I still have some records left over, thinking about doing another sale, would be game to test another means of customer acquisition and report back with results.
I really wish you could apply multi-armed bandit approaches to physical problems like these more easily; obviously its hard to do A/B testing when you only have one actual supply.
There are certainly a lot of forums out there too, but most of them aren't local to one area, so most of the sales are online (I did a lot of business through these forums in the past).
In retrospect, it would have been rad to run this as a fully fledged A/B test. I honestly just didn't know that the ads would have quite the impact that they did.