Part II: Diversity
Part of Apps and Their Stories: Volatility, Diversity, Policy
Evelien Christiaanse, Jason Park, Oana Patrichi, Vanessa Richter, Ziwen Tang (alphabetical)
Anne Helmond, David Nieborg, Fernando van der Vlist, Esther Weltevrede (alphabetical)
Summary of Key Findings
- China is a unique case in terms of its diversity across the Apple App store and Google Play Store as well as across countries.
- Different degrees of diversity can be identified but a specific entrypoint needs to be defined.
- Other factors need to be taken into account influencing metrics such as cultural proximity, consistent concentrations across platforms and countries, media concentration as well as the socio-economic affordances inscribed into the different app stores owned by Apple and Google influencing their availability.
- Overall, each country and app store showcases different levels of diversity in each metrics. Therefore, while a country or app store might be very diverse within a specific metric, it might be very uniform in another. Bringing all measures together can lead to a more complex understanding of diversity on the level of the app stores as well as the analysed countries.
App stores are generally used to analyze the connections between apps “as issue spaces for platform studies of developer engagement and innovation, media concentration and the conditions of possibility for practices” (Michael Dieter et al. 3). In other words, app stores are an important research space as they provide the suitable ground for exploring media concentration.
In light of current events stirring the media world, namely the antitrust case that Apple has recently faced (de Looper, 2018) the issue of app store monopoly arises. Charging developers seven percent of total revenues (on non-game apps) and allowing users to access apps from not only its own store but also from Amazon or other third party stores, Google Play - Google’s successful attempt to create an Android App Store equivalent (Moazed & Johnson 12) distinguishes itself from Apple - which imposes a thirty percent charge for all products and preventing iOS devices from “being able to install apps from other sources” (de Looper). Although Apple targets a limited consumer sector that is ultimately controlled by the company’s monopoly through a tight interdependence between its own products, Google has joined the e-giant, competing for popularity amongst developers and app users alike. Therefore, as these two operating systems are currently the dominating figures on the market, we can speak of a monopoly over the applications allowed on the phones of their users. Aiming to examine this monopoly further, we are interested in researching whether the top apps available in the two app stores in the United States, the Netherlands and China display a similar monopolistic pattern by analysing their diversity, taking into account four different metrics: app name, developer, category and country (headquarter of developer).
To make the case for the importance of diversity in media, Philip Napoli draws on the ‘marketplace of ideas’ metaphor (8). He contends that diversity is the foundation for a well functioning democracy, using the United States as an example. The First Amendment allows for freedom of choice from an extensive “range of ideas (content diversity), delivered from a wide range of sources (source diversity)” ( 9). This diversity (what Napoli names “exposure diversity”) aids citizens in the improvement of their overall knowledge and power of decision-making, allowing for a diverse overall perspective. He believes that, through diversity, well-informed citizens are developed, which augments the democratic process and renders a “self-governing society” whose members successfully carry out their democratic responsibilities(9). The three ‘types’ of diversity that he lists - content, source, and exposure - show that there are many ways to evaluate diversity and how diversity in each should help them become better informed participants in democracy. This research will largely focus on the content diversity of the two app stores.
As Dieter et al. have researched, apps are digitally native objects, “diversely situated, connected to and related with through different queries, topics, categories, developer categorisation as well as user behaviours” (3). Therefore, within this paper, we are following the previous research conducted by Weltevrede and Borra, who have coined the term “device cultures” - “the interaction between users and platform” (2) - of the media platforms, pointing out that “devices are not mere tools; they are also complex and unstable arrangements that bring together a variety of people and objects with particular purposes” (2). While their study focuses on Wikipedia’s device culture, we are taking app stores as a starting point in order to fill the current academic gap in the field of app store study: since little to no research has been done on diversity, we wanted to find metrics to methodologically look at the diversity within the app stores as well as determine whether different degrees of diversity or monopoly, mirroring the one outside of the apps (Google Play, Apple Store), exist.
2. Initial Data Sets
The Raw Data
has been collected on 8-1-2019 2pm (GMT+1), the rankings thus reflect the values on that day. It must be noted that rankings can differ if collected at another time. Using App Annie, we collected data on the top 50 apps of the Google Play and Apple App store namely the app name, developer, category, country (developer headquarter) as well as the image URLs for visualisation.
3. Research Questions
How diverse are Top 50 free apps listed in Google Play and Apple’s App Store across the Netherlands, US and China (distinction between market and the owning company)?
- Looking at the apps, how diverse are the Top 50 apps between Google Play and Apple Store in each country as well as across the three countries within each app store?
- How diverse are the results for each and across app stores regarding the distribution of categories in each as well as across countries?
- How diverse are developers providing the applications within Google Play and Apple in the three countries and across stores?
- How diverse are the developers based on the company’s headquarters?
For a comparable analysis, we chose to concentrate on the Top 50 free apps from the two largest app stores (Statista) - Google Play Store and Apple App Store. As we want to focus on the diversity of the general app store in this research, we decided on the dataset of the Top 50 free apps (dubbed as Top Apps and Top Charts on Google Play and Apple App Store, respectively) as those represent the most downloaded body of apps within each store. As the app store is thought to follow a power law distribution, like the search engine results have shown (Ifrach and Johari 8), we decided to collect the first 50 results as the “lion share” of downloads after the power law falls to the first apps within the ranking. Furthermore, we chose to investigate three countries, namely the United States (US), the Netherlands (NL) and China (CH) in an effort to have a broader look at app stores around the world. Both China and the US represent large markets, while the Netherlands serves as a European country for comparison, however, we are aware that that the Netherlands does not represent the European Union or the countries within. We decided on four metrics to test the diversity of the app stores (app names, app categories represented, app developers, country of the developers headquarters) to establish different ways of thinking and analysing diversity within the space of the app stores.
As a first step, we gathered six datasets on the 8th of January 2019 (2pm GMT+1) through AppAnnie
- a website for app market data and analytics providing insights into different app store rankings, countries and metrics (App Annie - The App Analytics and App Data Industry Standard) - on the Top 50 apps for each country in each of the two app stores. As there is no function to download the rankings, we manually copied the data from the website and pasted into a Google Sheet to collect the raw data sets. We discovered that the app rankings are incredibly volatile. App Annie takes a snapshot from that exact moment in time, so the app lists can change several times each day. Keep this in consideration when trying to do similar research; it would be wise to gather all your data at one time, and, if collecting for several countries, consider collecting data at the same local time.
Following our four metrics, the data for each metric was analysed separately. All apps listed in Top 50 of both app stores have been retrieved by country. The similarities and differences of apps between the two app stores within each country as well as across countries have been examined by using the DMI Triangulate Tool. Furthermore, we scraped the app image URLs from AppAnnie
to visualise the distribution of applications across the two app stores and countries (see Images 1 and 2). As we did not retrieve the image URLs on the same day as the actual data, we came across the problem that the ranking had changed within the Top 50 for a high number of apps. Therefore, we had to adjust the image URLs according to the previous dataset as well as manually retrieve missing image URLs.
We also compared the categories of the apps to interrogate their diversity by establishing the number of categories present within Google Play store and Apple App store in each country comparing it to the total number of categories. The categories are provided by the app stores and assigned by the developers. There are 28 apps available in the iOS store, including 3 categories specifically for macOS (Categories and Discoverability - App Store - Apple Developer) and not including subcategories for Games, Kids and Magazines and Newspapers. We decided to set the number of overall main categories to 25 as the focus of the study is on mobile applications hereby excluding apps targeting macOS or iPad. Similarly, the number of main categories available in the Google Play Store are 33 including Games as a main category summarizing the 17 subcategories for visualisation and analysis under Games (Select a Category for Your App or Game - Play Console Help). Moreover, we visualised the distribution of apps per category within each app store per country to discover concentrations before we compared the data across stores in each country as well as across countries within each app store.
The third metric focused on the developers listed by the stores within each application. As developers name the companies owning the apps this does not necessarily signify that they developed the app, as bigger companies hire digital agencies to produce apps. Similar to the app names, the developer names were triangulated across two and all three countries within each store as well as comparing the two app stores within each country. Finally, a list with unique developers across all three countries and two app stores was produced by filtering the developer names and counting their occurrences within the dataset of 300 applications. Besides the number of occurrences of developers, the corresponding visualisation includes the related countries of the developers headquarters to include the the diversity between countries.
To analyze the diversity of the developers headquarters country we filtered the specific column for both Apple and Google Play for the United States, the Netherlands and China, counted them and compiled the data into charts to create a clear visual representation of the diversity present across each country (see Image 4). However, as, while sorting the data, a number of countries were returned unknown, we have conducted additional research through Crunchbase, Google and LinkedIn
to diminish the error margin of our research and (strive towards the highest level of data accuracy achievable). 18 values out of 300 could not be identified; therefore, we cannot assert with certainty how the ‘unknown’ countries could affect the final results, were they to be known. An interesting observation is that while the US and the Netherlands presented a limited number of unnamed headquarter countries (1-2), China’s Google Play displays a prominent 12 unknown app geo-locations.
Analysing each metric (app names, app categories, developers, countries of developer headquarters), we focussed on the distribution within each country across the Google Play Store and the Apple Store as well as across each store across country. Within the first metric of the apps themselves, the United States and the Netherlands showed a similar degree of diversity within the Top 50 apps across the two app stores. It was found that 19 apps are listed in the Top 50 of both OS (operating system) platforms in the United States. Among these, Game (9 out of 19 apps) is the most prominent category, followed by entertainment (2), social networking (2) and photo and video (2). Similarly, 23 apps are found in the Top 50 of both Apple and Google Play app store in the Netherlands. QQ, a Chinese messaging app, was the only common app found in both Apple and Google Play in China.
Image 1: Diversity in the top 50 apps in each country across platforms (please refer to Appendix for full-size images).
In Google Play, there are four common apps found across the three countries, Stickman Hook (game), Instagram, Whatsapp, and Google Play games. In the Top 50 free apps in US and the Netherlands, they share more similarities by having 21 common apps in their rankings. Pinterest, Polysphere (game), and Snowball.io (game) are the three apps in common between the Chinese Google Play store and either the American or the Dutch Google Play Store. By contrast, more diversity was found in the Apple Store in this cross-country analysis, which presents no overlapped app among the three countries. While looking at Apple Store in the US and the Netherlands, 20 apps were found as overlapping in their rankings.
Image 2: Diversity in the top 50 apps in each platform across countries.
While in the Apple store between 14 to 17 categories out of 25 are represented, there is a clear concentration on the category Games in the United States and the Netherlands with 14 and 15 apps falling in this category. However, the distribution of apps in China is more diverse. Similarly, in Google Play the majority of apps fall into the games category for the United States and the Netherlands. China’s Google Play Store, however, is less diverse than its Apple counterpart with a high concentration on the category tools. As the Google Play store is banned in China, the different usage of the two major app stores might be relatable. Looking at common apps across countries and platforms, Games (7 out of 23 apps) is the most popular category, followed by entertainment (3) and productivity (3).
Image 3: Categories of the top 50 tops in each country across platforms.
Focusing on the developers, a high number of original developers can be found within each app store across the three countries as well as across Google Play and Apple (between 35 to 44 out of 50 in each). However, there are seven common developers across the United States, the Netherlands and China and twenty common developers accounting for any combination of two countries. Herein, a concentration of developers can be seen coming from the United States, mostly belonging to GAFA (Google, Apple, Facebook and Amazon) and the Chinese equivalent - BATX (Baidu, Alibaba, Tencent and Xiaomi). Furthermore, while there is a high number of common developers in the Google Play Store (19) and Apple (14) in the United States and the Netherlands, only few common developers can be found within the Chinese Google Play Store (US: 5, NL: 6) and Apple Store (US: 1, NL: 2). As the Chinese market seems to be overall differing from the United States and the Netherlands, the Google Play store is more comparable as only two common developers (Voodoo, Digiart) can be found within the Apple App store. Moreover, it should be noticed that Voodoo is a French developer that acts as a publisher for developers, which helps explain their dominant position occurring 15 times within all three countries app stores and therefore the second most, right after Google (16). As they provide services for international developers, the company seems to serve a comparably international market.
To investigate the metrics of developers further, we secondly included the distribution of developers’ countries headquarters. The app stores in the United States and China offer predominantly local apps, meaning that the developers of the top applications are largely centered in the country where they are ranked in the Top 50. Distinctively, the Netherlands is constructed as a more varied market of app distribution. While the American apps are still in the lead numerically, Dutch, Chinese, as well as apps developed by other European countries (e.g. France, Germany) contribute to the development of a diverse marketplace.
Shifting the focus on the two app stores in question, in all three cases (US, NL and China), Google Play has more countries represented in the Top 50 apps (16 US as opposed to 11 in Apple, 15 in the NL as opposed to 14 in Apple, 9 in China contrasting the 4 countries distinguished in Apple). As the e-giants have purposefully created a close interdependence between their developed operating systems and distinct devices, this has brought implications on both users and developers - the availability, the cost as well as the policies established create varied levels of use and a clear division between the two dominant companies.
Image 4: Headquarters of developers of top 50 apps in each country across platforms.
The findings showcase different degrees of diversity within each metric and country as well as within each country overall. Therefore, the three countries highlight different degrees of diversity through the represented apps within their two app stores, the represented categories of apps, the developers present and their headquarters country. Furthermore, we discovered four different factors influencing the metrics.
Firstly, based on the different measurements in our research, China is worthy as a unique case to be analyzed separately. The significant difference between Google Play Store and Apple App Store in China shed light on the current legal restrictions in the Chinese app store market. In the Chinese Apple App Store, there is no VPN software in the top 50 apps, which reflected the ban of the Chinese government since February 2018 (Haas). In contrast, in the Chinese Google Play store, VPN still accounts for a large portion of the top 50 apps. This could also explain that the other popular apps such as Facebook and Twitter, are those which cannot be downloaded due to the Great Firewall in China. One group of the highly arguable users would be the expats living in China, who still use those Western communication and entertainment softwares. Another perspective when looking at China, is the origin countries of the app developers. As indicated above, apps in the Chinese Apple app store are unique across the app stores as well as across the three countries. 48 apps herein have a chinese developer and consequently headquarter. The localisation in the Chinese Apple App Store indicates that, due to no access to foreign apps, Chinese developers have created their own to meet local demands. Most of the apps are named in Chinese characters, and only support Chinese language. Therefore, the popularity of the apps remain significant merely in China.
While the app market in the United States is dominated by local developers as well as local apps just like within the Apple App Store in China, the developers and apps themselves are diverse. However, the categories of the apps in the Top 50 within both the United States Google Play Store and Apple Store concentrate within on the Games section. The United States therefore, seem to overall showcase the least degree of diversity across all three countries in the combination of metrics analysed, as it has the least unique apps across app stores, category distribution, unique developers as well as a low diversity within the developers headquarters countries. In contrast, the Netherlands, display a high number of unique apps within both app stores from different developers. While the categories just as in the United States highlight a concentration within Games, the Netherlands are the most diverse with the apps developers origin, herein the developers present as well as the apps employed in the country.
Besides the general degrees of diversity according to the chosen and analysed metrics, we found four factors influencing the chosen metrics. The categories, besides the chinese Apple App Store, all showcase a concentration on one category such as Games (US, NL) or Tools (CH) seemingly highlighting a certain usage concentration within the app store on a specific kind of application.
Moreover, we discovered that in the developers’ distribution across countries according to their headquarters country of origin, a cultural proximity could be identified. Cultural proximity is oftentimes used to “explain media preference across national boundaries” (Ksiazek and Webster 2008). It comes into play as an important element which shapes the availability of culturally distinct apps on the two platforms in the US, the NL and China. Our data reveals that geographically proximate countries build upon similar markets. To exemplify, while the Chinese Google Play hosts Chinese, Japanese, Korean and Singaporean developers as well as apps designed in Hong Kong, the dutch Top 50 of the same platform include European developers and hereby a culturally proximal market (Germany, France, Belarus, Denmark, Finland, Sweden, Switzerland).
While we found a certain degree of diversity within apps across the three countries, the analysis of developers, as noted earlier, showed a media concentration. Similar to what Nieborg and Poell previously argued in discussing the dominance of internet platforms in the cultural production (15), the prominent presence of GAFA as developers is also apparent in our analysis across the three countries. Considering that the average frequency for a developer to be featured in the Top 50 on app stores across three countries is 1.95 times, the frequent appearance of developers from these media giants (Google; 16 times, Facebook; 17 times including daughter companies, Amazon; 6 times) is particularly significant. It is also worth noting that the subjects of our project – Google Play and App Store – are also run by Google and Apple. The Chinese app market, while it shows a distinctive diversity in apps compared to the other two countries due to the aforementioned governmental restrictions on the country’s media landscape, illustrates a similar tendency in terms of media concentration. Developers from Tencent and Alibaba, two Chinese media companies out of BATX appear 7 times and 5 times respectively in the Top 50 of two OS platforms, clearly showing a similar degree of dominance to previously highlighted Western media conglomerates.
Ultimately, the diversity of the two stores and hence of the apps within is deeply rooted in how the platforms function and develop within a social context. The wider availability of Google as opposed to Apple devices, its higher affordability which targets a more extensive developer base (for instance, developers are charged 99USD per year to contribute with apps to the Apple Store (Choosing a Membership - Support - Apple Developer) while Google Play (How to Use the Play Console - Play Console Help) charges a one-time fee of 25USD), the differences in policies and barrier to entry (e.g. Is it easy to contribute to? Does the app have stability? etc.) divert developers and consumers, depending on the socio-economic background and hereby lifestyles, towards one store or another, influencing the balance and perchance diversity of apps.
In conclusion, we found that different degrees of diversity can be identified with each metric but it is important to specify the metrics and their interpretation as entry points into the diversity of app stores such as Google Play Store and Apple App Store. Furthermore, within our country-specific research China has critalised itself as a unique case in terms of its diversity across the Apple App store and Google Play Store as well as across countries. Due to its distinct setting, as the Google Play Store is banned and the Apple App Store censored, it showcased the cultural significance of research into the diversity of app stores uncovering new questions for further research. Moreover, we realised throughout the analysis that other factors need to be taken into account influencing metrics such as cultural proximity, consistent concentrations across platforms and countries, media concentration as well as the socio-economic elements inscribed into the different app stores owned by Apple and Google influencing their availability. Therefore, while analysing the diversity of one metric within different countries or within different app stores, it is important to acknowledge these factors and their significance. Overall, we found that each country and app store showcases different levels of diversity in each metrics. Therefore, while a country or app store might be very diverse within a specific metric, it might be very uniform in another. Bringing all measures together can lead to a more complex understanding of diversity on the level of the app stores as well as the analysed countries.
In the future, it would be interesting to include more metrics such as the languages apps are offered in, also the data is only available for the Google Play Store right now, as well as the parent companies of developers listed. As we found a media concentration within the developers as is, it would be interesting to look further into their distribution. Along this line, we realised that actual developers of apps as individuals and producing companies and the companies stated as developers do not always align highlighting the owning companies. Therefore, another step might be to look into the diversity of the workforce within the employees creating the applications. Furthermore, China as a unique case has highlighted the possibility to use app store studies as an entry point for cultural studies into for example cultural proximity of app markets or the circumvention of censorship. Lastly, studying the diversity across app stores more closely in relations to the restrictions and differences between them as seen in the socio-economic lifestyle they promote and expect from users, would allow another layer of research into the diversity of app stores possibly combining the study of app store diversity with their volatility and policy development.
Finally, while we managed to collect all the app rankings and conducted comparisons across two different app stores and three countries, there are two limitations of the research data. Firstly, as noticed above, all the products and services of Google are banned in China. Therefore, the data of Google Play store in China may not represent the majority of the population. Further research in Chinese app stores could shed light on the rankings of the domestic app stores. The other limitation is that, in the mobile phone version of Google Play store, there is none category for “top free apps”, but an overall ranking of all the apps. Thus, the rankings of top 50 Google Play free apps we retrieved in App Annie would have been edited by the website itself. However, in order to keep consistent with the data in both app stores, we kept using the data of “top free apps” in App Annie.
- App Annie - The App Analytics and App Data Industry Standard. https://www.appannie.com/en/. Accessed 15 Jan. 2019.
- Categories and Discoverability - App Store - Apple Developer. https://developer.apple.com/app-store/categories/. Accessed 16 Jan. 2019.
- Choosing a Membership - Support - Apple Developer. https://developer.apple.com/support/compare-memberships/. Accessed 16 Jan. 2019.
- de Looper, Christian. The Supreme Court Could Rule That Apple’s App Store Is A Monopoly | Digital Trends. https://www.digitaltrends.com/mobile/apple-v-pepper-news/. Accessed 16 Jan. 2019.
- Haas, Benjamin. “China Moves to Block Internet VPNs from 2018.” The Guardian, Guardian News and Media, 11 July 2017. 14 January 2019. < www.theguardian.com/world/2017/jul/11/china-moves-to-block-internet-vpns-from-2018>.
- How to Use the Play Console - Play Console Help. https://support.google.com/googleplay/android-developer/answer/6112435?hl=en. Accessed 16 Jan. 2019.
- Ifrach, Bar, and Ramesh Johari. ‘Content on the Go: The Economics of the Market for Mobile Apps’. Media X at Standford University, no. Fall 2013, 2013, p. 25. Zotero, https://mediax.stanford.edu/wp-content/uploads/2018/09/Johari_F2013.pdf.
- Ksiazek, Thomas B., and James G. Webster. “Cultural Proximity and Audience Behavior: The Role of Language in Patterns of Polarization and Multicultural Fluency.” Journal of Broadcasting & Electronic Media, vol. 52, no. 3, 2008, pp. 485–503., doi:10.1080/08838150802205876.
- Michael Dieter et al. Store, Interface, Package, Connection Methods and Propositions for Multi-Situated App Studies. Vol. No. 4, Collaborative Research Center 1187 Media of Cooperation, https://www001.zimt.uni-siegen.de/ojs/index.php/wps1187/article/view/29. Accessed 16 Jan. 2019.
- Moazed, Alex, and Nick Johnson. Modern Monopolies: What It Takes to Dominate the 21st-Century Economy. St. Martin's Press, 2016.
- Napoli, P. M. ‘Deconstructing the Diversity Principle’. Journal of Communication, vol. 49, no. 4, 1999, pp. 7–34. Wiley Online Library, doi:10.1111/j.1460-2466.1999.tb02815.x.
- Nieborg, David B., and Thomas Poell. “The Platformization of Cultural Production: Theorizing the Contingent Cultural Commodity.” New Media & Society, vol. 20, no. 11, Nov. 2018, pp. 4275–4292, doi:10.1177/1461444818769694.
- Statista. “App Stores: Number of Apps in Leading App Stores 2018.” Statista. 14 January 2019. <www.statista.com/statistics/276623/number-of-apps-available-in-leading-app-stores/>.
- Select a Category for Your App or Game - Play Console Help. https://support.google.com/googleplay/android-developer/answer/113475?hl=en. Accessed 16 Jan. 2019.
- Weltevrede, Esther, and Erik Borra. ‘Platform Affordances and Data Practices: The Value of Dispute on Wikipedia’. Big Data & Society, vol. 1–16, no. January–June 2016, 2016, p. 16, doi:10.1177/2053951716653418.