The Worse Bug

Rare Earth is not profitable. Let’s get that out of the way up-front. The game was released as an experiment to see whether the mobile games platform would be receptive to an unfinished (but frequently-updated) game concept. In many ways, the experiment was a failure. Users were extremely divided about playing an unfinished product (almost all the reviews were either 5 stars or 1 star), and there were nowhere near enough users to generate enough impressions to make any money off the game. Or so I thought.

Here’s a graph of my impressions for June and much of July. Rare Earth is just one of the pack among the 3 ad units from BuildDown:

However, when I was finalizing the latest Rare Earth release (v0.6), I noticed an oversight on my part.  had programmed logic into my ad control wrapper to detect errors, wait a bit, and try to recreate the ad control, in an attempt to recover from any crashing issues in Microsoft’s ad control. Unfortunately, I had forgotten a critical component – to actually call my update logic! Ultimately, any ad crash caused the ad control to stop functioning until the game was started up anew. I had no idea I was missing out on potential impressions until I fixed the bug. Here’s the same chart, with another two weeks added in:

Well that’s interesting. Flurry tells me there hasn’t been any significant uptick in my userbase or number of sessions. Isolating for the variable, that means I have a 5-20x upgrade in my number of impressions just by compensating for errors in Microsoft’s ad control! Extrapolating backward, Rare Earth could have been sitting on hundreds of thousands more impressions over the two months it’s been live.

I’m sure there’s a lesson in there somewhere.