Designing Chewy's Autoship Habit Loop: A 90-Day Cancellation-Prevention Rollout
Objective: Play a lifecycle marketer at Chewy across three decision checkpoints, choosing the cue, recovering from a bad launch decision if needed, and setting the reward structure for Autoship, then see how each choice changes 90-day cancellation and trust outcomes.
Chewy's Autoship subscription sees a cancellation spike in month 2, right after the novelty of the welcome discount wears off. You're asked to redesign the reminder cue and reward structure before the next quarterly cohort launches.
At each checkpoint, choose the option that builds a habit users would endorse on reflection, not one that manufactures panic to hit a short-term save-rate number.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, sufficient for documenting a decision log
The process
Simulation
Day 0: Choosing the Cancellation-Prevention Cue
Day 0Data shows most Autoship cancellations happen in the 3 days before the second shipment, when customers forget they're enrolled. Leadership wants a fix live before next week's cohort.
Month-1 Autoship retention: 71%. Support tickets tagged 'forgot I was subscribed': 340/week. Current reminder: none.
Which reminder do you ship before next week's shipment?
Final deliverable
A 90-day Autoship habit-loop rollout plan documenting the cue, recovery step (if needed), and reward chosen at each checkpoint, with the resulting retention numbers.
See a reference example
HelloFresh Weekly-Box Reminder, 90-day rollout log (excerpt) Day 0: Cue chosen - dated reminder 'Your box ships Tuesday, skip or swap by Sunday 11pm'. Verdict: optimal. Day 60: Reward chosen - variable recipe-card bonus (sometimes a chef video, sometimes a free side). Verdict: optimal. Result: 6-box retention projected to rise from 44% to 55% based on the variable-reward stage.
Success criteria
You're done when you can:
- Chooses the honest, dated cue over the fake-urgency option at stage 1
- Correctly identifies the variable reward as stronger than the fixed discount at stage 2b
- Rollout log names the actual retention numbers changed by each choice, not just the choice made