Ask how many domains and mailboxes a cold email setup needs and you'll get a confident number back, usually without any working shown. The number is downstream of two things — how much you want to send, and how much a single mailbox can safely send — and once you have those, the rest is arithmetic rather than opinion.
More useful than any specific figure is understanding what the structure is actually protecting you from.
Start from volume, not from a recommendation
Work backwards from what you need.
If you want to reach 1,000 prospects a month with a four-step sequence, that's around 4,000 sends spread over the month, so roughly 200 on a working day. Divide that by whatever you consider a safe daily volume per mailbox and you have your mailbox count. Divide the mailbox count by however many mailboxes you're willing to put on one domain and you have your domain count.
That's the whole calculation. The judgement is entirely in the two divisors.
Why per-mailbox volume is kept low
Common guidance for cold outreach lands well below what a mailbox is technically permitted to send — typically a few dozen messages per mailbox per day rather than the hundreds a provider would allow.
The reason is that the technical limit isn't the constraint that matters. Reputation is. A mailbox sending a modest, steady volume to recipients who mostly don't complain looks like a person doing their job. The same mailbox sending several hundred cold messages in a burst looks like what it is, and the filtering that follows applies to the sending domain, not just that mailbox.
A newer domain warrants more caution than an established one, and ramping up over weeks rather than starting at full volume is the standard approach for the same reason: sudden changes in sending pattern are themselves a signal.
Why the mailboxes-per-domain number is the important one
This is the part people get wrong, and it's worth being precise about.
Domain reputation is shared across every mailbox on that domain. If the domain gets flagged, every mailbox on it degrades together. So the number of mailboxes you put on a single domain is really a decision about blast radius.
Put ten mailboxes on one domain and you've built a single point of failure carrying your entire outbound operation. Spread the same ten across four or five domains and a problem on one costs you a fraction of your capacity while you diagnose it. The usual guidance is a small number per domain — two or three — and the reasoning is concentration risk rather than any per-domain sending cap.
Never send cold outreach from your primary domain
If there's one rule here that isn't a judgement call, this is it.
Your primary domain carries your transactional email: password resets, receipts, invoices, notifications, and the replies of everyone in the company. Damage to its reputation doesn't just slow your outreach, it means customers stop receiving things they actually need.
Cold outreach goes on separate domains bought for the purpose, usually close variants of your brand — the .co of your .com, or a hyphenated or prefixed version — so replies still look legitimate to a recipient who checks. Those domains are expendable in a way your primary never is.
Every domain needs the same authentication
A point that gets skipped when someone buys several domains at once: each one is a separate identity to mailbox providers, and each needs SPF, DKIM and DMARC configured independently.
One unauthenticated domain in a set of five doesn't fail quietly in isolation. It sends mail that fails authentication checks, which is among the strongest signals available that something is off. Buying domains is the easy part; configuring each one correctly is the part that determines whether the setup works.
Same applies to warmup. Each new domain and mailbox starts with no history and needs to build one before carrying real volume.
Where setups actually go wrong
Over-building before there's volume to justify it. Ten domains and thirty mailboxes for a list of 500 prospects is expensive, slow to warm up, and adds administrative surface for no gain. Build for the volume you have, then extend. Treating it as one-time setup. Domain reputation moves. A setup that worked in March can be quietly degrading by August, and without monitoring you find out from a drop in replies rather than from a signal. Concentrating anyway. Buying five domains and then putting most of the load on one of them recreates the risk you bought five to avoid. Spread the sending, not just the registrations. Reusing burned domains. A domain with a damaged reputation doesn't reset because you paused it for a month. Sometimes the honest answer is to retire it. Ignoring the reply side. These mailboxes need to be monitored. An outreach domain nobody reads is a way to miss the interested replies that justified the whole exercise.The part that matters more than the structure
All of this infrastructure exists to protect reputation, and reputation is mostly determined by whether recipients want your messages. Complaint rate is the strongest input, and complaints come from people who had no plausible reason to be contacted.
So the infrastructure is a floor, not a strategy. You can build a perfectly-structured five-domain setup and still burn it in a month by emailing a badly-targeted list, and no amount of domain separation compensates for that. Conversely, well-targeted outreach at modest volume rarely stresses the setup at all.
That targeting problem is the one worth the most attention, and it's what [AutoReach](/register) is built around — finding businesses that genuinely match your ICP and scanning them for real context, so the people receiving your outreach have an actual reason to read it. Get that right and the infrastructure question gets much easier, because you need less of it.