In my previous post, I discussed how I set up house ads for all of my apps. Today I give an overview of the results from doing that that.
Let's cover some of the statistics on the installed apps I currently have in the Amazon App Store and in Google Play. For these values, I'm listing the Amazon App Store's "Total" report which indicates the number of times an app has been downloaded -- but does not necessarily mean it is still installed on a device. For Google Play, I'm listing the "Total User Installs" report which reflects the same thing. Google Play further reports active device installs which more accurately reflects how many installs of an app are still active (ie. not uninstalled). However, since I'm trying to compare apples to apples and the fact that Amazon doesn't report this kind of statistic, I'm leaving that out.
|Game Name||Amazon App Store total||Google Play total user installs|
|DB42 Full Version||29||147|
|DB42 Lite Version||1365||18939|
You can easily see that the installs for MatchCard and Pyromancer's Laboratory on the Amazon App Store is roughly twice what they are on Google Play. All my other apps have more installs from Google Play.
I ran my Admob house ads from Saturday morning (9/1) into the US Labor Day holiday (9/3), totaling a little less than three full days. If you read the previous post, you'll see how I divided things up on a per-app and per-app store (Amazon App Store vs. Google Play) basis. Below are the impressions for each one of those, grouped by app store. The "Ad Group" column defines the app the ad is promoting, the impressions indicates the number of times that ad was seen, and the clicks indicates how many people actually clicked on the ad itself (which, in turn opened up either the Google Play or Amazon App Store page):
|DB42 Full Version||233||5|
|DB42 Lite Version||311||5|
|DB42 Full Version||2763||6|
|DB42 Lite Version||2867||9|
So, what can we extrapolate from this? The first thing is fairly obvious. The amount of ads being served to the Kindle Fire are GREATLY outweighing the number of ads that are being served to all other Android devices. I'm speculating here, but I believe the reason may be the demographics of the audience of the various devices. I think the majority of Kindle Fire users are probably just "regular users". They bought a Kindle Fire and use it "as-is" -- they have not hacked it to run any other ROMs and probably don't plan to. Because of this, ads served to the Kindle Fire have less of a chance of being blocked. Other Android-based device users, on the other hand, probably have more of an interest in being power-users, have loaded custom ROMs, and have ad-blocking software installed on their devices.
Another obvious data point from all this is that the MatchCard game seems to be really popular. If you look at the number of impressions served up, you see that MatchCard has the FEWEST impressions. This is because I don't serve MatchCard ads within MatchCard itself (why advertise for the game when it is already installed???) To somewhat reinforce my hunch about ad-blocking above, consider the number of MatchCard installs. On Amazon I'm sitting at ~7000 and on Google Play I'm sitting around ~3000, that's roughly a factor of 2x. However, look at the impressions. On the Kindle Fire, they come in at around 2800 impression. On all other devices, they come in around 300 or so. I'd expect that the impressions for all other devices would be roughly 50% of what is on Amazon (figure 1500 range), but it's more like 10%. Interesting.
Not only are there differences in impression rates, but click-throughs do not seem to be proportional to the number of impressions across markets. Google Play ads weigh in at 2052 impressions with 45 clicks on those. This is a 2.2% click through rate. Compare that to the Amazon App Store: 17114 impressions at 72 clicks comes out to be a 0.4% click through rate. So, even though there are more impressions being served to the Kindle Fire crowd, they're a bit more stingy when it comes to clicking on the ads themselves. If things had been proportional, I would have expected nearly 400 clicks for the Kindle Fire users.
I wish there was some big take-away from all of this, but I really need to gather more data before I can start making assumptions or start seeing any kind of useful trends here. Hopefully these comparisons help other developers out there looking to use Admob with either or both of these two app markets.