One of the reasons people often go with Free-to-play with In App Purchases is to try and combat piracy. The hypothesis is that since the game is free, people won't pirate it. For example, Dead Trigger went free on Android due to a high piracy rate. I'd like to discuss what we've learned about piracy for our free-to-play iOS game Chip Chain.

First, lets look at how we know we actually have a piracy issue. Each day, Apple reports how many sales of IAP have been made. This is the standard - if Apple is paying us, it must be real. Then, we compare with what our own tracking server thinks has been purchased.

In our game, when we get back the purchase receipt from Apple, we then do receipt validation (but evidently we didn't do a good enough job on our validation). Once the purchase is "validated" we then track the purchase on our tracking servers. We can compare the numbers that we see on our tracker vs the numbers that Apple reports: we use the formula (#tracked - #reported_by_apple)/(#tracked) to calculate the percentage of tracked IAP that is due to piracy. We get the following results for our Unlimited Gems IAP, selecting our top 4 markets. For more background, you can read my previous blog post about the importance of segmenting your users by country.

Country     Unlimited Gems Piracy
---------------------------------
UK          0%
US          10%
Japan       13%
China       91%

This method isn't perfect, because it looks like sometimes our tracker misses events, as it doesn't make any sense to have a negative piracy rate unless our tracker isn't getting every event. Here is what the data looks like for our Small Gems Pack IAP.

Country     Small Gems Piracy
-----------------------------
UK          -1%
US          -4%
Japan       -3%
China       50%

And overall, if you do this on a revenue basis, where we look at (revenue_tracked - reveue_reported_by_apple)/(revenue_tracked), we get

Country     Revenue Piracy
--------------------------
UK          -5%
US          3%
Japan       5%
China       86%

Again, this method isn't foolproof because it doesn't make any sense to have a -5% piracy rate, so we've got to have some error in our numbers. But we can clearly see that piracy is a much bigger problem with our Chinese players than our US players. So we can then dive a little deeper and look at the differences between how US and Chinese customers behave.

Keep in mind that the reason I'm singling out the US and China is they are the two biggest markets for Chip Chain, and that's the finest segment we can do with Apple's reporting data. I'm not making an ethical judgement on either country or its people.

One thing we can easily do is look at the time between when they first install the application, and make their first IAP purchase.

Time it takes for IAP Users to Download the IAP
-----------------------------------------------
Time      Percent Users US   Percent Users China
>1d       59.66%             27.81%
1h-1d     23.92%             14.41%
15m-1h    8.16%              6.87%
5m-15m    3.94%              13.23%
1m-5m     2.44%              13.07%
<60s      1.88%              24.62%

It seems like it's a safe bet that if you make a IAP purchase in less than 60 seconds from downloading the game, you probably are pirating the IAP. It seems like you could even draw the line around 15m based on this data, accompanied by the earlier data that shows Chinese players are much more likely to pirate the IAP. Now this doesn't generalize to all games…Chip Chain is pretty quiet about pushing IAP, and I expect this time threshold is pretty different depending on your game design.

If you were really interested in stoping pirates, one conclusion you might come up with would be to just not provide a way for people to download the IAP for the first 5 minutes. The users who just barge in, and try to take whatever they can find for free without even bothering playing the game, might be stopped if there isn't a way to even get into the IAP shop until you beat the tutorial and play a couple games.

But, just as a final note, this doesn't stop everyone. Some people wait a while before pirating. Look at this one user:

IAP            Time after Install
-------------------------------
small_gems     1 day, 4:30:18
small_gems     1 day, 10:05:24
small_gems     8 days, 5:49:25
small_gems     8 days, 5:49:34
large_gems     8 days, 10:06:51
large_gems     20 days, 8:11:56
large_gems     20 days, 8:12:01
large_gems     20 days, 8:12:05
large_gems     20 days, 8:12:09
large_gems     20 days, 8:12:12
large_gems     20 days, 8:12:16

It's exceedingly likely he's not really buying the game, because the large gems pack currently costs $4.99, and you could just buy the unlimited one for $9.99. And notice how the user buys them 4 seconds apart at the end? I'm not actually angry at this guy for stealing the IAP. I'm happy that he likes the game enough that he's willing to stick around and play it, and perhaps share it with his friends.

But then again, I designed a game that doesn't have server costs that scale with the number of users or plays. If you have costs per user, piracy can be deadly, as the Battle Dungeon sadly found out.

My personal belief is that indies can't effectively fight piracy head on…I think you need a very dedicated team that is constantly fighting against it, and with a lot of server side support. It can be quite expensive and it's likely futile anyway. I do think there is an argument that pirates can actually help spread word of mouth for your game, so it's not completely clear that they are exclusively bad if you don't have real costs per user. For example, Rovio has said that Piracy Isn't Such a Bad Thing.

What I do think is important for indies is to not make business or design decisions based on analytics without knowing the difference between your pirating users, your paying users, and your non-paying users…so that you don't pollute your statistics with data from people that aren't really your customers. It's pretty difficult to know exactly who is a paying user, and who isn't (if you knew, you'd be able to stop them from pirating), but we can use some statistics to get a sense of it.