“at the moment we're taking roughly 400 pairs at the first renewals which means like a massive amount of users like dozens of thousands of users and there are a lot of people who try to test way too early”

Don't A/B Test Pricing Until You Have ~400 Renewal Pairs — Most Apps Test Way Too Early

Thomas Petit argues most apps lack the cohort size required for statistically valid pricing tests. His team (with six dedicated people) needs ~400 pairs at first renewal — which translates to tens of thousands of users per variant. Before hitting that scale, make radical pricing changes and read directional signals rather than running underpowered A/B tests that will produce false confidence in noise.