Win-back emails: how to bring back customers who went quiet

Someone used your product for a few weeks and stopped. They did not cancel or complain, they just disappeared. They are still on your list, still get your campaigns, and never open one. Every month that passes lowers the odds they come back, and every ignored email teaches Gmail and Outlook that your messages are not wanted.

A win-back email has two jobs: bring back the people who can still come back, and let go of the ones who will not, before that part of the list hurts delivery for everyone. This article covers both, with real examples from ViralRecast and Renderster, Wunka products that send through Couryo.

Why "we miss you" doesn't work#

The classic win-back email says "We miss you! Come back!". The problem: it talks about you, not about them. They stopped for a reason, and that reason is still there.

The ViralRecast case: a new use in every email#

ViralRecast reads and summarizes social posts inside Claude and ChatGPT. For people who stopped using it, win-back emails do not ask them to come back. Each one shows a different use they may not know (translated from Portuguese):

Someone who stopped because "I already read the posts I wanted" gets a new reason to open the product. It is the difference between "come back" and "look what else you can do".

The Renderster case: own the problem#

Sometimes people left because something went wrong. Then the strongest email is the most honest one. Renderster, a Wunka video product, once sent an email after a service problem with the subject "50% off your first month of Renderster: our apology" (translated).

The lesson is not the discount, it is the order: first acknowledge the problem, then say what changed, and only then offer something. And a caveat: this kind of email works because it is rare. An apology every week is just another promotion.

When to send: define "inactive" for your product#

"Inactive" depends on your product's rhythm. A daily-use app with 7 idle days is already at risk; a monthly tool is not.

Use product usage as the main signal, not email opens. Opens mislead: Apple Mail opens emails on its own for privacy, so someone can "open" everything and never use anything.

A three-email win-back sequence#

Email 1: a new use

• Subject: the result of the new use ("Your weekly content plan, built in 5 minutes").
• Body: one situation, the result and a button straight to doing it.

Email 2: what changed

• Subject: "What's new in {{product}} since your last visit"
• Body: two or three real changes, one line each. If nothing relevant changed, skip this email.

Email 3: the final question

• Subject: "Do you want to keep getting our emails?"
• Body: "It's been a while since you opened our emails or used {{product}}. To keep your inbox tidy, we'll stop sending updates. If you'd like to keep getting them, just click below."
• Button: "Keep me subscribed".

The third email matters most for your deliverability, and it is the one almost everybody forgets.

The sunset policy: when to stop#

Sending campaigns month after month to people who never open or click has a hidden cost. Mailbox providers measure engagement: when a large share of recipients ignore everything, more of your email goes to spam, including for people who want it. Addresses abandoned for a long time can also turn into spam traps.

The practical rule:

1. People who clicked nothing and did not use the product in your chosen period enter the win-back sequence.
2. People who did not answer the final question leave your marketing campaigns.
3. Transactional emails (password, billing) keep going as usual.

It looks like losing contacts. In practice you lose people who were already gone, and the rest of your list lands in the inbox.

What not to do#

• Do not buy them back with a discount every time. You teach the list to disappear to get a discount.
• Do not trick with the subject ("Re: your account", "Urgent"). It drives spam complaints, the signal that hurts delivery the most.
• Do not send win-back to people who cancelled and asked to leave. An unsubscribe is final; they come back only by signing up again.

Which metric to watch#

• Return to the product: how many inactive users used it again within 14 days of the first email.
• Clicks per email: which new use sparks the most interest.
• Campaign delivery after the sunset policy: with fewer inactive contacts, complaints and bounces tend to drop.

How to build it with Couryo#

1. Store the last-use date on the contact as an attribute (last_used), updated by your app (see Contacts and events).
2. For win-back, create a campaign for a segment: last_used before a date and last_clicked_at not within the last 90 days. Contact dates (created_at, last_emailed_at, last_opened_at, last_clicked_at) work in segments (Campaigns). Or trigger a sequence from a user.inactive event your app sends.
3. In the sequence, use the conversion event user.active_again: whoever comes back exits on their own.
4. For people who did not answer the final question: remove them from campaigns with a topic (for example updates) or a tag, and target future campaigns only at people in the topic (Topics and preferences).
5. Before sending, the campaign check counts the real audience, warns when many people never got email from the project and reviews the content with AI.

FAQ#

After how long is someone inactive?#

It depends on your product's rhythm. For daily use, 7 to 14 days without use; weekly, 3 to 4 weeks; monthly, 2 to 3 months. Use product usage as the signal, not email opens.

Why stop emailing people who don't open?#

Providers measure whether people care about your emails. A list full of people who ignore everything sends more of your email to spam, including for people who want it. Stopping protects delivery for the rest.

Do discounts work for win-back?#

When they are rare and have a reason, such as an apology after a real problem. A discount in every win-back teaches your list to disappear to get one.

Do people removed from campaigns still get password emails?#

Yes. Transactional email (password, billing, account notices) keeps going. The sunset policy only applies to marketing.