Every team running cold email automation at volume eventually sees the same pattern: reply rates that held steady for weeks fall by half in a few days, and the dashboard still says "delivered." Nothing bounced. The messages simply stopped reaching the primary inbox. By the time someone notices, three weeks of pipeline are already gone.
A damaged sending domain is recoverable more often than most operators assume, but only if you stop the damage first and follow a sequence rather than improvising. Retiring a domain too early wastes a clean asset. Rehabilitating one that is beyond saving wastes a month. This guide gives you a decision framework, a four-week recovery protocol, and the controls that keep it from happening again.
Before you touch anything, rule out the cheaper explanations. A reply-rate drop has at least four common causes, and only one of them is a burned domain.
- Authentication failure. A DNS change, a new sending tool or an expired record can break SPF, DKIM or DMARC alignment overnight. Check this first. Our SPF, DKIM and DMARC setup guide walks through the records to verify. - List quality. A fresh batch with a high invalid-address rate produces bounces that look like a reputation problem. Bounce rates above roughly 2% on a send are a list issue first. - Copy fatigue. If every mailbox sends the same template to the same segment, filters learn the fingerprint. The drop may follow the template, not the domain. - Provider policy changes. Gmail and Outlook tightened bulk-sender enforcement in recent cycles; see our breakdown of the Gmail and Outlook deliverability changes.
To isolate reputation, run a seed test: send the same plain message from the suspect domain and from a known-clean domain to a set of test inboxes across Gmail, Outlook and one or two smaller providers. If the clean domain lands in the inbox and the suspect domain lands in spam or promotions, the problem is the domain. Also check Google Postmaster Tools for domain reputation and spam rate. Google's published guidance asks bulk senders to keep the user-reported spam rate under 0.3% and, ideally, below 0.1%. A domain sitting above 0.3% is in enforcement territory.
The most expensive mistake in recovery is continuing to send "just a little" while you diagnose. Every additional message to an engaged-negative audience deepens the damage.
Do the following immediately:
1. Pause all campaign sending from the affected domain. Not a reduction, a full pause. 2. Pull the list of every contact the domain emailed in the last 14 days and suppress anyone who bounced, complained or unsubscribed across all your domains, not only this one. 3. Check for a blacklist listing on the major public lists. A listing on a high-impact blocklist changes the recovery plan, because you must resolve the listing cause before delisting requests will hold. 4. Move the affected mailboxes' scheduled follow-ups to healthy domains only if the sequence is high-value and the contact is verified. Do not simply dump a paused queue onto a clean domain; that transfers the problem.
Our platform does the pause and suppression automatically when a domain crosses a bounce or complaint threshold, which is the practical argument for deliverability monitoring with alerts. If you run this manually, assign an owner and a daily check, because the people who notice the problem are rarely the people who control the sending.
Not every domain deserves a recovery attempt. Use this scoring rule.
Rehabilitate when: - The domain is more than 60 days old and had a clean history before the incident. - The drop was gradual and tied to an identifiable cause such as a bad list or an over-aggressive volume ramp. - Postmaster reputation is "Low" or "Medium," not "Bad." - No major blocklist listing exists, or the listing has a clear, fixed cause.
Retire when: - The domain is under 30 days old. Time you spend rehabilitating it is better spent warming a replacement. - Spam complaints were sustained over several weeks. - Reputation is "Bad" on Gmail and inbox placement is under roughly 30% across providers. - The domain is also your primary brand domain. Never rehabilitate the domain your employees use for regular business mail; protect it by keeping cold sending on separate lookalike domains from the start.
A practical shortcut: if replacing the domain costs less than two weeks of one person's time, replace it. Domain and mailbox costs are small compared with the pipeline value of a functioning sending infrastructure. The economics are covered in our cost per meeting analysis.
If you decided to rehabilitate, treat it as a warmup from near zero, not a gentle resumption.
Week 1: Fix and rest. Resolve the root cause (authentication, list source, template). Keep all campaign sending paused. Re-enable only low-volume warmup traffic with real two-way conversation between your own mailboxes and trusted external addresses. Target 10 to 15 messages per mailbox per day.
Week 2: Rebuild engagement. Continue warmup and add a handful of messages per day to your most engaged contacts: recent repliers, existing customers, people who have opened and clicked in the past. Positive engagement signals are what move reputation, and these are the recipients most likely to provide them. Keep volume under 20 messages per mailbox per day. The mechanics mirror our warmup playbook for domains and mailboxes.
Week 3: Controlled cold reintroduction. Resume cold sending at roughly 25% of your previous volume, only to contacts that are verified and in your highest ICP tier. Plain text, no tracking links, one short ask. If you use ICP scoring, send only to the top scoring band. Watch placement and complaint rate daily.
Week 4: Step up, with gates. Increase to 50% of prior volume only if three conditions hold: bounce rate under 2%, no complaint spikes, and seed-test placement back above 80% in the inbox. If any gate fails, hold volume for another week rather than pushing ahead.
Most domains that are recoverable show measurable improvement in Postmaster data within two to four weeks. If you see no movement after four weeks of clean behavior, retire the domain and move on. Sunk cost is the enemy here.
Recovery is a symptom of missing guardrails. The teams that rarely burn domains share a few habits:
- Volume caps per mailbox, typically 30 to 40 cold sends per day, spread across more domains rather than more messages per domain. - Automatic pause rules on bounce rate, complaint rate and reply sentiment. A burst of hostile replies is an early warning that arrives before Postmaster data does. - Domain rotation with reserve capacity. Keep 15 to 20% of your domains warmed and idle so a retirement doesn't interrupt campaigns. - Weekly audits. Run the checks in our deliverability audit framework on a schedule rather than after an incident. - Targeting discipline. Spam complaints come overwhelmingly from irrelevant recipients. Tight targeting is the cheapest deliverability tool you have.
This is also where manual SDR workflows struggle. A human rep has no visibility into per-domain complaint rates and no reason to check them, which is part of why automated prospecting with built-in monitoring tends to protect reputation better than ad hoc sending.
- Sending a "re-engagement blast" to the whole list to prove the domain still works. This is the fastest way to turn a medium reputation into a bad one. - Changing everything at once. If you swap domain, copy, list and tool in one move, you will never know what worked. - Buying warmup volume to mask the problem. Artificial engagement without real replies does not repair a complaint-driven reputation. - Ignoring the list source. If the same vendor supplied the contacts, the next domain will burn too.
A burned domain is a process failure you can contain. Confirm the cause, stop sending within 24 hours, score the domain honestly for rehabilitation versus retirement, and follow a gated four-week ramp if it's worth saving. Then put pause rules, volume caps and reserve domains in place so the next problem is caught in days, not weeks.
If you'd rather not run these controls by hand, OnyxSend monitors every sending domain, pauses risky ones automatically and rotates volume to healthy infrastructure. You can see OnyxSend pricing and start with a small pilot to watch the safeguards work on your own list.