Since February, I’ve made it a habit to track R1 anime releases on US Amazon, hoping to get an idea of how sales figures look for specific US series. I recently finished up taking data for the month of June (available here).
Recently, that study seems to have begun bearing fruit. Combined with assumptions about the positions of their blu-ray top 20 versus the overall top 20, I was able to come up with a power law approximating how much a day ranking on US Amazon was worth in terms of disks sold per day (#=300,000/rank). Recently, I got my first substantial bit of proof that this prediction method was at least somewhat viable (to within +/-20%), and I’m at the point where I’m ready to put out very basic estimates of what first-week sales figures for various releases should look like for those series released in June.