Download numbers

When reading other's developers blog, I always appreciate reading the posts about the number of downloads. It is always useful to know how many downloads to expect, and what impact different events can have.

The mobile landscape is changing fast though, and while you can get a pretty good idea of the dynamics of the Apple Appstore,  there's not a lot of information regarding the Android Market (AM) and the new Amazon Appstore (AA). Here's my humble contribution (please note that all the numbers given are approximations).

PP on Android Market

During the first two weeks, PP was getting 800 downloads per day. However it got featured by Google, and the downloads increased a hundred folds (it got up to 80k downloads on one day). During the one or two weeks it was featured, it got about 900k downloads and entered the most downloaded free app list. I am guessing that currently, every free app in the top 10 are getting at least 40k downloads per day.

Right now PP is getting 3k downloads per day.

One metric that people rarely share is the number of active installations. It is particularly relevant if your app is advertisement based, because you want people to keep using (and thus having) your application for as long as possible. Currently, 37% of the people that downloaded PP kept it, which I think is pretty good considering it's a free application (I download a lot of free apps, but most of the time uninstall them quickly).

Active installations of PewPew.

PP2 on Android Market

Shortly after it reached 1 million downloads, I released PP2. Two days later, I updated PP to include a link/advertisement for PP2. I was getting around two hundreds downloads per day. PewPew 2 was then featured, and there was a 50% boost which is nice, but really nothing extraordinary.

If I recall correctly, with 400 downloads per day, PP2 was the number 5 in the top selling games category. Right now it is around #40, with around 50 downloads per day.

Active installations of PewPew 2. Currently the rate is 85%

PP and PP2 on the Amazon Appstore (aka the Useless Appstore)

While I was featured on the AM, Amazon contacted me so that they could get my application on their store. I happily obliged, having heard that if you were lucky enough to be selected for their "free application a day" program, you'd earn a boatload of cash. The idea was that Amazon selects one application per day, makes it free for their customers but still pay 20% of the application's price to the developer instead of the usual 70%.

PP2 got selected for the "free application a day" operation, but I actually did not earn anything because Amazon makes you sign a contract that cancels the one where the 20% are mentioned. In the end the operation translated into 100k downloads of PP2, but 0$ in my pocket. Some might say that it's worth it because of the free advertisement you get, but on the other hand you do loose quite a lot of potential sales. More importantly, I believe that giving out your products (even for a short amount of time) devaluates their value, but that's a topic that deserve a separate post.

The last thing I'll say about the AA is that in June, PP2 got 18 downloads...

Conclusion

Getting featured on the AM is awesome when you are a free app, advertising a payed app from a free app that is featured is a good idea, and it appears that the top games aren't raking in as much money as the top games on iOS.

Also, currently the AA is useless. If the rumors that Amazon is going to release its own tablets are true, that might change, but for now the AA is not interesting for developers.

Interview

Here's an interview I gave for droidgamers.com.

Among other things, I talk about which games influenced PewPew, what it's like developing for Android, what impact does getting featured on the Android Market has, and what I think about ad-supported applications.

Success and crashes

One day has passed, and PewPew is doing well. It has entered the Hot android app category on AppBrain.com, and I am receiving a lot of positive feedback. While monitoring twitter I found 14n3 saying "Oh man, making a review of PewPew, this game rocks!". Always gratifying.

Unfortunately, there are a few people reporting crashes.

Since PewPew is very stable on the iOS devices, the crashes must stem from Android specific code I added. There are two areas where I am not sure I am doing everything right:

It's either the handling of the application life cycle, specifically the part where I have to reload the openGL data when the game is resumed, or the http requests to upload and download the scores and replays.

Those thing would pose no problem if PewPew was a Java application like most Android applications, but it is 99% a C++ app with a thin Java wrapper (and thus uses JNI). To reduce the amount of JNI code I cut some corners.

At first I started using the Android Java API for the http requests, but then I realised it involved doing a boatload of JNI functions and threads, so I decided to use a C++ library. Curl and Boost are the most well known C++ libraries for that kind of job, but they are also completely overkill and would require too much time to integrate in PewPew. I searched for a light alternative and found HappyHTTP.

HappyHTTP does all what I need, and can be integrated into a project simply by adding a .h and .cpp file. The only problem is that it uses C++ exceptions. I can't use exceptions in PewPew even if I wanted to, because currently the Android NDK does not support exceptions handling and STL usage at the same time! I quickly hacked an exception-free version of HappyHTTP, but I may have messed up some things up, which may cause the crashes on some devices under certain conditions...

I hope I fix everything quickly so that I can resume doing the fun stuff: working on new levels and game modes.

PewPew 1.5 available on the Android Market

The title pretty much says it all. I uploaded PewPew to the market today, and it is now spreading across the internet.

Here's the link to download it for Android.

From what I've seen, PewPew is by far the best shooter on the Android Market. On my Nexus One, I'd say it runs at approximately 4 time the frame rate of the current most popular shooter, Gun Bros.

BTW, Gun Bros is an absolute piece of crap, gameplay and UI wise. It's sickening to see the work of good artists go to waste with game designers like that. Sorry for the rant, but some games just revolt me!

Now go play PewPew on Android and fill the empty leaderboards :)

PewPew Beta, for Android

Currently my work has shifted from developing PewPew to porting it to several different phone OS. I'll try to write a post later about the process of porting, because there's lot's of interesting bits I've discovered.

A few weeks ago I submitted a version of PewPew for the Bada OS to Samsung, but it got rejected because my website contained an image of an iPhone and a link to the App Store. At first I was surprised, but they simply don't want to have iPhone advertisement flaunted at their customers, which is understandable.

In the meanwhile I ported PewPew to Android and improved many things, and because I got rejected, those improvements will get to be in the initial Bada OS release as well (I try to look on the bright side of things).

Regarding the Android version, I don't know how well my app works on other phones, so I am starting a beta process. I want to get PewPew perfectly right before I submit it to the Android Market: I am a firm believer that first impressions counts.

If you want to test it, here is the .apk.

PewPew 1.4 and PewPew 2 1.2 submitted to the app store

I finally submitted to the App Store updated versions of PewPew and PewPew 2! Both versions now support the iPad's and the iPhone 4's resolution.

Most changes are in PewPew 2:

• There's an improved Chapter One: among the things I did was adding visual feedback showing destructible objects were reacting to the players' bullets, so the players would know those objects were destructible . More importantly, I've corrected the bug where the game denied the completion of a level when a player died just after completing the level's objectives, but before the transition back to the menu.
• There's the long awaited Chapter Two ! It took a really long time to create because it's containing a lot of new enemies, and the levels are more complex than in the first chapter.
• I now display the best times for the campaign levels.

• I've added a new infinity mode that you unlock by completing the chapter Two. This mode is very different from the existing ones since you actually control 2 ships. It took a while to adjust the difficulty progression because I am actually not a very good player at this game mode =). It's tough to adjust the gameplay for players that are going to be much better than you.

• I've slightly improved user interface. I've made the buttons prettier because I've always felt the old buttons (just a rectangle) were boring. Also I've removed the explosions in the backgrounds since they just weren't justified, and in my opinion reduced the purity of the menus.

I've also re-done the website. The client side is done in GWT while the server side is in python. One of the visible change is that the high scores now include the colors of the nicknames :-)

Now that I think about it I was quite productive. I really hope you all like the new stuff. My goal is to make PewPew 2 the best twin joystick game on the App Store, so I am going to keep on working on it.

Lack of updates because...

I am porting PewPew to the Bada OS. There is a contest going on, and I believe I have a real chance of winning a price ;-)