How many cold emails can you send per day?

A provider's limit is not a recommendation

This distinction matters more than any single number on this page, and almost every guide blurs it. What Google and Microsoft technically allow is an anti-abuse ceiling: the point at which the platform stops accepting your mail. It is not advice about cold outreach, and it says nothing about whether a receiving provider will place your message in an inbox.

Sources: Google Workspace Admin Help, "Gmail sending limits in Google Workspace"; Microsoft Learn, "Exchange Online limits" service description. Both are the providers' own current documentation.

Read the gap. Google will accept 3,000 external recipients a day from one account. We send ten or twelve. That is not caution for its own sake: a mailbox sending at a platform's anti-abuse ceiling looks exactly like the abuse the ceiling exists to catch, and reputation is assessed by the receiving provider, which never agreed to that number.

Why the limit is per mailbox, not per campaign

Receiving providers build reputation against the sending mailbox and its domain, not against your campaign. Sending 500 emails a day from one mailbox looks like bulk mail no matter how good the copy is; sending 500 a day across a fleet of warmed mailboxes looks like a group of people doing their jobs. That is the whole reason agencies run mailbox fleets, and it is why the honest answer to "how many can I send" is a multiplication rather than a single figure.

Why the Microsoft number is so much lower

It is not a rounding difference, and it is not that Microsoft mailboxes are weaker. The binding constraint sits at a different level. On Google, sending headroom is largely a per-mailbox property. On Microsoft 365, Exchange Online also applies a tenant-wide external-recipient control in a 24-hour sliding window, which Microsoft says depends on the number of licences the tenant has. Spreading sends across more mailboxes or more domains inside the same tenant does not by itself lift that control.

So the low per-mailbox number is a reputation decision rather than a capacity trick: many mailboxes each sending very little spreads sending reputation across many senders instead of concentrating it in a few, within whatever tenant capacity Microsoft's rules allow. The full explanation, including what Microsoft itself says about bulk mail on Exchange Online, is on cold email infrastructure.

The capacity maths, worked

Daily capacity is your mailbox count multiplied by the per-mailbox number. Two worked examples on the Google pool at 10 and at 12:

• 5 mailboxes at 10 a day = 50 sends a day. A 14,000-lead list takes about 280 days for a single pass. Most people discover this after building the list.
• 60 mailboxes at 12 a day = 720 sends a day. The same list takes about 20 days.

The practical consequence: if your list is larger than roughly 50 times your mailbox count, your list is not the constraint, your infrastructure is. Work out that ratio before buying more data, because more leads against fixed capacity changes nothing.

Working backwards is the useful direction. Decide how many conversations a month you want, divide by a realistic reply rate to get the sends you need - our measured campaign benchmarks put the median campaign at 2.12% of contacted leads - then divide by 10 to 12 to get the Google-pool mailbox count, and order those mailboxes two to three weeks ahead of when you need them.

The capacity planner: from reply target to mailbox count

The same arithmetic as a working tool. Enter the replies you want in a month, pick which measured reply rate to plan on, and it returns the contacted leads, the sends, and the mailbox and domain counts each provider pool needs. Every constant in it is published on this site: 10 to 12 sends per Google Workspace mailbox per day, 1 to 2 per Microsoft 365 mailbox, 1.77 sends per contacted lead across the measured book, and the reply-rate quartiles from the campaign benchmarks. It runs in your browser and sends nothing anywhere.

Volume is only one of eight things that decide capacity

Per-mailbox volume is the number everyone asks about and roughly the least interesting one. A programme that is sized properly accounts for all of these, and each can become the binding constraint before volume does:

• Number of mailboxes. The actual growth lever. Capacity scales here, nowhere else.
• Number of domains. Reputation and blocklisting are domain-level events, so domain count caps how much you lose to any single one.
• Provider mix. Google, Microsoft or both. Different limits, different architectures, and independent failure.
• Recipient-provider mix. Who you are sending to matters as much as who you send from. A list weighted to corporate gateways behaves nothing like one weighted to Gmail.
• Infrastructure health. A mailbox that is deteriorating has negative capacity: it spends domain reputation for nothing.
• Sender reputation. Earned per mailbox and per domain, slowly, and lost quickly.
• IP and provider pools. Whether your sending sits in one reputation environment or several independent ones.
• Redundancy. What still sends when one pool has to be paused.

All eight are the subject of cold email infrastructure. This page is about the first number only.

Warmup: the two to two and a half weeks nobody plans for

A brand-new mailbox cannot go straight to its steady-state number. It needs two to two and a half weeks of gradual warmup, sending small volumes of genuinely-engaged mail, before it carries campaign traffic. This is the single most commonly skipped step, and skipping it is expensive: the first impression a provider forms of a new domain is the one that persists. Two implications worth planning around: order mailboxes two to three weeks before the campaign that needs them, and never attach a still-warming mailbox to a live campaign to "get more volume" today. The full ramp is on email warmup.

The thresholds that actually get you throttled

• 0.3 percent spam complaints. Google holds bulk senders to this, and recommends staying below 0.10 percent. Above the line, placement degrades quickly and it is hard to recover.
• 5,000 messages a day to Gmail, and separately to Outlook.com, is where bulk-sender rules apply as a hard gate: SPF, DKIM and DMARC all required, plus one-click unsubscribe on marketing mail at Gmail. Outlook.com began rejecting non-compliant high-volume mail outright in May 2025.
• 2 percent bounce rate. Our own working ceiling. Sustained rates above 5 percent invite throttling and blocks regardless of volume.
• 10 DNS lookups in your SPF record. Exceed it and the record permanently errors, which means every message you send counts as unauthenticated no matter how careful your volume is. Our deliverability checker reports the live count.

Volume is not the only reason mail fails

Worth knowing before you blame your send rate: a meaningful share of failures are recipients refusing you on policy rather than anything you did. Corporate gateways such as Mimecast and Proofpoint return a hard rejection to senders they do not recognise, and in most reporting that lands in the same bucket as an invalid address. Lowering your volume does not fix it. Reading the actual rejection string does, because it tells you whether to change your list or your sending setup.

How to scale without losing placement

1. Decide the daily number of conversations you actually want, then divide by a realistic reply rate to get the sends you need.
2. Divide that by 10 to 12 for Google Workspace mailboxes, or by 1 to 2 for Microsoft 365 mailboxes, to get the mailbox count each pool has to carry.
3. Spread those mailboxes across enough domains, and across more than one provider, so no single failure takes the whole programme down.
4. Authenticate every sending domain fully before its first send, then warm each mailbox for the full period.
5. Verify the list live immediately before sending, not when it was purchased.
6. Hold per-mailbox volume flat forever. Growth comes from mailbox count and pool count.

Common questions

How many cold emails per day per mailbox in 2026?

On ReplyLead's current infrastructure, around 10 to 12 per Google Workspace mailbox per day and around 1 to 2 per Microsoft 365 mailbox per day. Those are operating choices well below what either provider technically permits, and materially more conservative than most of the industry still publishes. Total volume comes from mailbox count, not from the per-mailbox number.

Can I send more if the mailbox is old and warm?

Somewhat, and cautiously. Established mailboxes on older domains tolerate more, but the return on pushing is small compared with the downside if you misjudge it. Adding a mailbox is cheap and reversible; rebuilding a deteriorated domain's reputation is neither. We hold per-mailbox volume flat and grow the fleet instead.

How many mailboxes per domain?

It depends on the pool. On Google Workspace, two or three mailboxes per domain is a defensible pattern: it keeps any one domain's exposure small. On Microsoft 365, where per-mailbox volume is roughly 1 to 2, a domain may carry around 49 to 50 mailboxes and still sit at a modest total. The architecture, not a universal number, decides this - see cold email infrastructure.

Does a bigger list mean I should send faster?

No. A larger list means a longer campaign or more mailboxes, and those are the only two honest options. Sending faster than your reputation supports converts a good list into a deteriorated domain, and you lose both.

How do I know if my current volume is safe?

Watch bounce rate and reply rate together, on the sends made since your last change rather than on cumulative averages, which hide recent behaviour behind old data. If bounces sit under 2 percent and replies are stable as you add mailboxes, your ramp is working.