# Fictional example output

**No clear primary winner; rollout is not justified by this evidence.** Control is 100/1,000 = 10.00%; treatment is 110/1,000 = 11.00%. Difference is +1.00 percentage point (+10.00% relative).

Under the supplied independent binary fixed-horizon assumptions, the unpooled normal approximate 95% interval for the absolute difference is −1.6867pp to +3.6867pp. It includes zero and material harm as well as benefit. This is inconclusive, not proof of equality or a 95% probability that treatment wins.

Qualification per assigned user is 4.00% control and 3.90% treatment (−0.10pp). No predeclared guardrail margin was supplied, so acceptance remains unverified. The maturity and assignment checks are supported by the fictional design; any different actual design would require reassessment.

Retain the control and decide whether a separately planned follow-up can answer the question; do not repeatedly extend this test until significance appears. No experiment applied or traffic shifted.