Two-proportion z-test. Enter total observations (N) and successes (K) for each group.
E.g. Suppose you have two cohorts - In the first cohort, 1000 clicked and 100 purchased.
In other cohort, 2000 clicked and 210 purchased. How confident can you be that second
cohort outperformed first cohort due to a systemic improvement, rather than just the noise?
You enter N1 = 1000, K1 = 100, N2 = 2000, K2 = 210.
We should strive for 80% confidence in concluding our experiments.