T + 2 T + 3 Average impact (A) A A A 1.5A Larger scope Experiment s Cadence (X) 3 1 experiment/month 4 1 experiment/3weeks 6 1 experiment/2weeks 6 Require more design/data Success Rate (S) 33% lucky shot! 25% More experiments might lead to lower success in short term 30%
More experiments lead to higher quality hypotheses 30% Impact (T.A.X.S) 1A uncertain 1A 1.8A 2.7A Impact = T.A.X.S