At USV product summit last week we had an interesting discussion about pricing and how pricing is not something you could A/B test like other features of your product. Because if you do, the word will get out and user will complain and you'll have some explaining to do. In addition, it might be against the regulation in most countries to charge different price for your product just because user fell in a different testing bucket. But if you're operating in many countries, like most USV portfolio companies are, you can use that to your advantage and do experiments with pricing in one country first, before applying a pricing change worldwide. Due to varying purchasing parity, currencies used, and other cultural and socioeconomic issues between populations of different countries, these are not randomized tests and therefore not proper A/B tests. Blindly transferring decisions from one country to another is therefore not advisable, but it's still better to experiment on a limited set of users instead of your total population and risk loosing either substantial chunk of your users due to overpricing or revenue due to underpricing.