Most re-engagement advice tells you to send a win-back series and then clean up whatever bounces. Run in that order, you find your dead addresses by bouncing off them and your spam traps by hitting them — and you learn both things after the damage.
The sequence matters more than the copy here, which is unusual for a marketing topic. A well-written win-back email sent to the wrong list in the wrong order will hurt you more than a mediocre one sent carefully.
So this is the version organised around the mechanics: what the actual thresholds are, why dormant lists are dangerous in a specific technical sense, and a ramp that lets you find out whether anyone still wants you without betting your inbox placement on the answer.
The threshold that makes this risky
Start with the number, because it reframes how big your first send should be.
Google asks bulk senders to keep spam complaints below 0.1% and treats 0.3% as the enforcement line. It's measured daily in Postmaster Tools. Cross 0.3% and you become ineligible for delivery mitigation, and you typically need to sit back under the threshold for several consecutive days before that's restored.
Bulk sender status applies to anyone sending around 5,000 or more messages a day to personal Gmail accounts. It counts across your subdomains, and once you cross it the classification doesn't expire even if your volume drops later.
| Send size | 0.1% (target) | 0.3% (enforcement line) |
|---|---|---|
| 1,000 | 1 complaint | 3 complaints |
| 5,000 | 5 complaints | 15 complaints |
| 25,000 | 25 complaints | 75 complaints |
| 100,000 | 100 complaints | 300 complaints |
Look at the 25,000 row and imagine emailing people who last heard from you eighteen months ago. Seventy-five of them hitting "report spam" is not a remote scenario — it's a normal Tuesday for a cold list. And because the rate is daily, a single large send can breach it in one day rather than degrading gradually.
That's the whole argument for everything below. The rest of this is about not putting yourself in that position.
Clean first, then send
The order correction that most guides get backwards.
Before anything goes out, run the list through validation and remove what's obviously dead — syntax failures, non-existent domains, addresses that no longer resolve. Then remove role accounts, known complainers, and anyone who previously unsubscribed and got re-added by a form import at some point.
The reason the order matters Cleaning first turns your campaign into a test of interest. Sending first turns it into a test of your infrastructure, conducted in public, with your sender reputation as the stake.
Validation costs a small amount per thousand addresses and removes the highest-risk portion of the file before it can do anything. Set against the cost of a blocklisting, it's the cheapest insurance in email marketing.
One caveat worth knowing: validation catches dead addresses, not traps. Nothing catches traps reliably, which is why the next section matters.
Why dormant lists hit spam traps
The mechanism most articles skip, and the one that explains why this is riskier than it looks.
A recycled spam trap is an old, abandoned mailbox that a provider has repurposed to catch senders who don't practise list hygiene. The address was real once. Someone used it, stopped, and after a long dormancy the provider turned it into a detector.
Now consider what a dormant list is made of: addresses that nobody has opened in a long time. That's the same population. Dormant doesn't only mean uninterested — it frequently means the mailbox no longer belongs to anyone, and mailing it is precisely the behaviour the trap exists to catch.
Which is why "we've got 40,000 old subscribers, let's wake them up" is a riskier plan than it sounds. The older the segment, the higher the trap density, and hitting the wrong one can affect delivery well beyond the campaign that caused it. This is the same underlying dynamic behind why legitimate email still lands in spam.
InboxAlly
Trains mailbox providers to inbox your emails by simulating genuine opens, clicks and replies.
Best for: Senders fixing inbox placement and warming up domains or sender reputation
Don't re-engage everyone
The instinct is to work the whole dormant file because it feels like recoverable value. The economics rarely support it.
Segment by recency of last genuine engagement, and treat the tiers differently:
- 3–6 months dormant. Worth a proper attempt. Low trap risk, reasonable response rates, these people remember you.
- 6–12 months. Worth a smaller, careful attempt. Expect single-digit response and plan the sunset before you start.
- 12–24 months. Marginal. Only if the average customer value genuinely justifies it — which is a lifetime value calculation, not a hunch.
- 24 months and beyond. Usually not. High trap density, negligible response, and the reputation risk is carried by your whole programme rather than by this segment.
One important complication: you probably can't define engagement by opens any more. Privacy protections mean opens are inflated and unreliable as an individual signal, so a "hasn't opened in a year" segment may include people who read everything. Define dormancy by clicks, purchases, logins or site visits instead — the shift covered in what Apple and Gmail privacy changes mean for email marketers. Getting this wrong means you sunset engaged readers and keep dead weight.
Also set dormancy by your own cycle rather than a generic number. Ninety days is dormant for a weekly newsletter and perfectly normal for a business that sells something people buy twice a year.
Ramp, don't blast
Having cleaned and segmented, the send itself should be gradual and ordered by recency — freshest cohort first.
The freshest cohort is both most likely to respond and least likely to contain traps. Starting there means your early batches generate positive engagement signals, which is exactly what you want in the record before you send to the riskier portion.
Keep the re-engagement sending separate from your normal programme while this runs. Two reasons: you want to read the signals cleanly, and if it goes badly you don't want it dragging your core automated flows down with it. Where the volume justifies it, a dedicated subdomain gives you real isolation — noting that bulk sender thresholds count across subdomains, so this isolates reputation, not classification.
Make leaving easy on purpose
The counterintuitive part of the email itself.
A complaint damages your sender reputation. An unsubscribe doesn't. So during a re-engagement campaign, a prominent, obvious unsubscribe link is a feature — every person who takes it is a person who didn't hit "report spam" instead.
Practically, that means:
- A visible unsubscribe in the body, not only in six-point grey text in the footer. Plus the one-click header, which is a bulk sender requirement for marketing mail anyway, with unsubscribes processed within two days.
- A preference option alongside it. "Less often" retains people who'd otherwise leave entirely, and it's the single most useful addition to a win-back — the logic behind proper segmentation applied at the moment someone is deciding whether to stay.
- Say why they're hearing from you. "You signed up in March 2024 when you downloaded X" prevents the "who are these people" reaction that produces complaints.
- Ask for a click, not just an open. You need a signal you can actually read, and opens no longer are one. Make the click the point of the email.
On the subject line, resist the guilt-trip framing that dominates this genre. "We miss you" performs less well than a specific reason to look, and the principles in writing subject lines that get opened apply here more than anywhere — this is the least forgiving audience you will ever write to.
Decide the stop rule before you start
Almost nobody does this, and it's the difference between a controlled experiment and an incident.
Write down, before the first send, the conditions under which you halt:
- Complaint rate above 0.2% on any batch — pause, don't expand.
- Complaint rate approaching 0.3% — stop entirely and suppress the remaining segments.
- Hard bounce rate above 2–3% on a batch — stop, your cleaning didn't work.
- Any deferral or block message from a major provider — stop and investigate before another batch.
Agree this with whoever will be watching the revenue number, because the pressure to keep going comes from exactly that direction. A campaign that recovers 400 subscribers and costs you inbox placement for the other 30,000 is a bad trade that looks like a good one on the campaign report.
The infrastructure check before you start
Worth confirming rather than assuming, because a re-engagement campaign is the worst possible time to discover an authentication gap.
Bulk senders need SPF, DKIM and DMARC all present and aligned, TLS on the connection, valid reverse DNS, and RFC 8058 one-click unsubscribe on marketing mail. Microsoft has been rejecting non-compliant bulk mail outright since May 2025 with a permanent SMTP error rather than filtering it to junk — which means a failure here doesn't land you in spam, it means the message never arrives and you may not find out unless someone tells you.
Two things people commonly miss. The one-click unsubscribe header is the most frequent gap, and it applies to marketing mail rather than to transactional messages like receipts and password resets — you don't want an unsubscribe header on those. And a sudden volume jump from a domain that has been quiet reads as risk to filters that value consistency, which is a second reason the ramp above matters beyond complaint management.
Microsoft also weighs IP reputation more heavily than Google does, so if you're on shared sending infrastructure your placement there can suffer for reasons that have nothing to do with your own list. If you don't know which IPs your mail goes out on, find out before the first batch rather than during the third. The wider picture on how these signals interact is in how AI and privacy are reshaping the inbox, and the template side of it in building templates that render everywhere — a win-back that breaks in a client is a wasted send you only get one of.
What to do with the non-responders
The part people avoid because it feels like losing.
Anyone who doesn't respond to the sequence gets suppressed — moved to a list you never mail, not deleted. Deleting them means they can re-enter through a future import and you'll mail them again, which is how brands end up hitting the same traps twice.
Reframe what's happening: you're not losing subscribers, you're removing the dead weight that was suppressing placement for the people who do want you. A smaller list with strong engagement reaches the inbox; a large list with weak engagement doesn't, and the second one costs more to send. That's the core argument in landing in the inbox consistently.
Then make it structural. A sunset policy — automatic suppression after a defined period of no engagement — means you never accumulate a dormant list large enough to need a campaign like this again. It's a one-line rule that removes a recurring problem.
What to do instead of a big win-back
Worth saying, because for many brands the honest answer is that this campaign shouldn't be the priority.
Re-engaging a cold list is expensive, risky and low-yield. Keeping people from going cold is cheap, safe and compounding. If your dormant segment is growing faster than you can win it back, the problem is upstream: send frequency, relevance, or the fact that nothing arrives between the welcome sequence and the next campaign.
The flows described in post-purchase sequences and the broader logic of retention prevent dormancy far more cheaply than any win-back reverses it. If list health is a persistent problem rather than a one-off cleanup, that's an e-commerce lifecycle question rather than a campaign one.
And once the list is healthy, keep it that way by testing systematically rather than occasionally — the approach in an A/B testing programme that compounds is what stops engagement quietly decaying again.
The short version
Gmail measures complaints daily, wants you under 0.1%, and enforces at 0.3% — which at 25,000 sends is seventy-five people clicking one button. Clean and validate before you send anything, because sending first means discovering your dead addresses and traps by hitting them. Understand that dormant lists carry recycled spam traps precisely because dormant often means abandoned. Segment by recency and don't bother past about two years. Ramp in small batches from freshest to oldest, checking complaints and bounces between each. Make unsubscribing obvious, because an unsubscribe is infinitely better than a complaint. Write the stop rule down before you begin. Suppress the non-responders rather than deleting them, and set a sunset policy so this never becomes necessary again.
The campaign that recovers a few hundred subscribers and costs you the inbox for thirty thousand looks like a success on the report. That's the failure mode worth designing against.
Sitting on a big list you're afraid to email?
We clean, segment and ramp it properly — so you find out who's still there without risking the inbox.
Revive my list safely →Frequently asked questions
What spam complaint rate is safe when re-engaging an old list?
Google asks bulk senders to stay below 0.1% and treats 0.3% as the enforcement line, measured daily in Postmaster Tools. At 0.3% you become ineligible for delivery mitigation and typically need to sit under the threshold for several consecutive days to recover. In practical terms 0.1% is roughly five complaints per five thousand delivered messages, so a single large send to a cold list can breach the limit in one day.
Should you clean a list before or after sending a re-engagement campaign?
Before, which is the opposite of most advice. Sending first means you discover your dead addresses by bouncing off them and your spam traps by hitting them, both of which damage reputation before you have learned anything. Validating and removing obviously dead addresses first costs a small amount and removes the highest-risk portion of the list, so the campaign that follows is a test of interest rather than a test of your infrastructure.
Why are dormant email lists more likely to hit spam traps?
Because a recycled spam trap is an abandoned mailbox that a provider has repurposed to catch senders who do not practise list hygiene. A dormant list is largely made of addresses nobody has opened in a long time, which is the same population. Dormant does not only mean uninterested — it frequently means the mailbox no longer belongs to anyone, and mailing it is the exact behaviour the trap exists to detect.
How large should the first re-engagement send be?
Small, and to your most recently active dormant subscribers rather than the oldest. A few hundred to a low thousand, with a wait of a day or two to read complaint and bounce rates before expanding. The recency-ordered ramp matters because the freshest cohort is both the most likely to respond and the least likely to contain traps, so early batches build positive signal instead of spending reputation you have not yet rebuilt.
Is it better for someone to unsubscribe or to ignore your email?
An unsubscribe is far better than either a complaint or continued silence. Complaints damage your sender reputation directly; unsubscribes do not. Silence is worse still, because continuing to mail people who never engage drags down the engagement signals providers use for placement. Making the unsubscribe link genuinely easy to find during a re-engagement campaign reduces complaints, which is the trade you want.
