- Trigger win-back emails from your own repurchase data, not a generic 90 days: start shortly after the point where most of your repeat buyers have already come back.
- Lead with reasons to return (new products, replenishment, what they bought before) and save the discount for the last message, sized from contribution margin.
- Set a hard stop: customers who ignore the whole flow move to a lower-frequency track and, if they stay silent, out of your active email list.
What is a win-back email, and how is it different from a sunset flow?
A win-back campaign targets past customers who have stopped buying. The goal is a new order. It is triggered by purchase inactivity: a customer who usually reorders has gone quiet for longer than expected.
A sunset flow targets subscribers who have stopped engaging with your emails. The goal is list health. It is triggered by email inactivity, and it ends by asking people to confirm they still want to hear from you, then suppressing those who don’t.
A customer can click every email and still not buy: a win-back candidate, not a sunset candidate. Mixing the two sends discounts to people who were never customers, or suppresses customers who are still reading.
| Win-back flow | Sunset flow | |
|---|---|---|
| Audience | Past buyers | Any subscriber, buyer or not |
| Trigger | No order within the expected repurchase window | No clicks or site activity within an engagement window |
| Goal | Next order | Clean list, protected deliverability |
| Typical offer | Reasons to return, then a margin-safe incentive | A “stay subscribed?” choice |
| Exit | Places an order | Clicks, or is suppressed |
When should a win-back email be sent?
The common advice is “90 days after the last order,” which only works if your customers repurchase on a 90-day rhythm.
Find your own repurchase window
- Export orders from Shopify (or your store) with customer ID and order date for the last 24 months.
- For every customer with two or more orders, calculate the days between the first and second order.
- Sort those gaps and find the point where about 75 to 80 percent of second orders have already happened. That is your “most people who will come back have come back” line.
- Repeat for later orders (second to third) if you have the volume, since loyal customers often reorder faster.
Your win-back trigger sits just after that line. Before it, you discount people who were about to buy anyway. Long after it, the customer has likely moved on.
An illustrative example: a skincare brand finds that 80 percent of second orders arrive within 75 days of the first, so it starts the flow at day 80. A furniture brand might land on 300 days or more, or decide win-back isn’t the right tool at all.
This is the same cohort logic behind your first-to-second order rate; our guide to the first-to-second purchase rate walks through the setup.
Use predicted dates if your platform offers them
Some ESPs estimate when each customer is likely to order next. Klaviyo’s predictive analytics, for example, includes an expected date of next order once an account has enough order history to qualify. Triggering a set number of days after that date personalizes the timing. Keep a fixed-window fallback for customers without a prediction.
If one product runs out in 30 days and another lasts a year, a single trigger fails both groups. Split the flow by the category of the last order and give each its own window.
What should a win-back flow look like?
A practical default is three to four emails over three to four weeks, with SMS only for opted-in customers with a strong order history. It’s a starting structure, not a benchmark.
| Step | Timing (illustrative) | Message | Offer |
|---|---|---|---|
| 1 | Trigger day | “We’ve added things since your last visit”: new arrivals or best sellers in the category they bought | None |
| 2 | +5 to 7 days | Personal reminder: their last product, a refill or upgrade, reviews from similar customers | None, or free shipping if shipping cost is a known objection |
| 3 | +7 days | Direct ask: “Come back for X” with a clear deadline | Margin-safe incentive (see below) |
| 4 (optional) | +7 days | Last chance on the incentive, plus a preference link (less often, different topics) | Same incentive, expiring |
Two rules make this work:
- Exit on purchase, everywhere. Stop the flow the moment the customer orders on any channel you sync. “We miss you” the day after a purchase burns trust.
- Suppress people in other flows. A customer in a post-purchase or browse abandonment flow isn’t lapsed. Use flow filters to keep them out.
What to write when there’s no discount
The first two emails should give a reason to return that isn’t price:
- What’s changed: new products, sizes or restocks of something they bought.
- What they liked: their last product, with a one-click reorder link.
- What similar customers bought next: the most common follow-up purchase.
- Proof: reviews from customers who bought the same item.
Segmentation helps here. In our Value × Intent (RFMB) model, which adds a Buying Stage layer to classic RFM analysis, a lapsed high-value customer who viewed a product this week (B2) gets a personal, no-discount nudge instead of the generic flow. See the RFM segmentation playbook.
What should a win-back offer be?
Ask a better question first: what can you afford to give to get one more order from this customer? The answer comes from contribution margin, not from the discount your competitors run.
Size the incentive from margin
Worked example with illustrative numbers. Your average lapsed customer’s order is $70 with a 45 percent contribution margin, so a full-price order contributes $31.50.
| Incentive | Cost per redeemed order | Contribution left | Notes |
|---|---|---|---|
| None | $0 | $31.50 | Always test a no-offer control |
| Free shipping (actual cost $8) | $8.00 | $23.50 | Strong when shipping is the stated objection |
| 15% off | $10.50 | $21.00 | Cost lands on every redeemer, including those who would have returned anyway |
| 25% off | $17.50 | $14.00 | Rarely justified unless the customer’s expected future value is high |
| Gift with purchase (product cost $6) | $6.00 | $25.50 | Feels generous, costs less than its perceived value |
The trap is customers who would have come back anyway: every one who uses the code is margin given away. That’s why the offer belongs in step 3 and why a holdout group matters. Our article on contribution margin covers how to build the per-order number.
Match the offer to customer value
- High-value, lapsed: a personal message, early access or a gift. Blanket codes teach your best customers to wait.
- One-time buyers: a modest incentive on a second product, since the second order moves lifetime value most.
- Discount-only buyers: consider no offer. Another code usually just books another low-margin order.
When should you stop sending win-back emails?
Customers who ignore the whole flow shouldn’t keep getting your full campaign calendar. A simple stop rule:
- Finished the flow, no order, no clicks: move to a reduced-frequency segment (for example, one campaign a month with your strongest content).
- No clicks or site visits for a further period: enter a sunset flow with one or two “do you still want these emails?” messages.
- No response to the sunset flow: suppress from marketing email. Keep them as customers in your store data and in ad suppression lists; just stop emailing.
Use clicks, site activity and orders as engagement signals, not opens. Apple Mail Privacy Protection loads images automatically, so opens overstate engagement for many subscribers.
Why bother? Gmail and Yahoo’s bulk sender requirements include one-click unsubscribe and low spam complaint rates (Google’s stated threshold is below 0.3 percent). Mailing people who never engage raises complaints and lowers inbox placement for everyone else. Our guide to email deliverability for Shopify stores covers list hygiene, and a strong post-purchase flow keeps fewer customers from lapsing in the first place.
How do you measure whether win-back is working?
Platform-attributed revenue flatters any win-back flow, because it credits orders that would have happened anyway. Measure it this way instead:
- Holdout group. Randomly exclude 10 to 20 percent of eligible customers from the flow. Compare their reorder rate over the following 60 to 90 days with the group that received it. The difference is the flow’s real effect. This is a basic form of incrementality testing.
- Contribution, not revenue. Subtract discount cost from the orders the flow generated. A flow that lifts orders but gives a code to everyone may still lose money against the holdout.
- Reactivated customer retention. Check whether reactivated customers place a further order. If they only ever buy with a code, the flow is buying orders, not customers.
- Impact on retention rate. Track your customer retention rate by cohort before and after launching the flow.
Win-back launch checklist
- [ ] Repurchase window calculated from your own order data, per category if needed
- [ ] Trigger set just after the point where most repeat orders have happened
- [ ] Flow exits on any order, from any channel, and skips profiles in other flows
- [ ] Incentive sized from contribution margin, placed in step 3 or later, with a no-offer branch for discount-only buyers
- [ ] 10 to 20 percent holdout group in place
- [ ] Non-engagers move to reduced frequency, then sunset, then suppression
- [ ] Reporting built on contribution and holdout lift, not attributed revenue
FAQ
How many days after the last purchase should a win-back email go out?
There’s no universal number. Calculate how many days it takes for most of your repeat customers (around 75 to 80 percent) to place their next order, and trigger the flow just after that point. For fast-replenishment products that might be weeks; for durable goods, many months.
Should a win-back email always include a discount?
No. Start with non-price reasons to return and add an incentive only in a later message, sized from your contribution margin. Test against a no-offer holdout, because some lapsed customers return without one.
What’s the difference between a win-back flow and a re-engagement or sunset flow?
Win-back targets past buyers who stopped ordering and aims for a new purchase. A sunset or re-engagement flow targets subscribers who stopped engaging with emails and aims to clean the list, ending with suppression for those who don’t respond.