Required sample size

Before Tokopedia launches a test, someone has to answer "how long do we run it?" — and an underpowered test is the worst outcome available, because weeks of traffic produce a result nobody can act on.

The sample size per variant for a two-proportion test is:

n = (z_alpha + z_beta)^2 * 2 * p * (1 - p) / mde^2

where p is the baseline rate and mde is the minimum detectable effect in absolute percentage points. power_z holds the two z-values: alpha_two_sided_05 for 5% significance and beta_power_80 for 80% power.

From experiment_plan, return each experiment with users_per_variant — the formula above, rounded up to a whole user. Order by experiment.