Thread by @narseo

Can mobile users (and investors) trust Google Play statistics like app ratings, and app install counts assessing app quality, growth, user base, and popularity?

The answer is: "not really"

To know the reasons behind this answer, read the following thread:

At #IMC2020 we presented the 1st study on incentivise app install programs "running" on Google Play. We show how these schemas work, how money flows, and how they can successfully manipulate app ratings (and for potentially deceitful reasons)

This project started because we had some apps on the Play Store and we received many emails from companies offering us services to improve our app's presence and ratings on Google Play. We couldn't resist looking into it. We are curious minds.

Soon after we found that these practices are prohibited by Google Play's policies (support.google.com/googleplay/and…) Yet, many companies openly and broadly advertise these services so we decided to take a deeper look into this world.

This led us to "Offer Wall" apps, which are affiliate apps offering economic rewards to users for performing actions like installing an app, making in-app purchases, or interacting with it for some time. The schema below provides an overview of this ecosystem and its stakeholders

By monitoring these services for months, we could characterize their behaviour, find developers using such products, measure the monetary rewards offered to end users, and, finally, measure their impact on app store rankings.

While Pay-per-install services (PPI) were well-known mechanisms in the Desktop world to distribute malware, we did not found this type of abuse on Android as we initially expected. Instead, we found that incentivised installs seem to be driven by economic reasons in most cases.

We show that these services allow app developers to improve their app store ratings. Some apps even get listed in the top app ranks provided by Google Play. Note that apps' presence in top charts might organically attract more users, thus increasing their user base.

But some developers might abuse these schemes to (deceitfully) monetize their apps through in-app ads or obtain VC funding based on inflated user engagement or revenue metrics. Indeed, we identified some apps that successfully attracted VC funds after hiring these products.

Some apps (detailed in the paper) raised multiple rounds of funding after launching incentivized install campaigns. One particular app raised $30 million two weeks after launching the campaign, and $110 million two months after.

Does it sound familiar? :)

cnbc.com/2016/06/27/sil…

For more information, including some more anecdotes, you can find the full paper at: arxiv.org/pdf/2010.01497…

Or if you are short of time, watch @shehrozeFarooqi's IMC presentation video:
homepage.divms.uiowa.edu/~sfarooqi/File…

This paper was done by @AlvaroFeal and myself from @IMDEA_Networks in collaboration with @shehrozeFarooqi from @uiowa, @zubair_shafiq from @ucdavis, and T. Lauinger and D. McCoy from @nyuniversity.