A sample size of more than a handful of sales per month (since the software cost "between $200 and $500") would make it more convincing. A randomized experimen

A sample size of more than a handful of sales per month (since the software cost "between $200 and $500") would make it more convincing. A randomized experiment would also be more convincing, and not hard to do.