Hi, I am trying to understand the sample size calculator. As with the increase in the lift percentage, the no of days to complete the test is less. Example - if Lift is 15% then Days to complete test is 2 if daily visitors are 1242. So, how it is calculating this no of days. Could anyone explain how this field nos. are calculating(no of days)
The number of days are based on number of visitors/day.
Lift is directly related to the number of visitors of the page.
If you want the A/B test to have a high probability of detecting a very small lift, the number of visitors that must be included in the test would be prohibitively large. The reason for this is that, if the difference in conversion rate is small, both conversion rates must be estimated with high accuracy to identify the difference, which requires a large number of visitors.
I read it like this: the higher the lift, the quicker the test finds a winner. Lift is a calculated metric based on the typical conversion rate that you provided in accordance with the new offers performing the same conversion metric as control. If an offer is providing 15% lift, then the experiments wraps up with the winner in short order.
Side note: you must be doing something right because I think a 2% conversion rate is a good baseline for click-thru banner sort of stuff.