Both, in that order. Re-engagement campaigns are short email sequences aimed at subscribers who have gone quiet, asking them plainly whether they still want your mail. The honest way to run one is as a decision you hand to the subscriber. A few will raise a hand and come back onto your active list. Most will say nothing, and that silence is an answer you can act on: stop mailing them, stop paying for them, and stop letting them decide where your next campaign lands. A win-back campaign that only tries to save people is doing half its job.
So the title question has an uncomfortable answer. You are not running this campaign to keep everybody. You are running it to find out who is left.
Who belongs in the campaign
Two tests, and an address has to pass both.
They engaged at least once. Clicked a link, replied, bought something, logged into your app, at any point in their life on your list. That person has a relationship worth reviving. Someone who has never responded since the day their address landed in your database has no relationship to revive, and mailing them is cold outreach wearing a friendly subject line.
The address still works. More on that further down, because it is the step people skip.
Everyone below stays out:
- Never-engaged imports. The trade show scan, the inherited spreadsheet, the list somebody bought in 2023. No amount of clever copy fixes a name that never asked for anything. The repair for that happens at signup, where the double opt-in versus single opt-in argument gets settled.
- Anyone already past the sunset window who has had one win-back attempt. A sunset policy is your written rule for when a silent subscriber stops receiving marketing mail. Running the same person through a second sequence is mostly a second chance for them to reach for the spam button. The windows by sending frequency, and the policy itself, live in our post on the email sunset policy.
- People who unsubscribed or complained. Obvious, and yet it happens every time somebody rebuilds a segment from a raw export.
- Recent signups. Six weeks on the list is not a fair run at your emails.
- Paying customers who are quiet because they are busy using the product. Their engagement signal sits in your app, not your email reports.
While the segment is open, sort it by how recently each person did anything. That order matters when you start sending.
The shape of the sequence
One to three emails across two to four weeks. Longer sequences do not persuade a disinterested person. They just hand a distracted one more chances to report you.
Email one: the plain question. Short, plain text, from a real person’s name. One question, which is whether they still want these emails. Two obvious links, one to stay and one to go. No banner, no hero image, no discount. The ask is the message.
Email two: the alternative. About a week later, to whoever did nothing. This is the email that recovers the most people, because a good share of the silence on any list is about volume rather than interest. Offer a monthly version, a single topic, a digest, a pause until January. Give them something to choose instead of a door to walk through.
Email three: the last notice. One to two weeks after that. Name the date they will be removed, and then remove them on it. This email usually pulls the strongest response of the three, for the unglamorous reason that it is the first one with a consequence attached.
Stop the sequence the instant somebody responds. Nothing sours a recovered subscriber faster than a “we are removing you” notice landing two days after they told you to stay.
Check your one-click unsubscribe header before any of this goes out. Of all the mail you will send this year, this is the batch where leaving needs to be easiest.
Six subject lines that work because they are honest
Clever hurts here. The recipient half-remembers your brand, so a playful line reads as a stranger being cute, and the quickest way to deal with a stranger being cute is the spam button. Plain lines get the question answered:
- Still want these emails?
- Should we stop sending?
- Do you still want to hear from us?
- Want fewer emails from us instead?
- Last email unless you tell us otherwise
- We are removing you on the 14th
Leave out the fake “Re:” and “Fwd:” prefixes, the sad face next to “we miss you”, and any warning about an account closing when no account is closing. Those tricks do get responses. They also get complaints, and this is the one segment where you cannot absorb them.
What counts as a win
Decide this before you build anything and write it into the segment rule. The temptation to move the line arrives about two days into the campaign. A win is:
- a click on a link in the email
- a reply
- a purchase or a renewal
- a preference update, including a downgrade to monthly
- a login or any action inside your product
An open does not make the list. Opens are recorded when an invisible image loads, and since Apple introduced Mail Privacy Protection, Apple Mail preloads that image whether or not a human ever looked at the message. It is widely documented that this inflates open rates on any list with a decent share of Apple Mail readers. Count opens here and you will keep a pile of addresses that did nothing, then congratulate yourself for rescuing them.
The riskiest send you will make all year
By design, this campaign goes to the people least interested in hearing from you, all at once. Complaints run higher than on any ordinary broadcast. Plan around it.
- Send in small batches. A few thousand at a time over several days, so trouble shows up while you can still stop.
- Send the most recently quiet slice first. Those people respond best, which gives the campaign a healthy opening before the older, harder batches go out.
- Watch the rate between batches. Google’s sender guidelines say to “Keep spam rates reported in Postmaster Tools below 0.10% and avoid ever reaching a spam rate of 0.30% or higher.” Postmaster Tools is Google’s free dashboard for domain owners, and for Gmail recipients it is the only place that number appears. Our guide to the spam complaint rate covers how to read it.
- Keep it away from your big campaign. If complaints climb, you want quiet days afterward rather than the largest send of your quarter landing in the middle of the damage.
Verify the segment before the first email goes out
Skip this and the campaign meant to clean your list damages it instead.
Your inactive segment is the oldest, least-touched part of the database, which makes it the part where the most addresses have quietly died since you last looked. People change jobs, abandon accounts, let domains lapse. Our post on email list decay covers how fast that happens. Mail the segment without checking and email one produces a bounce spike arriving at the same moment as the complaint bump: two bad signals, same day, same sender.
Verification checks each address against the receiving mail server without sending anything. Reoon Email Verifier returns a verdict per address: safe, invalid, disposable (a throwaway address from a temporary-mail service), catch-all (a domain that accepts mail for anything, so the individual mailbox cannot be confirmed) or unknown. Suppress the invalid results before you write a word of copy. They were never going to reply.
What to do with everyone who stayed quiet
Suppress them. Do not delete them. A suppression list is the do-not-send list your platform checks automatically before every campaign, so the record survives even though the sending stops.
Store a reason with the date. “No response to win-back, 2026-11-25” answers the question somebody asks next spring about where eight thousand contacts went. Deleting throws that history away, and the address tends to walk back in with the next CSV import anyway.
One annual check-in is the most I would do afterward, verified again first, and only where the original permission still stands. Do not quietly fold suppressed contacts back into the regular broadcasts because a quarter looked thin.
An honest opinion about what re-engagement campaigns achieve
Most of them win back a small share of the segment. That is normal and it is fine, because the recovered subscribers were never really the product of the campaign. The product is a list you can describe accurately: every name still on it has done something deliberate, and every name that has not is off the send. That is worth more over a year than the handful of orders the sequence brings in.
If a campaign brings back a large share, read it as information rather than a triumph. It usually means your definition of inactive was too aggressive, or that you had been sending more often than people wanted and they tuned you out without ever leaving. Both of those get fixed upstream of the campaign.
The sequence on one page
| Timing | Message | The ask | What a response means | |
|---|---|---|---|---|
| 1 | Day 0 | Plain text from a person: are these emails still useful to you? | Keep me, or take me off | Either click is a decision. Keep the keeps, suppress the leavers |
| 2 | About a week later | You can have less of us, or something different | Pick a frequency or a topic | Volume was the problem. Move them to the lighter list |
| 3 | About 2-3 weeks after email 1 | Last notice, with the removal date in it | Stay, or do nothing | Silence is the answer. Suppress on the date you named |
Where to start
Start with the segment. Pull everyone who has done nothing inside your chosen window, then find out whether those mailboxes still exist before you write a single subject line. A free account comes with verification credits that renew every day and needs no card, which is enough to run your quietest few thousand addresses this afternoon and see what you are actually working with.
FAQ
How many emails should a re-engagement sequence have?
Two or three suits most senders, spread over two to four weeks. One is plenty if you send rarely and the segment is small. Beyond three you rarely convert anyone the earlier emails missed, and every extra send gives a disinterested recipient another chance to complain. Stop the sequence the moment somebody clicks, replies or updates their preferences.
Should I offer a discount in a win-back email?
Only if discounts fit your business, and not in the first email. An offer tests whether somebody wants a deal, which is a different question from whether they want your emails. Ask the plain question first and keep the offer for email two if it suits you. Subscribers recovered by a code often go quiet again once it expires.
Can I count an open as re-engagement?
Better not to. Apple Mail Privacy Protection preloads the tracking image in many messages, so an open can be logged without a person seeing anything, and this is widely documented as inflating open rates. Keep addresses on the strength of opens and you keep the same dead weight the campaign was built to clear. Require a click, a reply, a purchase or a preference update.
How often should I run a re-engagement campaign?
Once or twice a year works for most senders, or run it continuously as an automation that triggers per subscriber when they cross your inactivity window. The continuous version spreads the complaint risk over months, which is easier on your sending reputation. What you should avoid is putting the same people through the sequence again after they ignored it once.
