Reply rates fall off a cliff. Test sends to your own Gmail account land in spam. Bounce messages start to mention policy or reputation. Something has gone wrong with how mailbox providers see you, and every email you send while it stays wrong makes it worse.
This is a field manual for that situation, in working order: stop, diagnose, reduce, fix, and only then decide whether the domain is worth keeping.
One warning first. Nobody can promise a recovery date. Providers change their view of a sender when they see sustained better behaviour, on their own schedule.
Step 1: stop the bleeding
Pause every campaign on the affected domain today, not after the current batch finishes. Then:
- Do not move the campaign to another mailbox on the same domain. Reputation attaches to the domain, so this spreads the damage.
- Do not move the same list and copy to a fresh domain either. Whatever caused the problem will travel with it.
- Keep warmup and normal human correspondence running. Going silent and then restarting creates the burst pattern filters distrust, and wanted mail is the one thing still working in your favour.
Step 2: diagnose before you change anything
Authentication
Send a message from the affected mailbox to a Gmail account you control and choose "Show original". Check that SPF, DKIM and DMARC all pass, and that the passing domain is your own.
Common breakages: a second SPF record added for a new tool, a DKIM key never switched on, an SPF record past the 10 DNS lookups the SPF standard allows, or a DNS migration that dropped records. How to set up SPF, DKIM and DMARC has the correct values, and the free domain checker shows what your DNS returns now.
Blocklists
Check your sending domain against the major public blocklists, Spamhaus first. Check any domain that appears in your links too, including redirects. A listed link domain can sink a message from a clean sender.
If you are listed, read the reason and fix the cause before you request removal.
Google Postmaster Tools
If you send enough volume to Gmail for data to appear, open Postmaster Tools and look at the spam rate by date and at the compliance status dashboard. Find the day the rate rose, then what you sent just before it. For mail to Outlook and Hotmail addresses, Microsoft's Smart Network Data Services gives a similar view by sending IP, when you send from IPs you control.
Bounce messages
Read the bounce text, not just the count:
- "User unknown" or "address does not exist" means the list is bad.
- References to policy, reputation, blocked content or rate limits mean the provider is refusing you as a sender, and the message often names the reason.
Complaint sources
Break down unsubscribes, angry replies and spam-rate spikes by campaign, list source, sequence step and sending mailbox. The cause is often concentrated: one purchased list, one aggressive fourth follow-up, one segment the offer never fitted.
What changed
List every change from the two weeks before the drop: a new list source, a jump in volume, a new template, a new link, a mailbox added without warmup. One of those is your lead suspect.
Step 3: reduce and re-ramp
Once the cause is found and fixed, restart at a small fraction of your previous volume.
- Send first to your best-fit, most recently verified prospects. You need this mail to be wanted.
- Use plain text, with no links and no images in the first email. Run the draft through the spam checker before it goes out.
- Raise volume in small steps. After each one, watch bounces, spam placement on your test accounts, and the Postmaster Tools spam rate.
- If any of them worsens, hold or step back.
It is the same discipline as warming a new domain, with less benefit of the doubt. Domain warmup: what actually moves sender reputation explains the signals you are rebuilding.
Fuzegenai applies a version of this automatically. A mailbox whose reputation drops to BAD keeps sending a reduced share of its daily volume while warmup continues, and a mailbox at VERY_BAD is paused for campaigns. How Fuzegenai checks an email before it sends has the detail, and mailbox health and sender reputation explains the grades.
Step 4: fix the list and the copy
A re-ramp with the same inputs produces the same result.
The list
- Re-verify every address before it goes back into a campaign.
- Drop role accounts such as
info@andsales@, and hold back catch-all domains until you are healthy again. - Suppress everyone who bounced, unsubscribed, complained or asked you to stop, permanently and across all campaigns.
- Stop using the source that caused the trouble.
The copy
- Cut links to one at most, and none in the first email.
- Remove images, attachments and heavy HTML signatures.
- Rewrite subject lines that overpromise or imitate a reply thread.
- Shorten the sequence. Every extra follow-up to someone who has ignored you is another chance at a complaint.
- Make the opt-out obvious. An easy exit is cheaper than a spam report.
The anatomy of a cold email sequence covers this in more depth.
When to retire a domain
Sometimes the domain is not worth saving. Consider retiring it when:
- Authentication is correct, the list and copy are fixed, you have re-ramped patiently, and clean plain-text mail still lands in spam.
- The domain sits on a major blocklist that will not remove it.
Retiring a domain is not a reset button. The new one starts with no history and needs a full warmup. If the list, copy and volume habits move across unchanged, it ends up in the same place.
This is also the argument for never running cold outreach from your primary company domain. With separate outreach domains, the worst case costs you an outreach domain and not your invoices, support replies and password resets.
The takeaway
Reputation damage is a symptom. The cause sits in one of four places: authentication, list quality, content, or volume. Stop sending, find which one it was, fix it, and come back slowly with your best mail to your best prospects. Skipping the diagnosis and buying a new domain only repeats the problem.