Re-Engagement Campaigns: Winning Back Inactive Subscribers Without Wrecking Deliverability

A re-engagement campaign is the riskiest send in email marketing, and it is usually run by people who think of it as a low-stakes cleanup exercise.

Consider who is in the segment. By definition these are people who have not opened anything in twelve or eighteen months. Some have changed jobs and the address is dead. Some forgot they subscribed and will report spam. Some are mailboxes that were abandoned and have since been recycled, which is where traps come from. You have deliberately assembled the highest-bounce, highest-complaint, highest-trap-density segment in your database and you are about to mail all of it at once.

Done in the wrong order, a win-back campaign damages the deliverability of the engaged list you were trying to protect. Here is the order that limits that.

01Step 1: verify before you send, not after

This is the step that gets skipped, and skipping it is what turns a cleanup into an incident.

The instinct is to send the campaign and let the bounces tell you which addresses are dead. That works, in the sense that you find out. It also means you deliberately generated a large hard-bounce event against the segment most likely to produce one, and hard bounces are one of the strongest negative reputation signals mailbox providers use.

Verify first. Remove everything that comes back undeliverable before a single message goes out. On an eighteen-month-inactive segment expect a substantially higher dead rate than your list average, because inactivity and invalidity correlate: people stop opening partly because they stopped existing at that address.

That single step converts your bounce rate on the campaign from alarming to ordinary, and it costs one verification pass.

02Step 2: decide about risky and unknown deliberately

After verification you will have four buckets. Deliverable and undeliverable are easy. The middle two need a decision, and a re-engagement campaign is the context where that decision most deserves thought.

An address marked risky or unknown is one where the evidence did not resolve: a catch-all domain that accepts everything, a server that would not complete the conversation, a role-based address on a shared inbox. On a normal campaign to engaged subscribers, sending to these is usually fine. On a re-engagement campaign it is not, because you are already accepting elevated risk on every other axis.

My recommendation is to exclude risky and unknown from the win-back entirely. You are trying to limit exposure on a send that is inherently exposed. Adding a bucket of addresses you cannot confirm to a segment you cannot vouch for compounds two uncertainties for no benefit.

03Step 3: send in tranches, smallest first

Do not mail 40,000 inactive addresses on Tuesday morning.

Split the segment into tranches and send over several days, starting small. Something like 2,000, then 5,000, then 10,000, then the rest. After each tranche, stop and read three numbers before continuing:

• Spam complaint rate. The one that matters. Under the bulk sender requirements Gmail's hard ceiling is 0.30% and the practical target is under 0.10%. On a 2,000-message tranche, 0.30% is six complaints. That is a small enough number that you should be counting them individually.
• Bounce rate. Should be low if step 1 was done. If it is not, your verification pass was stale or the segment is older than you thought.
• Unsubscribe rate. Counterintuitively, a high unsubscribe rate here is good news. It means people are using the correct exit instead of the spam button.

If complaints exceed 0.10% on any tranche, stop. Do not proceed to the next one on the assumption it will average out. It will not, because the tranches are drawn from the same population.

The tranche approach is the whole safety mechanism. It converts a single irreversible event into a series of small ones where you can stop after the first bad signal.

04Step 4: write it as a decision, not a plea

Most win-back copy underperforms because it asks for a feeling rather than an action.

"We miss you" gives the recipient nothing to do. "Is this still useful to you? Yes, keep sending / No, unsubscribe" gives them a binary with two clear buttons, and either answer is a good outcome for you. A confirmed stay is an engaged subscriber. An unsubscribe is a complaint you did not receive.

Three things that reliably help:

Make the unsubscribe more prominent than you are comfortable with. Every person who unsubscribes instead of reporting spam is a direct protection of your complaint rate. You want to make the correct exit easier than the destructive one.

Say when you will stop. "If we do not hear from you we will remove you from this list on the 30th" is honest, it creates a real reason to act now, and it commits you publicly to the sunset in step 5.

Keep it to one message, or at most two. A five-email win-back sequence to people who have ignored you for eighteen months is five opportunities to be reported. The marginal return on emails three through five is close to zero and the marginal risk is not.

05Step 5: actually sunset the non-responders

This is where most re-engagement programmes fail, and the failure is organisational rather than technical.

You run the campaign. Some people re-engage. The rest ignore it. And then nobody deletes them, because deleting 30,000 contacts feels like destroying an asset, and there is always an argument for waiting one more quarter.

They are not an asset. They are a liability that has now demonstrated, twice, that they do not want your mail. Keeping them means paying your ESP to store them and carrying their complaint risk into every future send.

Decide the sunset rule in advance, before you see the results, and write it down. "Anyone who does not open or click within 30 days of the final win-back message is removed." Deciding afterwards means negotiating with yourself while looking at a number you do not want to delete.

06What verification can and cannot tell you here

Worth being explicit, because re-engagement is where the two get conflated most often.

Verification tells you whether a mailbox exists and accepts mail. On an inactive segment that is genuinely valuable, because a large share of the segment is inactive precisely because the address died.

Verification tells you nothing about whether the person wants your mail. Someone who has ignored you for two years at a perfectly functional Gmail address comes back deliverable, which is correct, and they are still the most likely person on your list to report you. That signal exists only in your ESP's engagement history.

So the two datasets do different jobs and you need both. Verification sets the bounce rate. Engagement history sets the complaint rate. A re-engagement campaign is the one send where you are gambling on the second, which is exactly why you should not also be gambling on the first.

07The sequence, condensed

1. Verify the inactive segment and remove undeliverable addresses
2. Exclude risky and unknown from this particular campaign
3. Write the sunset rule down before you send anything
4. Send in ascending tranches, checking complaint rate after each
5. Stop immediately if complaints exceed 0.10%
6. One message, two at most, with a prominent unsubscribe and a stated deadline
7. Execute the sunset on the date you committed to

The whole thing is a controlled version of something people usually do uncontrolled. The tranches are the control, verification is what makes the first tranche survivable, and the sunset is the part that makes it worth having done at all.

Verify the segment first. 500 free credits a month is enough to sample it and see what you are dealing with.