Is there any way to reduce the confidence of an experiments to get results faster? For example, our marketing folks said that if we can get an A/B test with 75% confidence, that's enough for them to try more tests instead of waiting for a 90 or 95% result.