Calculator / Sell & retain
Read the result before calling a winner.
Enter two groups from the same measurement window. Compare the observed rates after refunds. The tool does not decide whether either offer caused a change or whether the difference is reliable.
02 / Follow the result
Your modeled scenario
- A: readers with kept purchases
- —
- B: readers with kept purchases
- —
- A: observed kept-purchase rate
- —
- B: observed kept-purchase rate
- —
- B minus A (percentage points)
- —
For the assumptions entered. This is a scenario, not a forecast.
One question, two scenarios
What changed?
How to read this result
Kept readers = paid readers − fully refunded readers. Kept-purchase rate = kept readers ÷ eligible exposed readers × 100. B minus A is an absolute difference in percentage points, not a relative uplift. Zero exposure leaves that rate and the difference undefined.
Example inputs are invented. This is a descriptive reconciliation, not a significance test, sample-size recommendation, forecast or automatic winner selector. One reader can change a small cohort result materially.
Groups can differ in audience, timing, assignment, offer visibility, price or support. A higher rate does not establish causation or higher contribution. Record the assignment method, primary outcome and stopping rule before starting.
Use a common cutoff that allows comparable refund observation. Partial refunds, chargebacks, multiple purchases, taxes, future renewals and delivery costs require a more detailed ledger; they are not included here.
Inputs run locally in this browser tab. They are not sent or saved automatically. Download your work before leaving if you want to keep it.
The research behind the workings