Part III: Store Policies and Developer Conditions

Part of Apps and Their Stories: Volatility, Diversity, Policy

Team Members

Participants: Michael Dieter, Mike Huntemann, Gemma Newlands, Anna Teresova (alphabetical)

Lead: Anne Helmond, David Nieborg, Fernando van der Vlist, Esther Weltevrede (alphabetical)

Contents

Summary of Key Findings

Studying Apple App Store Review Guidelines and Google Play Developer Policy Center, we developed three different approaches in order to contextualise the amount of data we had collected during our research. Firstly, we focused on the word counts, which showed us to what extent the documents have been getting longer over time, currently having about 11.000 words respectively. Secondly, we visualised the formatting and the textual changes within the policy documents, which revealed when the documents have been most edited, highlighting the overall evolution of the app stores' policies. However, in order to focus only on the textual additions to the texts we also compared the tables of contents, which helped us to explore how the documents have responded to the new changes in law, such as COPPA and GDPR. Additionally, we close-read the sections 'Introduction' and 'Conclusion' in Apple App Store Review Guidelines from 2010 to 2019, which showed Apple's emphasis on the user experience over the developer experience and it highlighted how Apple gradually removed the notion of the Review Board from its document.

1. Introduction

As the global downloads across iOS and Google Play surpassed 113 billion in 2018, with the average smartphone user in the United States spending nearly 3 hours each day in apps (App Annie), the study of apps offers more data for an analysis than ever.

With the project 'Apps and Their Stories' being divided into three sub-projects - volatility, diversity and policy of applications - we decided to focus on policies which the app developers need to follow in order to get their app successfully released on either Apple App Store or Google Play. Through this project, our main goal was to explore the evolution of these policies and consequently contextualise these changes. As Apple states, 'apps are changing the world, enriching people’s lives, and enabling developers (...) to innovate like never before' (App Store Review Guidelines, 2019), with Google Play adding that with this innovation, however, comes 'responsibility' (Developer Policy Center, 2019). Doing both quantitative and qualitative research, we aimed to look into how Apple App Store and Google Play set boundaries in order to ensure that the developers remain 'responsible' when designing the apps.

2. Historical Background

An app is an abbreviation for any application program, at least from a programmer's point of view (Fagerjord, p. 93). However, within the context of everyday life, an app is a small program for a mobile device, downloaded from a central distributor, an app store (Ibid). In comparison with a Web browser, an app is a short-cut that guarantees direct and immediate access to the information stored in a database on the cloud (Matviyenko, pp. xvii-xix). Since it is much easier to access services on mobile devices from apps rather than from browsers, the demand for effective apps has been steadily growing as every user needs many apps to get the most out of his mobile device (Fagerjord, p. 91).

The industry of mobile apps significantly changed between 2006 and 2008 with the introduction of the first wave of smartphones and the possibility of broadband connections (Feijoo et al, p. 213). When Apple consequently introduced a new iPhone in late 2007, combining touch screen, motion sensor, enhanced display etc, the app market was changed forever (Ibid). The role of developers, however, was still rather limited as it was only possible to develop application within the platform (Ibid). But this evolved between 2008 and 2011 when the developers became the main engine in mobile software innovation, with the suppliers, platform and network operators competing to build strong developer communities ever since as their success is measured in terms of the number of applications and downloads (Ibid). On July 10, 2008, Apple introduced its own App Store that contained not only 'first-party' applications developed by Apple, but also apps published by third-party developers (Dieter et al, p. 1). Google reacted by launching Android Market on October 22, 2008 - later being rebranded as Google Play - and offered apps designed for the Android operating system (Ibid).

In 2010, Anderson and Wolff declared the death of the World Wide Web as the apps, offering simpler, sleeker services, were more about 'the getting' rather than 'the searching' and thus attracting more and more users (Anderson and Wolff, 2010). Indeed, since its beginnings in 2008, the app economy has become a billion-dollar global industry (Dieter et al, p. 14). Apps are so thoroughly insinuated into everyday life, they are often imperceptible: we seamlessly chat, take pictures, listen to music, play games, check our bank balance, and so on, without any moment of reflection (Ibid). The increasing demand for apps has consequently lead not only to more creative ways in developing them, but also to new policies which developers need to follow in order to get their apps released.

Many scholars have published articles on the app phenomen, focusing on topics such as policy and mobile advertising (Evans, 2016), separately examining the Google Play developer policy (Kywe et al., 2016) and the Apple App Store policy (Hestres, 2013), or comparing platform ecosystems of Google and Microsoft (Angeren et al., 2016). However, none of the studies has directly focused on the developer policies of Google Play and Apple App Store and has not contextualised their evolution from their early beginnings in 2008 up until now, to 2019 respectively.

3. Initial Data Sets

We were provided with the latest versions of App Store Review Guidelines and of Google Play Developer Policy Center. From there we conducted our research using Wayback Machine in order to explore how the developer policies have changed over time in both Apple App Store and Google Play. We established, scrapped and collected all the data between January 08-10, 2019.

4. Research Questions

• What are the changes in developer policies over time and how often do they take place?
• What is the ecosystem of developer policies (such as its structuring and length, tables of contents, what are the focus points of the app stores)?
• Do the guidelines put more emphasis on the developer or the user experience?

We hypothesized that Apple App Store would have more regulations regarding the process of developing an app, while Google Play would present itself rather 'moderate' with its policy documents.

We also believed that the length of the documents would change over time as more regulations have been added, including the latest changes in law, specifically with the EU General Data Protection Regulation (EUGDPR) from May 2018 (EUGDPR Information Portal, 2019). However, we were unsure to what extent the policy documents would potentially get denser.

5. Methodology

5.1. Establishing and collecting the material

• Firstly, we had to establish the different URLs of all the available versions of the guidelines, both the current and the preceding. We did so with the help of the Internet Archive Wayback Machine.
• When it came to Apple Developer Policy, through Wayback Machine we retrieved the earliest version dating back to September 3, 2014. Before then, Apple did not provide an open access to its developer policy and one was able to read it only with a valid developer ID (Apple Statement, 2010). However, Apple released the App Store Review Guidelines to public in 2010 (Ibid), which we accessed via Wayback Machine, dating back to September 11, 2010.
• It was more complicated when collecting the Google Play's URLs as the Wayback Machine retrieved the previous versions only back to July 27, 2016. However, we traced the developer policy guidelines back to October 21, 2008, when it was titled 'Android Market'. In order to further enrich the range of our material, we also included 'Android Developer Policy' that covered the policy back to March 09, 2012. Combining these three different URLs, all referring to the developer policy of Google app store from 2008 up until now, we were able to study the majority of the available material.
• Once we established the URLs, we input them into the Wayback Machine again. We then used the Internet Archive Wayback Machine Link Ripper tool to retrieve the links to the URL's archived versions at the Wayback Machine. We opted to scrape only one result per month because otherwise we would have potentially ended up with too many results to analyse.
• After saving the results as CSV files, we downloaded the archived web pages to our computers with a command line script created by Erik Borra. This gave us the developer policy documents from both Apple App Store and Google Play from 2008, and 2010 respectively, to the beginning of January 2019 in form of .html files, ready for further analysis.

5.2. Comparing the versions and curating the collected material

• Although we scraped only one result per month, we still retrieved a lot of material which needed to be curated before being analysed. We thus compared all of the collected documents on DiffChecker, which visualised the differences between them.
• For example, we started with Apple, comparing the policy document from September 2014 with the one from October 2014. Since there were changes between the documents, we kept both of them and then repeated the same with October 2014 and with November 2014. As the November policy was showing no differences compare to its preceding version, it was eliminated from the list, and so on. We manually noted down the changes between the documents and created a new list with the developer policies that differentiated from each other.
• This method provided us with one curated list for each app store, featuring 14 Apple App Store policy documents and 10 Google Play policy documents to analyse.

5.3. Further analysis

5.3.1. Length of the policy documents

• We noted down the word count of each policy document from our curated lists in order to see the changing length of the texts over time as well as to compare the length of Apple App Store Review Guidelines with Google Play Developer Policy (see Appendix 5 and 6).

5.3.2. The formatting and textual changes within the documents over time

• We took screenshots of the side bars of the DiffChecker which show the formatting and textual changes between the two input documents - for example, comparing the Apple policy documents from September and October 2014 showed 1 removal and 29 new additions.
• We then visualised all of these changes in two graphs (see Appendix 3 and 4).

5.3.3. Table of contents

• In order to highlight the focus points of both Google Play and Apple App Store in their policy documents, we compared the tables of contents and the changes within them.
• We manually retrieved the tables of contents from the curated lists of the policy documents of both Google Play and Apple App Store.
• We inserted them into an Excel Sheet and colour-coded the changes, visualising the changing focus points of the app stores (see Appendix 7 and 8).

5.3.4. Textual analysis

• We close-read and compared the introductions and conclusions in Apple App Store Review Guidelines in order to determine whether the store puts more emphasis on the user or the developer experience. We also paid attention to the changes in notion of repeal policy (see Appendix 9 and 10).

6. Findings

6.1. The length of the compared documents over time

• The overview of the word counts from both Apple App Store and Google Play revealed that the policy documents were getting longer over time (Appendix 5 and 6). Whereas at the beginning Apple had a longer and more specific policy document with over 3.000 words (Appendix 5), Google had significantly shorter version which was factual but rather simplified, with less than 1.000 words (Appendix 6).
• Comparing the latest policy documents of Google Play and Apple App Store, they are similarly long, with both having over 11.000 words (Ibid). While the Google Play documentations have a rather steady grow in its word counts, Apple App Store has a sudden jump in May 2018 when it adds about 2.000 words (Appendix 5), suggesting a potential correlation between these changes with the revision of the EU GDPR law.

6.2. The formatting and textual changes within the documents over time

• Overall, Apple has been having more formatting and textual changes in their policy documents compare to Google Play (Appendix 3 and 4). Although the one in September 2014 was mainly about adding double spaces to the text, in 2016 Apple completely reformatted their policy document, including a different outlay of the table of contents as well as of the sections (Appendix 3). In 2017 it also added further subcategories, which is visible in the graphical visualisation (Ibid).
• Google Play made major changes in July 2017, adding sections such as 'Spam', 'Permissions', 'App Visibility' or 'Content Ratings' (Appendix 4 and 8). Although the visualisation from January 2018 suggests significant modifications as well, these were mainly about formatting. However, in July 2018 Google Play added further sections like 'Daily Fantasy Sports' or 'Enabling Dishonest Behavior' (Ibid).

6.3. Table of contents

• The tables of contents helped to reveal the density of the policy documents of both app stores as well as the changing focus points of the app stores over time.
• Both Google Play and Apple changed the outlines of their tables of contents, with Apple in 2016 and Google Play in 2017 (Appendix 7 and 8). The tables also got longer and more detailed over time, adding sections such as 'gambling', 'child-directed apps' or 'data privacy'.
• Through these comparisons we saw that, for example, Apple introduced a new subcategory titled 'Data Security' in the same month when the EU released new General Data Protection Regulation, in May 2018 respectively (Appendix 7, and EUGDPR Information Portal, 2019). Also, in July 2017, Google Play enlarged a subcategory by adding a new paragraph on 'child-directed apps' and in the upcoming months required that the developers would determine whether their apps were primarily child-directed or not (Google Play, Updates and Other Resources). As a result, many apps were rejected because they did not comply with 'The Children's Online Privacy Protection Act' (Welch, 2018), suggesting that Google Play was responding to the increasing number of young children downloading apps (Nielsen’s Mobile Insights Survey, 2017).

6.4. Relation between the OS releases and the updates in policy documents

• We compared the timelines of changes with the dates of OS releases of both Android OS and iOS (Appendix 1-4). The changes in Apple App Store Review Guidelines did correlate with the releases of Apple iOS. For example, in September 2014 the Review Guideline added new sections on HealthKit, HomeKit and Passbook, alongside the new iOS 8 being introduced the same month and launching new apps with the same names (Appendix 1 and 3). The new iOS also featured Apple Pay, which was added to the Review Guideline in October 2014 (Ibid).
• However, comparing the Google Play timelines with the releases of Android OS, we did not detect any significant connections between the changes (Appendix 2 and 4).

6.5. The evolution of sections 'Introduction' and 'Conclusion' in App Store Review Guidelines

• Comparing the changes in sections 'Introduction' and 'Conclusion' in App Store Review Guidelines between 2016 and 2017 revealed that Apple gradually stopped referring to the policy as a 'living document' (Appendix 9). Moreover, we found that Apple has been putting a significant emphasis on the user experience. For example, in May 2016 it stated that '[they are] committed to [their] users and making sure they have a quality experience with [Apple's] products'; repeating in April 2017 that the app needs to look great rather than just 'cobbled together in a few days' (Ibid).
• The notion of 'Review Board' is also replaced by 'Resolution Center'. Thus whereas in 2014 the developer can find a section saying that 'if [his] App is rejected, [Apple] ha[s] a Review Board that [the developer] can appeal to', this bit is removed in 2016, and the developer is asked to communicate with the Resolution Center instead (Ibid).
• Overall, comparing the 'Introductions' and 'Conclusions' from May and August 2016, Apple is no longer 'pleased' with the developers coming to them but rather puts emphasis on the importance of apps and users, setting a clear tone of what they are focused on - the user experience (Appendix 10).

7. Discussion

We were not surprised to find out that over time the developer policies have been getting longer, now being established on more than 11.000 words per document for both Apple App Store and Google Play. However, it was interesting to see that although now both of the documents are similarly extensive, that was not always the case. In 2010, Apple's policy was significantly longer, having about 3.000 words, while back then Android Market had less than 1.000 words and was not so detailed. The structuring of Apple's policy was clear, including a table of content which helped to find specific sections. Google, on the other hand, was so short that it did not even introduce its table of content until much later. Therefore looking at these differences in the early beginnings of the app stores, we conclude that Apple realised the importance of developer policy earlier on, being very specific and even more restrictive compare to Google Play, which started putting the same emphasis on developer policy significantly later.

Moreover, in the case of Apple App Store, we detected changes within the policy documents that took place around the same time as new releases of Apple's iOS systems. We thus deduce that Apple's App Store is closely tied to Apple's iOS. However, doing the same comparacion with Google Play, we saw almost the opposite, with Google Play adding new changes to its developer policy documents independently on the launches of its operational systems. In our opinion, Google's app store then functions more as an independent app store compare to Apple App Store, which is partially shaped by the releases of Apple's other products.

Looking at the formatting and textual changes within the documents over time, we assume that Apple generally updates its policy more often than Google Play. Further comparing the tables of contents of both Google Play and Apple App Store, we contextualised the specific changes with the modifications of other laws outside the companies - such as the revision of the GDPR; or we linked them with the increasing number of young children downloading apps - like it was the case of Google Play in July 2017. This suggests that compare to the early beginnings from 2008 to 2011, when the policy documents were more focused on the demands from the specific app stores on the developers, recently the documents have been also subjects to the influences and changes from outside.

Lastly, we were surprised to learn to what extent Apple puts emphasis on the user experience over the developer experience in their policy documents. Whereas in the beginning, Apple started its document by saying how pleased they are that the developers want to invest their talent to develop an application for iOS, it gradually started highlighting the importance of the user experience, demanding the developers to deliver an app that would surprise, delight and excite the users. Although Google Play also asks the developers to create safe apps, their policy document is more focused on the developers rather than the app users. This suggests that overall, Apple focuses more on users and usability while Google is aware of the importance of developers and consequently pays more attention to focusing on them in the policy documents.

8. Conclusions

Our research used digital methods in order to reflect on the evolution of changes within the policy documents of both Apple App Store (Apple App Store Review Guidelines) and Google Play (Google Play Developer Policy Center). We contextualised the collected data and showed how the app stores have been reacting not only to changes within the companies but also to new laws outside of them. Whereas in 2008 the app stores were slowly navigating themselves through the new field of market, more than ten years later they are aware of the increasing demand for apps and react to this in their policy documents addressed to the developers, who have the responsibility to design a safe but exciting app.

For future research, we would suggest to focus more in detail on specific sections within the policy documents and explore how they have changed over time - such as the evolution of gambling policy in Google Play or 'Human Subject Research' section in Apple App Store. Furthermore, it would be interesting to study the links from the documents in order to see how the app stores interlink the policy documents with other instructions for the developers and thus elaborate on the overall helpfulness and effectiveness of these documents.

9. References

• Anderson, Chris, and Wolff, Michael (2010). 'The Web is Dead. Long Live the Internet.' Wired. Retrieved from https://www.wired.com/2010/08/ff-webrip/.
• Angeren, Joeyvan, and Alves, Carina, and Jansen, Slinger (2016). 'Can We Ask You to Collaborate? Analyzing App Developer Relationships in Commercial Platform Ecosystems'. Journal of Systems and Software Volume 113. pp. 430-445.
• App Annie (2018). 'A Year in Review: Mobile Highlights of 2018'. App Annie. Retrieved from https://www.appannie.com/en/insights/market-data/a-year-in-review-mobile-highlights-of-2018/.
• App Store Review Guidelines (2019). Apple. Retrieved from https://developer.apple.com/app-store/review/guidelines/.
• Apple Statement (2010). 'Statement by Apple on App Store Review Guidelines'. Apple. Retrieved from https://www.apple.com/newsroom/2010/09/09Statement-by-Apple-on-App-Store-Review-Guidelines/.
• Developer Policy Center (2019). Google Play. Retrieved from https://play.google.com/about/developer-content-policy/#!?modal_active=none.
• Dieter, Michael , and Gerlitz, Carolin, and Helmond, Anne, and Tkacz, Nathaniel, and van der Vlist, Fernando, and Weltevrede, Esther (2018). 'Store, Interface, Package, Connection. Methods and propositions for multi-situated app studies.' Collaborative Research Center 1187 Media of Cooperation Working Paper Series No. 4. pp. 1-16.
• EUGDPR - The EU General Data Protection Regulation (2019). 'EUGPDR Information Portal'. EUGDPR. Retrieved from https://eugdpr.org .
• Evans, David (2016). 'Mobile Advertising: Economics, Evolution and Policy'. Global Economics Group; University College London. pp. 1-57.
• Fagerjord, Anders (2015). 'The Cloud, the Store, and Millions of Apps'. In I. Kaldrack & M. Leeker (Eds.). There Is No Software, There Are Just Services (pp. 91–102). Lüneburg, Germany: Meson Press.
• Feijoo, Claudio, and Gómez-Barroso, José-Luis , and Aguado, Juan-Miguel, and Ramos, Sergio (2012). 'Mobile Gaming: Industry Challenges and Policy Implications'. Telecommunications Policy 36. pp. 212–221.
• Google Play (2019). 'Updates and Other Resources, April 2017'. Google Play. Retrieved from https://play.google.com/about/updates-resources/updates/.
• Hestres, Luis (2013). 'App Neutrality: Apple’s App Store and Freedom of Expression Online'. International Journal of Communication, 7. pp. 1265–1280.
• Kywe, Su Mon et al. (2016). 'Dissecting developer policy violating apps: characterization and detection'. 11th International Conference on Malicious and Unwanted Software: “Know Your Enemy” (MALWARE). pp. 10-19. Retrieved from https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7888725.
• Matviyenko, Svitlana (2014). 'Introduction'. In Miller, P. D., & Matviyenko, S. (Eds.). The Imaginary App. pp. Xvii-xxxvi. Cambridge, MA: The MIT Press.
• Nielsen’s Mobile Insights Survey (2017). 'Mobile Kids: The Parent, The Child and The Smartphone.' Nielsen. Retrieved from https://www.nielsen.com/us/en/insights/news/2017/mobile-kids--the-parent-the-child-and-the-smartphone.html.
• Welch, Chris (2018). 'Google took down over 700,000 bad Android apps in 2017'. The Verge. Retrieved from https://www.theverge.com/2018/1/30/16951996/google-android-apps-removed-security-2017.

10. Appendices

Note: All the graphs disclosed below were designed by Sofia Chiarini based on our collected data.

10.1. Overview of iOS Releases

10.2. Overview of Android OS Releases

10.3. Overview of changes in Apple App Store Guidelines

10.4. Overview of changes in Google Play Developer Policy Centre

10.5. Word Count of Apple App Store Guidelines

10.6. Word Count of Google Play Developer Policy Centre

10.7. Apple App Store's Tables of Contents over time

10.8. Google Play's Tables of Contents over time

10.9. Apple: Overview of changes in the sections 'Introduction' and 'Closure' (which is referred to as 'Living Document or later as 'After You Submit')

10.10. Screenshots of changes in the sections 'Introduction' and 'Conclusion' (comparing May and August 2016)

10.10.1. Introduction

10.10.2. Conclusion