A Global Piracy Heat Map Print E-mail



With the release of Google's License Verification Library (LVL) and the subsequent cracking of applications that make use of LVL, the Android development community has been buzzing.  The Android Developers Google Group has several active threads discussing how LVL can be improved and what the best approach to solving piracy on the platform might be.

One Google engineer made an interesting statement in response to my claim that I thought a "culture of piracy" might be developing in the Android user community:

"If you are saying that [there's a culture of piracy] because you think most people are pirating Android apps...  I think your perception of things is probably pretty off.  I know lots of people who have Android devices, and none of them even think of turning on the option to install from external sources, let alone go out and find pirated apps." 

That's actually a really good point.  I don't know anyone that pirates apps.  The statement raised a very interesting question in my mind:  Who actually is pirating Android apps?  In my experience for my apps such as Screebl Pro, I've seen piracy rates that are very high (as I described in this blog posting). Things had gotten so bad, in fact, that we actually developed our own license verification solution some months before LVL in an attempt to stop piracy of our applications. 

But the question remains.  If I'm not pirating apps, and you aren't, and no one knows anyone who is, where is the piracy coming from?  In the case of my app, I thought it would be interesting to try and identify which parts of the world were spawning the most pirates.

In doing this I figured that I could test one of the most prevelant claims around piracy on Android, namely:  "Piracy comes primarily from countries that do not have an option for purchasing apps legitimately on the Android Market."

The image at the top of this post depicts a "piracy heat map", showing the percentage of installations that are pirated.

Here's what I did.  First I collected anonymous Flurry stats from my application for a 90 day span.  I also collected Google Checkout records for the same time span.  Both sources include coarse-grained location information at the country level as to where the purchase or installation was made.  My assumptions:

  1. Google Checkout records should be a subset of Flurry records.   All app installs whether purchased or pirated are registered with Flurry, but only purchases are registered with Google Checkout.
  2. Given assumption #1, I define the number of pirated installs as Flurry Count - Google Checkout Count.
  3. I counted any engagement with Google Checkout as a "legitimate purchase".  Even those that were canceled by the user or Google (for insufficient funds, etc.)
  4. Google Checkout "Buyer Country" property represent the country from which the app was purchased.  It's possible that this isn't true 100% of the time, and indeed there were checkout records that showed purchases from countries such as Kuwait, Nigeria, and Ukraine, but the numbers where very low.

With these assumptions in hand some very interesting observations can be made from the raw data.  Here's the overview.

Over the course of 90 days, the app was installed a total of 8,659 times.  Of those installations only 2,831 were legitimate purchases, representing an overall piracy rate of over 67%. For my app, the largest contributor to piracy, by far, is the United States providing 4,054 or about 70% of all pirated installations of Screebl Pro.  Since the U.S. has supported paid apps from the very beginning, this would tend to argue against the idea that piracy comes primarily from countries that don't support the sale of apps. The other biggest contributors include Australia, Germany, and the U.K., also countries that support the sale of apps in the Android Market.  In fact, of the total 5,828 pirated installs only 844 or about 14% came from countries that don't support the purchase of apps in the Android Market.  The Piracy Contributors figure below represents the distribution of pirated installs by country.  The category "Other (Paid Apps)" represents all countries (other than those listed explicitly) in which users can purchase paid apps from the Android Market.  "Other (No Paid Apps)" covers countries in which apps can not be purchased.



The two bar graphs below represent the piracy rates for the thirteen countries in which apps may be purchased from the Android Market.  Notice that in no country other than Japan are more than 50% of the installs of my app actually purchased (more on Japan in a minute).  In fact, the average piracy rate in these countries is 66%.  Two out of every three installs is pirated in markets where the user could purchase the app!  To me, this is starting to look more and more like a "culture of piracy", and not just in countries where the user can't purchase due to Google restrictions.

An interesting exception is Japan.  I was doubtful of the numbers when I first saw the results.  In fact, there were some anomalies -- Flurry reported slightly less installs than Google Checkout had records for (about 2% lower), which violates assumption #1 described above.  I'm tacking this up to the Flurry Agent not being as good at reporting installs in Japan as it is for other parts of the world.  According to this data, 100% of installs of my app are purchased legitimately in Japan.  Wow!  If there's a lesson to take away here, it is localize your app for Japan!  I will be interested in hearing people's theories on why piracy is so low for that locale... 

But what about those "no paid apps" countries?  Well it does appear that piracy is significantly higher as a percentage of total installs in those geographical areas where apps can't be purchased, as one would expect.  In fact it should be 100%, but the average was more like 99%.  That difference is likely due to some kind of error in my assumptions.


So there you have it, a global view of piracy for one Android app.  I found this information very enlightening, and it dispelled several assumptions that I had been holding about what motivates piracy and who is actually engaging in it.

I'm interested to hear your thoughts...