Win-back emails and sunset policy: when to try again and when to stop

A win-back campaign is a short series of messages to customers who have stopped buying. It often ends in a decision: re-engage them or suppress them. Everything depends on the first step, which is deciding who counts as lapsed.

Without subscriptions, nobody cancels. Customers just stop. A commerce platform's blog on churn in ecommerce counts a customer as churned when they have not reordered within twice the average time between repeat purchases¹. That is one platform's rule of thumb, not a standard. An email vendor's win-back guide says there is no universal definition of lapsed: it depends on your category's natural purchase cycle². That is the useful part. The rest of this page assumes you take it seriously.

Example (illustrative): A coffee brand whose customers reorder every month has a customer who is overdue after two or three missed cycles. A furniture brand whose customers buy once every few years has nobody who is "lapsed" after 90 days. Same label, different clocks.

How to find your clock:

• Look at the gap between first and second orders, and between later orders, for your own customers.
• Take the point where most repeat buyers have already come back. Anyone past it is overdue.
• Split by product category if your range mixes consumables and one-off items.
• Use RFM analysis (recency, frequency, monetary value) if you want to segment lapsed customers by past value, so your best former buyers get the most attention.

Our view: a lapsed-customer definition you cannot explain from your own order gaps is a guess. Write the rule down, with the date you set it.

Keep it short. A win-back is a few messages, not a drip that runs for months. The same email vendor's guide lays out a short sequence and moves non-responders to a suppression list afterwards². We agree, for a reason that has nothing to do with copy: every extra message to an unresponsive address is another chance for a complaint or a delete without opening.

A workable shape:

1. Reminder. A plain message that says what they bought, what is new, and that you noticed the gap. No guilt.
2. Reason to return. Something relevant to their past purchases: a restock, a new variant, a replenishment prompt. If you use an offer, make it the last step, not the first. A discount in every message teaches customers to wait.
3. Last call. Say clearly that you will stop emailing unless they click. Then do it.

Rules for each step:

• Send to people who have consented to marketing email. Consent rules differ by country, and your counsel decides what applies to you.
• Exclude anyone who has bought since the sequence started.
• Segment by past value where you can. Your top customers may deserve a longer, more personal attempt than a one-time buyer.
• Put an easy unsubscribe in every message.

These flows sit alongside your other email automation flows, such as welcome and post-purchase. Check that a customer in a win-back is not hit by a promotional blast at the same time.

An email sunset policy is a rule for suppressing subscribers who have not engaged for a set period, to protect deliverability and sender reputation. Email deliverability means whether your email reaches the inbox rather than spam, or gets rejected.

Why would a dead address hurt a live one? A trade-publication article argues that an unengaged list hurts deliverability and muddies campaign data³. That is an author's view, not a measured rule, so treat it as a reason to look at your own data. The firmer rules are on the provider side. Gmail's sender guidelines ask all senders to keep the spam rate reported in Postmaster Tools below 0.3%⁴. Senders of more than 5,000 messages a day to Gmail accounts should aim below 0.10%, and their marketing and subscribed messages must support one-click unsubscribe⁴. People who ignore you and never unsubscribe are the ones most likely to hit "report spam" when your next message surprises them.

When to sunset inactive subscribers

There is no verified universal number of days. Use a rule built from two inputs:

• Your lapsed definition. A subscriber who has not bought and has not engaged since well past your buying cycle is a candidate.
• A completed win-back. Run the short sequence first. Anyone who does not respond is suppressed, not mailed forever.

Suppression is not deletion. Keep the record, stop sending marketing email, and let the address come back if the person opts in again or buys. One caveat: the same trade article warns that a policy based on opens alone keeps inactive contacts on the list after Apple's privacy changes, and suggests blending clicks, site visits and purchases³. See Apple Mail Privacy Protection for why opens became unreliable.

Before you change anything, check that your basics meet the provider rules. Our Gmail and Yahoo bulk sender requirements checklist covers them.

Our view: a sunset rule is a cost-benefit call. Mail to unengaged addresses tends to earn little and puts at risk the reputation your best customers' emails depend on.

A win-back email will show orders. That does not mean the email caused them. Some lapsed customers were about to come back anyway, and last-click reporting gives the email the credit.

Hold out a random slice of the lapsed segment. They get no win-back messages. Compare orders per person between the two groups over the same window. The gap is the incremental effect. This is the same logic as a holdout test, just applied to email.

Then count it in margin, not in opens or revenue. Our convention is to measure the result on contribution margin before marketing: revenue minus variable costs such as cost of goods, shipping, payment fees, returns and discounts, before any marketing cost. A win-back that works only with a deep discount may add revenue and still lose money.

Report the sunset side too: how many addresses you suppressed, and what happened to complaint and inbox signals afterwards.