Testing buyer paths before building tooling is the right call. Most devs do it backwards — build for months, then discover nobody will pay. Your endpoint list reads like you're solving real problems, not imaginary ones. If the validation loop fails, you saved yourself weeks. If it works, you've got a lean machine. Either outcome is faster than the competition.