This is the first calculation to get right when setting up cold email, and the one most often done backwards. People pick a mailbox count based on budget, then send whatever volume that supports. The correct order is the reverse: your target volume determines the mailbox count, and the mailbox count determines the budget.
The one number everything hangs on
A cold email mailbox can safely send 30 to 50 emails a day.
Google and Microsoft allow far more. That is the point people miss: the provider limit is a technical ceiling, not a safety threshold. What gets mailboxes flagged is the pattern, consistent volume to recipients who have never replied, from an account with no other activity. Staying at 30 to 50 keeps you in territory that looks like a person doing outreach rather than a machine.
Use 40 as your working number. Use 30 if the domain is new or the list is cold-sourced.
The math
Three steps.
1. Mailboxes = daily sends ÷ 40.
2. Add 30% headroom.
3. Domains = mailboxes ÷ 3.
| Target sends/day | Base mailboxes | With headroom | Domains | Monthly sends |
|---|---|---|---|---|
| 100 | 3 | 4 | 2 | ~2,200 |
| 250 | 7 | 9 | 3 | ~5,500 |
| 500 | 13 | 17 | 6 | ~11,000 |
| 1,000 | 25 | 33 | 11 | ~22,000 |
| 2,500 | 63 | 82 | 28 | ~55,000 |
| 5,000 | 125 | 163 | 55 | ~110,000 |
Monthly figures assume 22 working days. Sending at weekends is possible but reply rates are poor enough that most teams do not bother.
Why the headroom is not optional
Three reasons, all of which cost you if ignored.
Warmup rotation. New mailboxes take three to four weeks before they carry full volume. If you have no spare capacity, adding mailboxes means a month of reduced sending.
Attrition. Some mailboxes get flagged. Not many, but enough that assuming zero is optimistic. Headroom means a flagged mailbox is an inconvenience rather than a gap in the schedule.
Ceiling avoidance. Every mailbox sending exactly its maximum every single day is a pattern. Running at 70 to 80% of capacity looks more like human behaviour and gives you room for a heavier week.
Why three mailboxes per domain
Domain reputation is shared by every mailbox on it. Put twelve mailboxes on one domain and a single bad campaign takes out all twelve at once, along with the domain, which is the harder thing to replace.
Three per domain spreads the risk. More domains costs more, roughly $12 to $14 each per year, but a burned domain costs a month of warming and the campaigns that were running on it.
One exception: if you are sending very low volume from a domain you also use for real business, keep cold outreach off it entirely. Use separate domains for outbound, always.
What this costs
Take 1,000 sends a day, which is 33 mailboxes and 11 domains.
| Line | Google / M365 at $4 | Mailboundry from $2 |
|---|---|---|
| 33 mailboxes / month | $132 | ~$66 |
| 11 domains / year | ~$150, bought in-app up to $14 per TLD | |
| Dedicated IPs | Shared provider infrastructure | One per mailbox, included |
| Warmup | Unlimited, included in the plan | |
Which model to pick depends on recipients rather than budget alone. Real Google and Microsoft mailboxes carry more inherent trust and suit enterprise targets. Private infrastructure is cheaper per send and gives you reputation isolation, which matters more as volume grows. Plenty of teams run both.
Two mistakes worth avoiding
Buying all the mailboxes at once. Fifty new mailboxes all warming simultaneously is a month of near-zero output and a suspicious pattern. Add them in waves.
Sizing for your best month. If volume moves, size for the normal month and use monthly billing so you can scale up when a campaign needs it rather than paying for peak capacity year-round.
Next: how fast to ramp new mailboxes, or the full per-mailbox cost breakdown.
Frequently asked questions
How many emails can one mailbox send per day?
Technically the provider limit is much higher, but the safe cold email limit is 30 to 50 per mailbox per day. Beyond that you are relying on the provider not noticing a pattern that looks automated, and eventually they do.
How many mailboxes should I put on one domain?
About three. More concentrates risk, because a reputation problem on the domain takes out every mailbox on it simultaneously. Fewer mailboxes across more domains is the safer shape.
How much headroom should I buy?
Around 30% above your calculated need. Mailboxes rotate out for warmup, a few get flagged, and running every mailbox at its daily ceiling is itself a pattern worth avoiding.
Do I need to warm every new mailbox?
Yes, and budget three to four weeks before it carries full volume. Buying pre-warmed mailboxes shortens this but does not remove the need to ramp your own sending gradually once you take them over.
Is it better to have more mailboxes sending less each?
Generally yes. Spreading the same volume across more mailboxes lowers the per-mailbox rate, which is the number providers scrutinise, and it limits the damage when one mailbox has a problem.