Buy Again Rollout

Amazon: Decide whether a homepage module that suggests repeat purchases should be rolled out.

Assignment

A "buy again" module was shown on the homepage to a random half of a customer group for four weeks, and the product team now wants a recommendation on whether to roll it out to everyone. They have shared two exports: the assignment log for the test, and weekly activity for the customers in it.

Deliverables

• A single headline measure of whether the module worked, with the figure for each group.
• Two or three supporting measures that would tell the team if the module was causing harm elsewhere, and what each one protects against.
• Your recommendation on whether to roll the module out, and a short closing report for the product team covering the reasoning and anything they should check before acting on it.

Data Description

experiment_assignments.csv has one row per customer entered into the test:

• customer_id
• variant - which group the customer was placed in
• assigned_on
• is_prime - whether the customer held a Prime membership when the test began

customer_activity.csv has one row per customer per week of the test:

• customer_id
• week - week of the test, 1 to 4
• orders
• order_value_usd
• returns - orders the customer sent back
• support_contacts

Practicalities

You can solve this in Python, SQL, or a mix of both — StrataScratch notebooks support both languages in the same notebook.

Assume this data comes directly from the experiment platform and hasn't been cleaned, so review it carefully before relying on the numbers. When something looks off, work out what it is and decide how to handle it; when a definition or a final decision is genuinely ambiguous, choose a reasonable option and explain why. There's often more than one valid way to compute a result, so explain the approach you take and the reasoning behind your choices.

What's assessed: your results, and how clearly you explain what you see in the data and justify the choices you make.