How many mailboxes per domain for cold email? Our rule after Google pushed back

Sunny Ray · Sep 17, 2026 · 6 min read

Our rule of thumb is about 20 mailboxes per domain, not 40, and no more than a few hundred new mailboxes created in a week. Go denser or faster than that and you start to look like what you are, a sending operation, and the providers react. We learned it by doing the opposite.

Here is what went wrong, the rule we use now, and the account-level risk most people never hear about until everything on it goes dark at once.

What went wrong

On our team call on September 15 we looked hard at our own sending setup. On paper we had about 1,000 mailboxes. In practice many of them were disconnected or sitting on domains that were not connected to anything. Some of the ones that were live were going to spam.

The cause was not the copy. It was not the list either. We had created too many mailboxes too fast on too few domains, and Google was flagging the volume.

That is an easy mistake to make when the business is working. Volume is the lever. More mailboxes mean more invites, more invites mean more conversations, and the marginal cost of another hundred mailboxes is close to zero. Charles had made exactly that point a few weeks earlier. So you push. And if you push on density and speed at the same time, you hit a wall.

About 20 per domain, not 40

The rule we settled on that day: about 20 mailboxes per domain. Not 40.

It is worth knowing how we got to 20 in the first place. On our August 27 team call Charles had recommended different densities by campaign type: about 10 mailboxes per domain for direct sales outreach, and about 20 for podcast invites, because an invitation carries lower spam risk and gets a better response. On that same call we were creating new mailboxes at just 3 per domain to start, with the plan to ramp slowly. At 10 emails per mailbox per day, a domain with 3 mailboxes sends about 30 a day.

Twenty is a ceiling, not a target. Start lower on a new domain and grow into it. The trouble came when some domains got pushed well past that.

A few hundred new boxes a week, maximum

Density is one half. Speed is the other. Creating hundreds or thousands of mailboxes in a burst looks nothing like a real company hiring people. Real companies add a handful of addresses at a time.

Our cap now is a few hundred new mailboxes a week at most. If you need thousands, you get there over weeks, not days. That is slower than you want when you have a target in mind. It is much faster than rebuilding after a provider decides you are a problem.

And every new mailbox needs time to warm before it sends live campaigns. At the time of that call we had a batch of about 75 mailboxes that had been set up more than two weeks earlier and were still warming. Plan for that lag.

The account-level kill switch

This is the part most people do not hear about until it happens to them.

Your domains and mailboxes live inside a hosting account. If you breach that account’s limits, the provider can suspend the whole account, not just the offending domain. With Namecheap, which we use, a suspended cPanel kills every domain on that account at once.

Think about what that means. You might have done everything right on most of your domains, and one overloaded corner of the account takes all of them down. Every mailbox, every warmed reputation, every live campaign on that account, gone in one action.

The protection is boring. Know the limits on each account. Spread domains across accounts rather than piling them into one. And treat the hosting account itself as a single point of failure, because it is.

Build a dashboard before you test

The other thing that came out of the call: we could not see the problem clearly. Our goal was 2,500 active mailboxes, and we were not going to get there by guessing which domains were healthy.

Charles started building a local dashboard to show bounce rate and opportunities by domain and by campaign. That sounds like a nice-to-have. It is not. Without it you cannot tell whether a dip in replies is a list problem, a copy problem or a domain problem, and every test you run is noise.

We had a cheap-domain test on the list too, trying low-cost domains to bring the cost down. We held it until the dashboard was there. Run tests when you can measure them.

Delete the dead ones

One more lesson from our August 27 call, because it is the kind of mess that builds up quietly. We had 151 existing mailboxes sitting paused that could not be activated. They had been created incorrectly earlier, and they were taking up namespace on domains we wanted to use properly.

The decision was to delete all 151 and recreate them the right way with the mailbox tool Amir had built. It felt wasteful in the moment. It was not. Broken mailboxes are worse than no mailboxes: they make your counts look healthier than they are, and they block the domain space you need for working ones.

If your sending dashboard shows a big number with a lot of grey in it, do the same audit. Count only what is live and healthy.

What the volume buys you

Why push on mailboxes at all? Because volume is the lever. On August 27 we had about 200 active mailboxes generating 20 to 30 leads a day, and the goal was 500 to 1,000. More volume forces better filtering, and it makes you better at the calls faster. Thirty calls a day is a different game from three.

That is exactly why the rule matters. The upside of more mailboxes is real, and so is the temptation to get there too fast. About 20 per domain, a few hundred new boxes a week, and patience on warmup is how you get the volume without losing the reputation that makes it worth anything.

The rule, summarized

• About 20 mailboxes per domain at most for invite campaigns. Lower for direct sales.
• Start new domains with a few mailboxes and ramp.
• Create a few hundred new mailboxes a week at most.
• Give every new mailbox time to warm before it goes live.
• Know your hosting account limits. A breach can suspend every domain on the account.
• Spread domains across accounts.
• See bounce rate and opportunities per domain before you run tests.

Or skip the plumbing

Every founder who runs cold outreach at volume ends up learning this. Most learn it the way we did. The reason we built our system to send from separate, warmed sending domains is so that your main domain and inbox stay clean while all of this happens somewhere else. The system sets up the domains, DNS and mailboxes for you, so you are not buying domains, counting mailboxes per domain or watching hosting limits yourself.

Frequently asked questions