As part of my ongoing (quixotic?) quest to create a correlation between the number of user ratings for a Paid Game App on iTunes and its downloads, I've been doing some concept testing with sources and just spent way too much time building the spreadsheet above to guesstimate the revenue generated by the top 30 best selling games in the US store from iTunes Rewind 2009. I'm beginning to get more comfortable with the idea that in general each rating is worth about 75 downloads (as opposed to 100 that I've posited previously), or in other words, that 1.33% of game App purchasers submit a review. In looking at the spreadsheet you'll notice that I've had to tweak to the download to review ratio in order keep Apple's Rewind ranking intact... and these mostly make sense to me intuitively. Basically, what this shows, is that purchasers of more casual games (e.g. Bookworm, Radgdoll Blaster) are less likely to submit a review and those that buy more gamerly titles (e.g. Zenonia, Myst) are more likely to submit one. There are two anomalies here worth noting... one which I can explain, the other I cannot. You'll notice that the revenue associated with Rock Band seems be out of sequence, however this is an App that allows in-App purchasing, so there's a revenue stream that I'm not capturing. The one I can't figure out is Sally's Spa, which in this analysis has the characteristics of a hard core game title... which doesn't make any sense. There may be a point when titles get over a certain number of ratings (this one has 16,750) that a different paradigm kicks-in. I know some of you'll be thinking that I partied too much in college (and after?) and this is all evidence of some kind of long-lasting impairment... but if that's not the case, and this model can be refined a bit further, it should be a great benchmarking tool. Let me know what you think. I'd especially love to hear feedback from peeps who weren't sleeping during statistics and, of course, from publishers.
btw - I have noticed that in some instances Apps have lost ratings from previous versions despite iTunes claiming to have ratings from all versions of the product. This is the case here with Wheel of Fortune, for which I was able to estimate all ratings from data I had collected for an earlier post.