How many mailboxes do you need for cold email? A sizing formula
A step-by-step formula for sizing cold email infrastructure: daily volume, safe sends per mailbox, domains per mailbox and the buffer you actually need.
Ask ten cold emailers how many mailboxes you need and you will get ten confident answers, most of them a version of "25 a day per mailbox, three mailboxes per domain." Those rules of thumb exist because they are easy to remember, not because they are right for your list, your provider mix or your reply goals. This post gives you a formula you can actually reason about, then shows where the rules of thumb break.
Start from the outcome, not the mailbox
Nobody wants mailboxes. You want replies, and replies come from a certain number of delivered emails landing in the primary inbox. So the first number to write down is not "how many mailboxes" but how many emails do I need to send per working day to hit my pipeline target.
Work backwards:
- Meetings you need per month. Say 20.
- Positive reply rate you expect. Be honest. For a well-targeted list with decent copy, plan on a low single-digit percentage of delivered emails turning into positive replies.
- Emails delivered per month. Meetings divided by positive reply rate. At 20 meetings and a 1% positive rate, that is 2,000 delivered emails.
- Sequence length. Most sequences send three to five touches per prospect, so total sends are two to five times the number of prospects.
Only now do you have a daily send target. Divide monthly sends by roughly 21 working days and you have the number the rest of the formula depends on.
The formula
Here it is in one line:
Mailboxes needed = (daily sends ÷ safe sends per mailbox) × buffer
And the follow-on:
Domains needed = mailboxes ÷ mailboxes per domain
Each of the three inputs deserves a paragraph, because the defaults people copy from Twitter threads are where the sizing goes wrong.
Safe sends per mailbox
This is the number of cold emails one mailbox can send per day without reputation drifting toward the spam folder. It is not a constant. It depends on the provider (Google, Microsoft and independent SMTP behave differently), on how well the mailbox was warmed, on how clean your list is and on how engaged the recipients are.
A conservative planning figure for a warmed mailbox is 20 to 30 cold sends per day, on top of whatever warmup traffic it continues to exchange. If you are new to cold email or your list is unverified, plan at the low end. If you monitor placement per mailbox and can pull under-performers out of rotation, you can safely push individual mailboxes higher, which is exactly what a rotating pool does for you automatically.
Mailboxes per domain
Every mailbox on a domain shares that domain's reputation. If one mailbox gets flagged, the others are dragged with it. The usual guidance of two or three mailboxes per domain is about limiting the blast radius, not about any hard provider limit. More mailboxes per domain means fewer domains to buy and configure but more correlated risk.
Buffer
The number most spreadsheets leave out. Mailboxes get paused, warmup takes time, a domain lands on a blocklist, someone's inbox provider has a bad week. If you size infrastructure to exactly your daily target, any one of those events drops your volume below plan. A 20 to 30% buffer is a sensible starting point. If cold email is your primary channel, go higher.
A worked example
A team wants to send 1,500 cold emails per working day. They plan for 25 safe sends per mailbox, three mailboxes per domain and a 25% buffer.
| Input | Value |
|---|---|
| Daily sends | 1,500 |
| Safe sends per mailbox | 25 |
| Raw mailboxes needed | 60 |
| Buffer | 1.25× |
| Mailboxes to provision | 75 |
| Mailboxes per domain | 3 |
| Domains to provision | 25 |
Seventy-five mailboxes and twenty-five domains, all warmed before they send a single cold email, all monitored so that the weakest ones can be swapped out. That is the real footprint behind a "modest" 1,500 sends a day, and it is a good illustration of why people quietly halve their volume once they see the bill.
Where the formula breaks
A formula is only as good as its inputs, and the inputs move.
Volume is not constant. Sequences ramp, campaigns end, lists get exhausted. If you provision for peak, you overpay most of the month. If you provision for average, you cannot hit peak. Infrastructure that can be scaled up and down without a procurement step solves this in a way no spreadsheet can.
Mailboxes are not interchangeable. Two mailboxes on the same provider, warmed the same way, can perform very differently after a month of sending. Treating every mailbox as worth 25 sends leaves volume on the table from the strong ones and burns the weak ones.
Warmup is a cost in time, not just money. A fresh mailbox needs weeks before it should carry full cold volume. If you size correctly but only start warming when you need the capacity, you are still late by weeks. Provision ahead of need or use a pool that already has warmed capacity waiting.
The buffer needs to be replenished. When you pull a mailbox from rotation because placement dipped, your buffer just shrank. Somebody has to build the replacement, warm it and slot it in. In a hand-managed setup that is a weekly chore; in a managed pool it happens in the background.
A simpler way to think about it
If you are buying and managing infrastructure yourself, run the formula, add the buffer, and re-run it every time your volume plan changes. It will keep you honest about what the volume actually costs in mailboxes, domains and hours.
If you would rather not own the spreadsheet at all, the alternative is to buy delivered sends instead of mailboxes. Sender Pools builds the mailboxes and domains for you, warms them, monitors placement per mailbox and cycles under-performers out, so the sizing question turns into a single number: how many emails do you want delivered, and by when? You can see how that pricing works on the pricing page.
