronic Media, 50(3), 524–536.