Tuesday, August 6, 2013

Picking a banner for interstitial ads

I understand so little about marketing. It's clearly not my job, I have no experience and I definitely don't have enough time to put into it. Nevertheless, being an independent developer/publisher, I have to try.

I was looking today at some recent results of a cross promotion via Chartboost, and I thought I may share the experience.
The cross promotion consists in promoting Fractal Combat for Android in other games of ours, namely the free editions of Final Freeway (two variations) and Final Freeway 2R.

The results may be obvious to some, but here they are.
Considering the three alternative banners (or "creatives"):
Creative "A" (CTR: 22.32% IR: 0.68%)
Creative "B" (CTR: 24.77% IR: 1.57%)
Creative "C" (CTR: 21.95% IR: 0.55%)

We have the following data (higher is better):
  • A) CTR: 22.32% IR: 0.68%
  • B) CTR: 24.77% IR: 1.57%
  • C) CTR: 21.95% IR: 0.55%
CTR is slightly better for  "B",
the in-game screenshot.
(chart shows relative percentages)
CTR is the Click-Through Rate. Meaning the percentage of clicks received per impression (= ad view).
One creative gets almost to 25% CTR, meaning that the banner is clicked once every four impressions. Actually pretty good ! I think.
Interstitial ads lead to a much better CTR, because they practically take over the screen.
Here there isn't much of a difference between the three creatives. The screenshot-based creative ("B") does only slightly better.

IR is much better for "B",
the in-game screenshot.
(chart shows relative percentages)
IR is the Install Ratio. Meaning the percentage of installs / impressions, basically how many installs are obtained every 100 impressions.
Here "B" gets the lion share.
(Notice that Chartboost gives IR as percentage of installs / clicks. Both are useful things to look for, but in this case I just want to see the direct relation between impressions and installs).

So, the creative "B" clearly wins. One could argue that the illustration-based banner could be progressively improved. For example we already know that "A" performs better than "C", even though they are almost the same.
Still, "B" becomes the easy choice. The game already has good visuals, and an untouched screenshot seems to work well.
If anything, the screenshot where a flight combat scene is depicted leads to more likely users. In fact, while the CTR is only marginally better for "B", the actual IR is much better.
This means that although the screenshot-based banner is not terribly more attractive than the illustrations, the user that gets on the app store page via the screenshot is 2-3 times more likely to convert to an actual install.

The good news here is that if we keep on doing games with attractive in-game graphics, we won't have to put too much effort into making illustrations for marketing material.