Business Context

StreamWave, a video streaming app, wants to test a new push notification strategy aimed at reactivating dormant users. Users are considered dormant if they have not opened the app in the last 28 days.

Problem Statement

The product team ran a 14-day randomized A/B test. Control users received the current generic push reminder, while treatment users received a personalized push sent at the user's historically most active hour. Determine whether the new strategy significantly improves resurrection rate and whether the observed lift is large enough to justify rollout.

Given Data

Additional business inputs:

Requirements

1. State the null and alternative hypotheses for the experiment.
2. Run a two-proportion z-test using the observed data.
3. Compute a 95% confidence interval for the treatment-control difference.
4. Determine whether the result is statistically significant at $\alpha = 0.05$.
5. Compare the observed lift to the minimum detectable effect.
6. Estimate the required sample size per group for detecting a 0.40 percentage point lift with 80% power.
7. Briefly explain key design choices: unit of randomization, primary metric, and at least two guardrails.

Assumptions

• Randomization is at the user level.
• Each user is counted once.
• No interference between users.
• Push delivery and tracking are reliable and comparable across groups.