Cold email suddenly getting zero replies? Our postmortem

Sunny Ray · Sep 27, 2026 · 6 min read

When cold email replies drop to zero overnight, it is almost never the copy. It is the sending setup. Check, in this order: did anyone change the sending limits, did several people touch the same account, did you add a large batch of new mailboxes, and is a major provider like Outlook blocking your domains? In our case the answer to all four was yes, at the same time, and that is exactly why it was hard to see.

This is the blameless postmortem of the day our replies went to zero, and the change-control rule we adopted because of it.

The symptom

September 25, a few minutes before our weekly webinar. Charles and I were on the call early and I was looking at our sending account. Zero replies. Not low. Zero.

The campaigns were running. Emails were going out. Nobody was answering. When something goes from working to nothing that fast, you want one cause. We had five.

Five contributing factors

1. Emails per mailbox crept up to 20

Somewhere along the way, the setting for how many emails each mailbox sends per day had been increased to 20. Nobody had decided that on purpose. It had been changed by accident. For context, a few weeks earlier we were running about 10 per mailbox. Doubling the per-mailbox volume across the account is exactly the kind of jump providers notice.

2. Several people were changing one account

Multiple people on the team had been making changes to the same sending account at the same time. Each change made sense to the person making it. Nobody had the full picture of what had changed overall. That is how the 20 got in without anyone owning it.

3. 2,400 new mailboxes were warming

We had 2,400 new Namecheap mailboxes in warmup, not yet active. On its own, warming new mailboxes is normal. Doing it at that scale, on top of everything else, added to the load and to the risk.

4. Outlook appeared to have banned our domains

Outlook appeared to have blocked my domains entirely. If a big chunk of your list is on Microsoft mail, a block like that alone can make replies fall off a cliff.

5. A possible name-level flag

Charles had one more theory: that the sending platform might have flagged my name itself across accounts. We could not confirm it on the call, so we treated it as a hypothesis to check, not a fact. It went on the list of things to investigate.

The account-level limit nobody tracks

There was a sixth issue that was not causing the zero, but could have caused something worse. Our Namecheap account has a limit of 300 mailboxes per cPanel. Having 2,400 on one account is a red flag. If the provider decides you have breached the limits, it can suspend the cPanel, and every domain on that account goes down with it.

We had written about this risk ourselves a week or so earlier, after another deliverability scare. And here we were, with a much bigger version of it. That is the honest part of a postmortem. Knowing a rule is not the same as having a system that enforces it.

Recovery

The good news from the call: recovery was possible within hours if the mailboxes were reset. Put the per-mailbox volume back where it should be, sort out the overloaded account, and let the healthy mailboxes do their work.

Charles took on the investigation: check the cPanel limits, assess the name-level theory, and decide whether to remove or restructure the 2,400 warming mailboxes. I owed him the deliverability report I had pulled the day before so he had the full picture.

Change control: one gatekeeper

The real fix was not technical. It was a rule.

No more uncoordinated changes to live sending accounts. Every change to a live account goes through one person. We named the candidates on the call: Charles, Amir or me. One of us owns it, everyone else asks.

That sounds like bureaucracy for a small team. It is the opposite. A single gatekeeper is how a small team moves fast without breaking the thing that pays for everything else. It means that when something does go wrong, one person knows every change that was made and when. The postmortem is a short conversation, not a day of digging.

Credentials are part of change control

Change control is not only about who clicks the buttons. It is also about who can. Ten days earlier, on our September 15 team call, we had been passing hosting and sending-platform credentials to each other over WhatsApp and deleting the messages afterward. Everyone agreed that was a stopgap. The right fix was a shared password manager that also handles the login codes for the hosting provider, so nobody has to forward a code at the wrong moment.

When access is scattered, changes are scattered too. Putting credentials in one managed place is the quiet half of having one gatekeeper.

Pause before you scale again

The other lesson is about timing. A few days before the zero-reply afternoon, on a September 21 client call, we had already put further mailbox expansion on hold for about two weeks, because pushing too far risked getting flagged by the hosting provider. That was the right call. The trouble was that the warming batch already in flight was large enough to be a risk on its own.

When you decide to pause, check what is already in motion. A pause on new mailboxes does not undo the thousands that are still warming.

Blameless, on purpose

None of the five factors was anyone’s fault in the way people usually mean it. Each change was made by someone trying to help. The accident with the per-mailbox setting was exactly that, an accident. The warming batch was part of a plan to grow. Outlook does what Outlook does.

What failed was the system: no owner for the account, no log of changes, no limit checks before scaling. If the system allows a mistake, someone will eventually make it. Fix the system, not the person.

Your checklist when replies hit zero

1. Check the per-mailbox daily sending limit. Has it changed?
2. Ask who has touched the account in the last week, and what they changed.
3. Look at how many new mailboxes are warming, and on which accounts.
4. Check whether a major provider is blocking your domains.
5. Check your hosting account limits against your mailbox count.
6. Reset what you can, and expect recovery in hours if the mailboxes are healthy.
7. Then put one person in charge of every future change.

Or let someone else own it

Every team running cold email at volume has a day like this eventually. The difference is whether one person is watching the whole system. When you run your show on our system, it sets up the domains, DNS and mailboxes for you, and invites go out from separate, warmed sending domains, never your main domain. Your real inbox stays clean.

Frequently asked questions