How to use Win-Back Campaigns
A practical framework for re-engaging churned subscribers — how to segment by reason and recency, design offers that don't erode margin, and measure incremental wins instead of would-have-returned-anyway revenue.
Win-Back Campaigns
Targeted email and SMS sequences designed to re-engage churned subscribers and recover incremental revenue.
A win-back campaign is a structured re-engagement program aimed at subscribers who have cancelled, lapsed, or stopped opening — built around three decisions: who to target, what to offer, and how to sequence the touches.
The discipline that separates a real program from a discount blast is measurement. Without a holdout group, a portion of the people who 'win back' would have returned on their own — so the headline recovery rate flatters the channel and quietly erodes margin. A serious framework segments by churn reason and recency, varies the offer accordingly, and isolates incremental lift against a control.
Win-back is the highest-leverage retention lever most brands run badly. The list is warm — these people already bought once — and the marginal cost of an email is near zero, so even a 3% recovery rate looks heroic on a dashboard.
The problem: most of that 3% would have come back anyway, and the 40% discount you sent them just trained your active base to wait for cancellation emails. This guide walks through the four decisions that make the difference, and sits inside the broader Churn Reduction Playbook as the recovery layer underneath prevention.
1. Segment by reason and recency, not just 'lapsed'
The single biggest upgrade you can make is splitting 'churned' into segments that respond to different messages. A subscriber who cancelled after a damaged shipment doesn't need 20% off — they need to know you fixed the warehouse. Someone who lapsed silently after six months needs a reason to remember you exist.
Start with two axes: reason and recency. Reason comes from your cancellation flow (price, fit, product issue, lifestyle change, no reason given) and from support tickets in the 30 days before cancellation. Recency is how long since their last order — 0-60 days, 60-180, 180-365, 365+.
On a typical apparel or beauty subscription, four segments cover 80% of the volume: recent price-sensitive cancellers, recent product-issue cancellers, mid-recency silent lapsers, and deep-lapsed (12+ months). Each gets a different opening message, a different offer ceiling, and a different sequence length.
Don't win back everyone
Subscribers who churned with 3+ support tickets, returns over 40% of orders, or chargebacks should be excluded. Re-acquiring unprofitable customers is worse than losing them — and they often re-churn within one cycle, dragging your cohort retention curves with them.
2. Design offers that match the churn reason
Default to a non-discount offer first. For a beauty brand, that's a free travel-size add-on with the next order. For apparel, free exchanges or a sizing concierge. For supplements, a one-time skip or pause instead of cancel. These cost less per redemption than a percentage discount and don't reset price expectations on your active base.
When you do discount, ladder it: 15% in email one, 25% in email three, 40% only in the final touch and only for segments where margin allows. Most brands skip straight to 30-40% in the opening email — which is the exact behaviour that conditions cancellation.
Reactivation rate by offer type (90-day window)
Reactivation rate rises with discount depth — but so does the share of would-have-returned-anyway buyers, and so does cannibalisation of your active list. The 25% tier usually has the best incremental-margin profile once you measure against a holdout (covered in section 4).
3. Sequence cadence — three touches over 30 days, then stop
A useful default sequence: day 3 (we noticed, here's what's new), day 10 (specific product reason to return + small offer), day 25 (final touch with the deepest offer the segment allows). For deep-lapsed segments (365+ days), stretch to 5 touches over 60 days because the brand-memory cost is higher.
Stop hard after the sequence ends. Brands that keep poking lapsed subscribers monthly drive their deliverability into the ground — Gmail and Apple Mail both weight low-engagement sends heavily, and your active campaigns to current subscribers suffer for it. Suppress non-openers from broadcast for 90 days after the win-back ends.
Win-back benchmarks by vertical (Shopify subscription brands, €1-15M revenue band)
| Vertical | Open rate (touch 1) | Reactivation rate (90d) | Incremental lift vs holdout | Avg recovered LTV |
|---|---|---|---|---|
| Beauty & skincare | 38-44% | 5-8% | +2.1-3.4pp | €180-260 |
| Supplements & wellness | 32-40% | 4-7% | +1.8-2.9pp | €220-340 |
| Apparel basics | 28-34% | 3-5% | +1.2-2.0pp | €140-210 |
| Coffee & consumables | 35-42% | 6-9% | +2.5-3.8pp | €160-240 |
| Pet food & treats | 40-46% | 7-10% | +3.0-4.2pp | €280-420 |
The gap between headline reactivation rate and incremental lift is the whole game. A beauty brand showing 7% reactivation but only +2.3 percentage points of incremental lift means roughly two-thirds of the 'wins' would have come back without the campaign — and you paid in discount margin for revenue you already had.
4. Measure incremental lift with a holdout
Hold out 10-15% of each eligible segment from the campaign for 90 days. Compare reactivation rate, AOV, and 6-month retained LTV between the treatment and the holdout. The difference is what your program actually generated. The raw reactivation number is a vanity metric.
Refresh the holdout quarterly so seasonal effects don't bias it, and report incremental contribution margin — not revenue — to the leadership team. A 6% reactivation rate at 40% off can have lower contribution than 4% at 15% off once you net out the cannibalisation. Pair this with the broader Churn Reduction Playbook so prevention and recovery share one measurement frame.
What a healthy win-back program looks like
Four reason-based segments, three-touch sequence, ladder of offers starting non-monetary, 10-15% holdout per segment, quarterly measurement against incremental contribution margin. Reactivation rate is reported but never optimised against directly.
Win-back campaign FAQ
Three to seven days after the cancellation processes. Earlier feels desperate and often catches people mid-buyer's-remorse; later than 14 days and brand attention has decayed. For silent lapsers (no formal cancel), trigger at 1.5x their typical purchase cycle.
3-8% over a 90-day window is typical for DTC subscription brands, with beauty and pet food at the top of the range and apparel at the bottom. But the metric that matters is incremental lift versus a holdout, which is usually 1.5-3.5 percentage points lower than the headline rate.
Yes, for subscribers who opted into SMS while active — usually as touch two or three, not the opener. SMS reactivation rates run 1.5-2x email for the same offer, but consent rules mean your eligible audience is smaller. Don't add SMS for the win-back specifically; that's a consent violation in most jurisdictions.
If your win-back discount exceeds your average promotional discount to active subscribers, you're training cancellation behaviour. A practical ceiling is 25-30% for most categories; go deeper only for deep-lapsed segments (12+ months) where the alternative is full re-acquisition cost.
Touch one: a 'we changed X' message referencing a real product or service update since they left, no offer. Touch two: a personalised product reason ('your shade is back in stock' / 'new flavor matches what you ordered before') with a modest offer. Touch three: deepest offer with a deadline. Subject lines that name the gap ('It's been 4 months') outperform vague ones.
Re-engagement targets disengaged but still-active subscribers (low opens, no recent buy) — the goal is to keep them on the list. Win-back targets people who have already cancelled or fully lapsed — the goal is to recover revenue. They share creative DNA but have different audiences, offers, and success metrics.
Yes — high-return-rate customers (>40% returns), serial chargeback filers, anyone with 3+ unresolved support tickets, and customers who explicitly asked not to be contacted. Re-acquiring unprofitable customers damages cohort economics more than the recovered revenue helps.
Prevention (cancellation flow, pause options, proactive saves) sits upstream and is always higher ROI per touch. Win-back is the recovery layer for the churn that prevention couldn't catch. The two share segmentation and measurement infrastructure — if your churn reasons are coded well, both programs benefit.
Match the measurement window: a 90-day holdout for 90-day reactivation, but ideally extend to 6 months so you can also measure retained LTV and re-churn rate. Refresh the holdout each quarter to avoid the same subscribers being permanently excluded.
Yes, mostly for product recommendation in touch two (which specific SKU to surface based on prior order history) and for subject-line variant testing. Don't let AI write the full body without review — the tone needs to acknowledge the cancellation honestly, which generic models tend to soften into uselessness.
Test ideas before you ship them
Run unlimited A/B tests, attach hypotheses to outcomes, and build a searchable archive of what works — and what doesn't.