Example 1
- 8000 sessions/day
- 5 conversion %
- 2 MDE pp
- 1.96 Z
- 1.15 power
1.0 d | n~7,299
With 8000 sessions/day, 5 conversion %, 2 MDE pp, 1.96 Z, 1.15 power the result is 1.0 d | n~7,299. This is not an evanmiller.org sample-size tool.
The calculator uses the numbers you enter. With 8000 sessions/day, 5 conversion %, 2 MDE pp, 1.96 Z, 1.15 power the result is 1.0 d | n~7,299. This is not an evanmiller.org sample-size tool.
This is not an evanmiller.org sample-size tool. Incrementality is on Incrementality.
Enter data and click Calculate.
A/B sample duration in this calculator is days to n from conversion, MDE and Z. With 8000 sessions/day, 5 conversion %, 2 MDE pp, 1.96 Z, 1.15 power the result is 1.0 d | n~7,299.
n per variant comes from conversion, MDE and Z. days = n / sessions. A proxy, not a power test. The formula uses MDE and sessions.
1.0 d | n~7,299 is not an evanmiller.org sample-size tool. Another set of numbers gives 1.4 d | n~5,634, and a further set gives 0.4 d | n~5,027. Empty scenario fields do not change the result.
Incrementality answers a different question. Attribution window does too. Here 8000 sessions/day, 5 conversion %, 2 MDE pp, 1.96 Z, 1.15 power leave 1.0 d | n~7,299.
Type 8000 sessions/day, 5 conversion %, 2 MDE pp, 1.96 Z, 1.15 power and click Calculate. Two more ready sets sit under the form: 1.4 d | n~5,634 and 0.4 d | n~5,027.
Other numbers on the same labels move the result. The first example gives 1.0 d | n~7,299.
n per variant comes from conversion, MDE and Z. days = n / sessions. A proxy, not a power test.
With 8000 sessions/day, 5 conversion %, 2 MDE pp, 1.96 Z, 1.15 power the result is 1.0 d | n~7,299. This is not an evanmiller.org sample-size tool.
1.0 d | n~7,299
With 8000 sessions/day, 5 conversion %, 2 MDE pp, 1.96 Z, 1.15 power the result is 1.0 d | n~7,299. This is not an evanmiller.org sample-size tool.
1.4 d | n~5,634
With 4000 sessions/day, 3 conversion %, 1.5 MDE pp, 1.65 Z, 1 power the result is 1.4 d | n~5,634.
0.4 d | n~5,027
With 15000 sessions/day, 8 conversion %, 3 MDE pp, 1.96 Z, 1.3 power the result is 0.4 d | n~5,027.
1.0 d | n~7,299. This is not an evanmiller.org sample-size tool.
1.4 d | n~5,634.
0.4 d | n~5,027.
No. It is the result from the numbers you enter.
No. Empty scenario fields do not change the result.
That calculator answers another question. Here A/B sample duration stays at 1.0 d | n~7,299.
Yes. A comma and a dot mean the same in a number field.
No. A result from the fields. an evanmiller.org sample-size tool sits outside this calculator.
Yes. That is the second example under the form.
Agent cost is a sketch from your rates. Below are EU, OECD and NIST AI frames, not a contract quote.
Page updated in 2026.