Email Warm-Up: The 21 Days You Cannot Skip

Every founder who sets up outbound for the first time hits the same wall in week one. The sending domain is bought, the mailboxes are made, SPF, DKIM and DMARC are in place, the list is ready. Everything says go. And the right move is to send almost nothing for three weeks.

That is email warm-up, and it is the step most likely to be skipped, because it looks like waiting. It is not waiting. It is the only way a new domain earns the right to be read. Rush it and you do not save three weeks; you lose the domain, and start the three weeks again on a new one.

What is email warm-up?

Email warm-up is the practice of sending a small, steadily growing volume of mail from a new domain or mailbox, so that inbox providers build a positive history for it before you ask them to accept real campaign volume.

The reason it exists is simple. Gmail and Microsoft decide where a mail lands by looking at the sender's reputation, and a new domain has none. To a spam filter, no history is not neutral. It looks like a spammer's fresh domain, which is exactly what most fresh domains that send bulk mail turn out to be. A domain that goes from zero to 200 mails a day on day one matches the pattern perfectly, and the filter treats it accordingly.

Warm-up breaks the pattern. A domain that sends five mails on Monday, seven on Tuesday, and grows from there, to people who open and reply, looks like a small business starting to use email. That is the history you want on file before the first cold campaign goes out.

We covered why outreach must run on a separate domain in our email deliverability guide. This article is about what you do with that domain once you own it.

Domain warm-up and mailbox warm-up are not the same thing

People use "email warm-up" for two different jobs, and you need both.

Domain warm-up builds the reputation of the domain itself, the part after the @. Every mailbox on that domain shares it. A brand-new domain has no reputation at all, and it takes weeks of sending to establish one.

Mailbox warm-up builds the reputation of one address. Add a new mailbox to a domain that is already warm, and it still needs a shorter warm-up of its own, usually a fortnight, because the receiving side tracks the address as well as the domain.

The practical consequence: when you set up outbound, warm the domain and the first mailboxes together over 21 days. When you add a mailbox later, give it two weeks at low volume before it carries campaign mail. Never assume a new address inherits a warm domain's trust in full.

How to warm up an email address: the 21-day schedule

There is no official schedule from any inbox provider, but the practice that works follows one shape: start tiny, grow slowly, never jump. Here is the version we use.

Days 1 to 7: a trickle. Send five to ten mails a day per mailbox, to real people who will open and reply. Colleagues, friendly contacts, your own other mailboxes. Write normal messages and reply to the replies. The aim is to show the filter a mailbox having conversations, not broadcasting.

Days 8 to 14: a steady climb. Add two to five mails a day, so the mailbox is sending twenty to thirty a day by the end of the second week. Keep the mix conversational. If any mail lands in spam in your own test inboxes, move it to the inbox and reply to it; that correction is a strong positive signal.

Days 15 to 21: approaching campaign volume. Grow to thirty to forty a day. In the final days, send a handful of real, hand-written messages to warm prospects, people who know you, to check that mail from the domain reaches real inboxes outside your own network.

Day 22 onwards: campaigns, still climbing gently. Begin cold sending at a low share of the mailbox's capacity and grow over the following week. Keep the warm-up conversations running in the background for the first month of campaigns, because replies from strangers arrive slowly and the filter still needs positive engagement to weigh against the inevitable ignores.

Three rules sit underneath the schedule.

Never jump. The filter watches the rate of change, not just the volume. Doubling from 20 to 40 overnight can trip a limit that 20 to 25 to 30 never would.

Engagement matters more than count. Ten mails that get replies do more for the domain than fifty that get deleted. Warm-up mail that nobody reads is nearly worthless.

Watch bounces from day one. A bounce on a warm-up mail means you typed an address wrong or the mailbox is dead. Even a few bounces early in a domain's life leave a mark, so send warm-up mail only to addresses you know are live.

What email warm-up tools actually do

Search for warm-up and the results are almost all tools: Warmup Inbox, Instantly, Snov.io, GMass, and a dozen others, many with a free tier. It is worth being clear about what they do, because the marketing makes it sound like a shortcut.

A warm-up tool connects your mailbox to a network of other users' mailboxes. Every day it sends a small number of automated mails from your address to theirs, and theirs to yours. The receiving mailboxes are set to open the mail, reply to it, and if it lands in spam, pull it out and mark it as not spam. Over a few weeks this manufactures the engagement history the schedule above builds by hand.

The upside is real. It removes the manual labour of the 21 days, which matters when you are warming ten mailboxes rather than one, and the spam-rescue behaviour is something you cannot easily do by hand at scale.

The limits are real too. The engagement is synthetic, and inbox providers are not naive. Their filters are built to spot patterns, and a network of mailboxes that all open and reply to each other on a timer is a pattern. Providers have grown steadily stricter about mail that looks automated in this way, so a tool is a supplement to genuine conversations, not a replacement for them. A mailbox warmed only by a network, and then switched to a hundred cold mails a day, often lands in spam within a fortnight anyway.

Our approach is both at once: a warm-up tool running in the background from day one, and real human conversations layered on top through the three weeks. The tool builds the volume history. The humans build the reputation the tool cannot fake.

How many cold emails per day can a warm mailbox send?

The number most experienced senders settle on is thirty to fifty cold emails a day per mailbox, on top of ordinary replies. Above that, complaint rates and per-mailbox sending limits start to bite, whatever the technical cap of the platform says.

This is why outbound at any real volume runs on several mailboxes across two or three sending domains, not one address pushed to its limit. Two hundred cold mails a day is five or six warm mailboxes, each doing a modest, sustainable share. A mailbox that goes down takes a fraction of the campaign with it, and a domain that has a bad week can be rested while the others carry on. We explained why that domain must never be your company's own in our email deliverability guide.

It also means warm-up is a rolling job, not a one-off. New domains and mailboxes are always coming online and being warmed while the campaign runs on the ones already warm. Anyone selling you outbound that starts full campaigns inside a week is either sending on their own already-warm infrastructure, which is fine and is how our own service works, or is about to burn a domain, which is not.

What warm-up does to your cold email response rate

Founders judge cold email by its reply rate, and blame the copy when the rate is low. Often the copy never got a hearing. A cold email response rate is capped by deliverability first: a message that lands in spam for half the list has its reply rate halved before a word of it is read.

So when a campaign underperforms, check the sending before the writing. If open rates on a new domain are far below what the same message achieved elsewhere, the domain is not warm, or it was warmed too fast. Rest it, drop the volume back, and warm again. Only when mail is reliably landing does the conversation about subject lines and message mean anything.

Warm-up also shapes the list you send to first. The ideal first campaign after warm-up is a small one, to the people most likely to reply: accounts showing a live buying signal, decision-makers with a specific reason to hear from you now. Early replies from strangers are the best reputation a domain can earn, and they make every later campaign land better.

Why we do not shorten the 21 days

Every client asks whether the three weeks can be compressed. The honest answer is that you can warm faster, and it works right up until it does not, and the failure is expensive. A domain warmed in a week and burned in the second week costs you the domain, the mailboxes, whatever replies were in flight, and three fresh weeks of warm-up on a replacement. The slow path is the fast path.

It is also why our engagements are structured the way they are. The first three weeks of a campaign are domain and mailbox warm-up on infrastructure we own, running alongside list research, so the first real mail goes out in week four onto domains with a clean history. It is the part of the work that does not look like work, and it is the part that decides whether anything after it lands.

Frequently asked questions

Warm-up is the least glamorous three weeks in outbound, and the most decisive. Every domain we send from goes through them, every time, before a single cold mail leaves it. That is why our engagements are built with the warm-up already inside them, on domains we own, so that when your first campaign goes out in week four, it lands.