How many cold emails should you send per inbox per day?

Microsoft 365 permits 10,000 recipients per mailbox per day. Google Workspace permits 2,000 messages and 3,000 external recipients per user per day. All of those numbers are accurate, and using any of them as a target is the fastest way to burn a domain.

Platform limits exist to stop abuse of the platform. They are not deliverability guidance. The number that governs cold email is the volume at which a mailbox stops resembling a person using email, and that number is far lower than any published cap.

The answer

Three cold emails per Microsoft 365 mailbox per day, on a mailbox that has been warmed. New mailboxes should start at one or two. Google Workspace tolerates considerably more - twenty-five to thirty per inbox is common there - so the right number depends on which platform you are sending from, and this figure is the Microsoft one.

That is a fraction of one percent of what the platform would permit. The gap is the point: the constraint is not throughput, it is pattern recognition on the receiving side. Hitting a real volume target therefore means more mailboxes across more domains, which is a domain strategy question rather than a throughput one. Sequence length multiplies it further, since every follow up is another send against the same daily budget. Anyone doing this per client rather than once will want the agency version of the calculation, where the mailbox count repeats with every account.

What the receiving side is actually measuring

Spam filters are not counting your messages against a published threshold. They are building a behavioural profile of the sending mailbox and asking whether it looks like a person or a machine.

• Volume consistency - a mailbox that sends 3 a day for a month then 300 in an hour is announcing itself.
• Reply rate - real correspondence gets replies; a mailbox that never receives one looks like broadcast.
• Engagement - opens, replies, messages moved out of spam, senders added to contacts.
• Complaint rate - the single fastest way to destroy a domain. Above roughly 0.3% and placement collapses.
• Bounce rate - high bounces indicate a list that was never verified, which correlates strongly with spam.
• Content variance - a hundred byte-identical messages is a signature no volume limit will save you from.

None of these improve by sending more. Several actively worsen. This is why the answer to "how do I send more" is never "increase per-mailbox volume".

Scale comes from mailbox count

If you need 300 cold emails a day, there are two ways to get there and they are not equivalent.

Both produce 300 sends. Only one survives contact with a spam filter. The arithmetic of outbound is therefore mailbox count multiplied by a small safe number, not a small number of mailboxes multiplied by an aggressive one.

Working out what you need

Start from the outcome and work backwards. If you want twenty conversations a month, and your reply rate is 5% with a third of replies being positive, you need roughly 1,200 sends a month to get there. That is about 60 a working day, which is twenty Microsoft mailboxes at three a day.

1. Decide how many conversations a month you want.
2. Divide by your positive reply rate to get the sends required.
3. Divide by working days to get daily volume.
4. Divide by your per-mailbox rate - three on Microsoft 365 - to get the mailbox count.
5. Add headroom, because some mailboxes will always be in warmup or rotation.

One of our orders is 100 inboxes on one domain at three sends per inbox per day, which is 300 a day or roughly 6,000 a month. For most teams that is comfortably more than they need, which is the correct position to be in - running infrastructure at its ceiling leaves nothing for a bad week. Once you have a mailbox count, what cold email inboxes are and how to buy them covers what to look for in a provider and what the going rate per inbox is.

What the number looks like when it goes wrong

Sending too much per mailbox does not usually announce itself as a bounce. The mail is accepted, the sequencer reports it as delivered, and the reply rate quietly falls to nothing. Delivered and inboxed are different things, and only one of them is visible from the sending side.

The sequence is fairly consistent. Placement moves from inbox to Promotions or Other, then to spam, then the receiving side starts deferring - a 4.x.x temporary failure that the sequencer retries and eventually records as sent. By the time hard bounces appear the domain has usually been in trouble for a week.

• Reply rate falling while delivery stays near 100% is the first useful signal, because it is the one that moves before anything is rejected.
• A gap between opens and replies that widens over days rather than jumping suggests placement, not copy.
• Deferrals concentrated on one receiving domain - all the Outlook addresses, say - point at reputation with that provider rather than at the list.
• Bounces above roughly 5% mean list quality is the problem, not volume.

Seed testing is the only way to see placement directly. Put a handful of addresses you control across Outlook, Gmail and one corporate tenant into the send, and check where the mail lands rather than what the dashboard claims.

Why more mailboxes beats more per mailbox

Two setups sending 300 a day look identical in a report and are not remotely the same to a receiver. One mailbox sending 300 is a machine. A hundred mailboxes sending three each is a hundred people writing a few emails, which is what normal business mail actually looks like.

Receivers score reputation per sending address as well as per domain, so concentrating volume also concentrates the consequence. When a single mailbox is throttled it takes the whole days output with it; when one of a hundred is throttled the campaign loses one percent and keeps running.

The cost argument usually runs the other way from what people expect. At $0.30 per inbox per month, adding mailboxes to spread the same volume is cheaper than recovering a burned domain, which costs the domain, the warmup weeks, and whatever the pipeline would have produced in the meantime.

Warmup

A new mailbox with no sending history has no reputation, and no reputation is treated with suspicion. Ramp over two to four weeks rather than starting at full volume on day one.

• Week one: 1-2 sends a day, ideally to addresses that will reply.
• Week two: 20-30 a day, beginning real prospect sends.
• Weeks three and four: hold at target volume and watch placement.
• Ongoing: keep a warmup pool running alongside live sending so engagement continues.

Skipping warmup is the most common cause of a batch of new mailboxes performing badly. The mailboxes are not faulty; they are strangers. If placement drops after warmup has finished, work through the deliverability guide and confirm the domain is not listed with the blacklist checker.

The signals that mean stop

Reduce volume immediately if reply rate falls sharply without a change in copy or list, if test sends to seed mailboxes start landing in Junk, if bounce rate rises above 3%, or if any complaint rate above 0.3% appears. Recovering a damaged domain is slow. Not damaging it is much cheaper.