Understand your numbers & results

Read the break-even calculation

Average order value is net revenue divided by total orders. Contribution per order is that average multiplied by your contribution margin. Monthly program cost includes the service fee, extra platform costs, and setup fee divided by its allocation months.

The tool rounds the orders needed to cover that cost up to a whole order. It cannot calculate a meaningful break-even point with no positive contribution per order. This is a planning scenario, not a forecast; the extra orders must actually be caused by the program.

Record a completed comparison

1. Open Pilot results and name the test. Enter a completed start and end date.
2. Enter everyone assigned to treatment and control, plus the unique buyers in each group. A buyer is a person, not their number of orders; buyers cannot exceed participants.
3. Describe the design, source, exclusions, and limitations in at least 20 characters. Use the same observation window for both groups.
4. Check the randomization statement only if it is true. Choose Save comparison and review the result.

What the result means

The difference is in percentage points. If treatment converts at 12% and control at 10%, the displayed difference is +2 percentage points. It is not a 2% increase in revenue.

A non-randomized comparison is labeled Descriptive only. A randomized comparison can show a positive signal, negative signal, or an inconclusive result according to its approximate 95% interval. Randomization is your recorded assertion, not something the app independently verifies.

The interval does not correct for repeated checking, multiple tests, overlapping groups, or a poor design. These comparisons do not measure incremental profit or guarantee future results. Keep attributed revenue separate from evidence of a causal improvement.