How many mailboxes do you need for cold email?

Work it out from the bottom up: the daily sends your campaign needs, divided by a deliberately conservative per-mailbox daily volume, plus spare capacity. That is the whole calculation. The mistake that burns domains is dividing by the provider's per-mailbox ceiling instead, because the ceiling is a limit imposed by the platform, not a volume you can sustain into a cold audience.

There is no universal number for sustainable per-mailbox volume, and anyone who gives you one without asking about your list is guessing. It depends on your bounce rate, your complaint rate, the mailbox's age, how long it has been sending, and which providers your recipients use. What you can do is start low, raise it in small steps, and let per-mailbox reputation and measured inbox placement tell you when to stop.

The arithmetic

1. Start from replies, not sends. Decide how many conversations a month the campaign has to produce. Divide by your observed reply rate for that segment to get required sends. If you do not have an observed reply rate yet, run a small pilot rather than modelling one.
2. Convert to sends per working day. Spread across your actual sending days — not calendar days, and not weekends.
3. Divide by a conservative per-mailbox daily figure. Low enough that losing any one mailbox is an inconvenience rather than an outage.
4. Add spare capacity. Enough warm, idle mailboxes that you can pause one or two without pausing the campaign. Pausing a mailbox is a thing you will do.
5. Add nothing for follow-ups. Count follow-up messages as sends in step 2. A three-step sequence to 500 people is up to 1,500 messages, and people routinely forget the multiplier.
6. Check the per-domain shape. Spread the mailboxes across a small number of outreach domains, a handful each, rather than piling them all onto one.

Why the provider ceiling is the wrong divisor

Three reasons, all of which cost you more than the mailboxes would have.

Warm-up time. A new mailbox cannot carry its target volume for weeks. If your plan requires every mailbox at full volume from week one, the plan fails in week one.

Rolling reputation. Reputation is a judgement about recent behaviour and it decays. A mailbox at its ceiling has no headroom when something changes — a bad list segment, a volume jump, a provider-side policy change — and the only lever left is to stop.

Correlated failure. Mailboxes ramped identically, in the same tenant, on the same infrastructure, in the same week, are not independent. Running them all at the ceiling means they all cross the same line at the same time.

More mailboxes do not fix the problems people buy them for

A low reply rate is a targeting or copy problem. Pushing the same message through forty senders instead of ten multiplies the complaints, not the replies, and Google asks bulk senders to keep reported spam under 0.3%.

A high bounce rate is list rot. Verify addresses, drop role accounts and unverifiable catch-alls. No mailbox count changes the quality of an address.

Poor placement is a reputation, authentication or content problem. Adding senders to a domain whose placement is already drifting attaches new identities to a reputation heading the wrong way.

The honest version of "how many mailboxes do I need" often turns out to be "fewer than you have, sending to a smaller list".

Each mailbox has a running cost beyond the licence

• Warm-up, before and during live sending, continuously.
• Monitoring, per mailbox: reputation, bounces, complaints, placement.
• A monitored reply inbox. Mail arrives: replies, out-of-office notices, bounce messages. A mailbox nobody reads is a bad signal and a missed pipeline.
• A share of a domain's reputation. Mailboxes on one domain are not independent of each other.

That is why the answer is not "buy as many as you can afford". Each one is a thing you have to keep healthy.

What not to do

Do not divide by the platform's maximum. It is the number at which sending stops, not the number at which it works.

Do not put every mailbox in one tenant. A single outbound restriction then stops everything at once.

Do not use your primary company domain for any of them. A sending problem there takes your invoices, password resets and support replies with it.

Do not create them all on the same day. Stagger creation and warm-up so mailboxes have different ages and different histories.

Do not buy mailboxes described as pre-warmed. Reputation comes from a sending identity engaging with real recipients over time. A mailbox warmed inside a vendor's pool before you owned it has no useful history, and one that has already sent cold mail for someone else has the wrong kind.

Do not forget the sequence multiplier when you size the fleet.

Do not size from sequencer "delivered" counts. Delivered means a receiving server returned a 250; it cannot tell you the mail went to spam, in which case your real capacity is zero regardless of mailbox count.

Where MailPilot fits

The number you need depends on a per-mailbox volume you have to measure rather than assume. MailPilot warms every mailbox over OAuth on Gmail and Outlook, or over SMTP/IMAP for SendGrid, Amazon SES, Mailgun, Postmark and custom SMTP, tracks reputation per mailbox daily, checks SPF, DKIM and DMARC per sending domain, and measures real inbox placement per receiving provider through a seed-inbox network — so you can raise volume on evidence and stop before a mailbox turns into a liability. Every plan includes unlimited mailboxes, so the fleet size is your decision rather than a billing one.