Shamalo

SYS/COCKPIT

CRO tools

Pre-test, post-test reading and sequential peeking — the experimentation lab in one workspace.

Pre-testDuration

Test duration

Turns a sample size into a calendar length. The rate is entered per day, week or month, and the duration is read in the unit you pick. A sample size without an end date is a test that will get peeked.

SYS/IN

Visitors per arm. Total = this value × the number of arms.
Arms
2 for an A/B. 3 or 4 stretch the calendar when n is per arm.
Site-wide traffic. A Wednesday is not a Saturday.
Rate unit
Read the duration in

SYS/OUT

Weeks

5.9

Other readings: 41 days · 1.3 months

Weekday and weekend traffic are not interchangeable. A month here is 365.25 / 12 days, not a named month.

Stopping before this n the moment p crosses α is peeking — the test is no longer fixed-horizon. Open the peeking tool.

Duration against traffic rate
Duration against traffic rate
RateDuration
10023.4
20011.7
4005.9
8003
1,6001.6

Curve: n total ÷ rate, rounded up to a whole day, then converted into the unit you picked. The table is readable with JavaScript off.

How this is calculated

Method

Days = ceil(n_total / visitors_per_day). The rate can be entered per day, week or month: it is converted to a daily rate before the division (a week is 7 days, a month is 365.25 / 12). The display unit only changes the reading: weeks = days / 7, months = days / (365.25 / 12).

n_total = n per arm × number of arms, or the total n entered directly. This is calendar arithmetic, not a statistical model: the tool does not forecast traffic, it divides it.

Sources

The sample size itself comes from a two-proportion z-test — see Sample size (binary). This page only turns that n into weeks.

Limits

Traffic is not flat. A Wednesday is not a Saturday; a bank holiday is not a Tuesday. Round up to full weeks.

Stopping the moment p crosses α is peeking: the z-test assumes a horizon fixed in advance.

SYS/VIEW