Rebuilt our winback flow with this prompt — three emails, escalating incentive logic, honest exit conditions. The sequence reactivated 11% of lapsed subscribers in June.

## Winback flow: meal-kit subscription (paused/cancelled 60+ days) **1. Segmentation gate.** - Segment A — power users (20+ deliveries): highest reactivation odds, lead with what's new. - Segment B — barely activated (≤3 deliveries): they never formed the habit; lead with the onboarding they effectively missed. - **Suppress entirely:** involuntary churn (card failures — route to dunning, not winback), anyone with an open support ticket, cancellations citing dietary/medical reasons, and anyone winback-mailed in the past 90 days. **2. Email 1 — value, no discount (both segments, day 0).** - Subject A: "The three recipes people came back for" - Subject B: "We fixed the thing you told us about" - Preview: "No discount in this email. Just what's changed." - Body (Segment A, 118 words): opens with the two most-requested changes shipped since their last box — 20-minute recipe line and the swap-any-protein option — because their cancel-survey cohort's top reason was weeknight time pressure. Ends with a single CTA: "See this week's menu", no urgency device. **3. Email 2 — friction removal (non-engagers, day 6).** Your data says the top lapse reason is "too much food / waste guilt." The email leads with the honest sentence: *"Most people who pause tell us the portions outpaced their week — fair."* Then introduces the 2-recipe minimum plan and pause-anytime-in-app. One CTA: "Build a smaller week." **4. Email 3 — incentive (non-engagers, day 12).** Offer: 40% off one box (within your CHF 30 cap at AOV 72) — one box, not a multi-month discount, because multi-month deals are what teach lapse-and-wait behavior. Expiry: 7 days, real, enforced. Copy frames it as a decision-closer: *"If a smaller, faster box still doesn't fit your weeks, that's a real answer — this is the last email in this series either way."* **Exit condition:** no click → 180-day winback suppression; two full sequences ignored → move to quarterly digest list only. **5. Measurement.** E1: menu-page CTR. E2: plan-configurator starts. E3: redemption rate. Flow success threshold: ≥6% reactivation at 30 days post-sequence. Discount-training indicator: rising share of reactivations that ONLY convert on email 3 across cohorts — if that share passes 70%, the incentive is doing all the work and E1/E2 need rebuilding. **Positioning conflict flag:** your voice sample emphasizes abundance ("generous boxes, feed the whole table") while your top churn reason is waste guilt. Email copy can't resolve that — the portion-size product decision is the actual winback lever here.