How to Warm Up a New Email Domain for Cold Outreach

Most operators start the warmup clock the day they connect the mailbox, and that one habit burns more new domains than any warmup setting ever has. We run outbound for 50 plus B2B companies and have sent over 8 million hyper-personalized cold emails this year across hundreds of domains we warmed from zero. Below, the 2 clocks that have to run at the same time, the 30 day volume ramp that does not trip a throttle, and the 5 readiness signals to confirm before the first cold batch goes out.

How long does it take to warm up a new email domain?

The 14 day number came from a period when Gmail and Outlook were looser about new senders. That period ended. Google published bulk sender requirements in 2024 that made SPF, DKIM, and DMARC mandatory for anyone sending at volume, and Microsoft followed with equivalent rules for Outlook in 2025. Both changes raised the floor on what a brand new domain has to prove before its mail gets treated as legitimate, and the practical effect is that a domain warmed for 2 weeks now behaves like a domain warmed for 1 week used to.

Domain Warmup

The 30 day process of building positive sender reputation on a new email domain by sending a slowly increasing volume of messages that get opened, replied to, and marked as important. Most of that traffic comes from an automated warmup network. Mailbox providers read the engagement and gradually grant the domain higher inbox placement. A warmed domain can sustain cold sends. An unwarmed one cannot, at any volume.

Domain Age

How long a domain has existed since registration, independent of whether it has ever sent mail. Filters treat newly registered domains as higher risk because throwaway domains are the cheapest tool a spammer has. Age is not something a warmup tool can produce. It only accrues on the calendar, which is why domains should be bought well before the campaign that needs them.

Those 2 definitions are the whole article in miniature. Warmup builds engagement history. Age builds the absence of suspicion. They are different inputs into the same scoring decision, and running only one of them is why so many operators report a clean warmup dashboard and a cold batch that still went to spam. If you want the mechanism behind warmup itself rather than the new domain ramp, our full breakdown of how email warmup works covers the signals providers actually read and what warmup cannot fix.

Why warmup time and domain age are 2 different clocks

Filters score an inbound message on dozens of inputs. 2 of them are the ones a new sender controls: how long the sending domain has existed, and what its sending history looks like. Warmup tools only move the second one. The first one moves on its own, at the speed of the calendar, and it costs nothing to run except forethought.

The operational takeaway is simple and almost free. Buy the domains 30 to 60 days before the campaign needs them. Point the DNS, stand up the mailboxes, run the warmup, and let the age clock tick underneath all of it. By the time the warmup finishes, the domain is 30 days old instead of 30 days old and 30 days warmed being the same number. Teams that plan a quarter ahead register the next quarter's domains during the current one, and the ramp stops being a launch blocker entirely.

If you already burned a domain by launching on day 3, the ramp described here will not save it. A domain carrying a suspicious flag has to be diagnosed and, most of the time, replaced. Our recovery playbook covers what is actually salvageable and what is not, and the domain reputation primer explains how the flag gets applied in the first place.

What has to be true before day 1: DNS, mailboxes, and the sending account

Warmup does not start with the warmup tool. It starts with DNS. A tool generating engagement on a domain with no SPF, no DKIM, and no DMARC is producing signals that mailbox providers cannot tie to a verified sender, so the signals get discounted. 30 days of that produces a domain that still gets flagged on the first cold send, and none of the warmup history transfers once the records are fixed. The ramp restarts.

Every record below has to be published and verified before the warmup tool sends its first message. Not the same week. Before.

Google's email sender guidelines, Yahoo's sender best practices, and Outlook's high volume sender requirements all now converge on the same 3 records. There is no provider left that treats authentication as optional. If any of the syntax above is unfamiliar, the SPF, DKIM, and DMARC explainer walks through each record line by line, and the DNS records guide covers the ordering mistakes that break authentication without throwing an error anywhere visible.

One setup decision sits above all of these: never warm the primary company domain for cold outreach. Outbound runs on a secondary domain that is a close variant of the brand, with incoming mail forwarded to the real inbox. If a list turns out dirty or a campaign draws complaints, the damage lands on a domain you can retire. The secondary domain writeup covers naming patterns that do not read as throwaway, and the full domain setup guide covers registrar choice, forwarding, and mailbox naming.

The 30 day ramp: what to send, day by day

The volume curve is not a straight line. The first 2 weeks are deliberately tiny, because the point is engagement ratio rather than volume. Weeks 3 and 4 ramp. Cold traffic does not appear until day 22, and even then only in small batches beside the warmup traffic. Each phase depends on the engagement data the previous one produced, so skipping a phase makes the volume jump look exactly like the pattern filters are built to catch.

The line operators ignore most is the last column of the bottom 3 rows: warmup never turns off. A domain that flips to 100 percent cold traffic with no positive engagement underneath it drifts back toward suspicious over about 4 to 6 weeks, because cold lists reply at a few percent and warmup networks reply at 30 or more. Keeping 10 to 15 warmup messages per mailbox per day permanently is what holds the engagement ratio in a healthy band once the cold volume arrives.

Volume is measured per mailbox, never per domain. That distinction decides the entire infrastructure plan, and it is the one most send volume advice gets wrong. The sending volume breakdown goes deeper on where the per mailbox ceiling actually comes from, and mailbox rotation covers how sending platforms spread a day's volume across the set without spiking any one sender.

How many domains and mailboxes does the ramp actually need?

Work backward from the monthly send goal. Divide by 22 sending days to get the daily number. Divide that by 30 sends per mailbox, which is the conservative first 90 day cap. Then group mailboxes 3 to a domain, because 3 is where per mailbox cost and per domain risk balance out. Round up at every step.

Build the full set before the ramp starts. Every domain added later runs its own 30 day clock from its own registration date, so a campaign that adds 4 domains in month 3 is really 3 campaigns at 3 different maturity levels, with 3 different safe volumes. That is a reporting mess and an unnecessary one. The multi domain strategy guide covers how to group domains so one bad list segment cannot take the whole set down, and the infrastructure setup guide covers the full stack from registrar to sending platform.

The volume plan is downstream of the list, not the other way around. A tight list of 5,000 real buyers beats a loose list of 30,000, and it needs a fraction of the infrastructure. Defining the ICP as a deliverability constraint rather than a marketing document is the single highest leverage thing an operator can do before buying a single domain.

How do you know the new domain is ready for cold traffic?

Day 30 on the calendar is the floor, not the answer. The sharper test is signal based, and 5 of them have to be true at the same moment. Hit day 30 while missing any one, and the domain is not ready, no matter what the warmup dashboard says.

1. Inbox placement above 95 percent. Run a seed test on day 18 to 20 and again on day 28. A seed list delivers a test message to 50 to 100 inboxes across the major providers and reports where each one landed. Below 95 percent means another 3 to 5 days of warmup before anything cold goes out. How to run and read a placement test covers the setup.
2. Zero authentication failures across 7 days. Pull the last week of sends and confirm SPF, DKIM, and DMARC pass on 100 percent. A 2 percent failure rate at 20 sends a day becomes a visible pattern at 600, and by then the reputation damage is already priced in.
3. Warmup reply rate above 30 percent. The network should be returning roughly 1 reply per 3 sends. Under 20 percent means the network is too small or the engagement is too templated, and mailbox providers have gotten good at fingerprinting both. Switch tools before sending cold, not after.
4. Zero blacklist appearances. Check against Spamhaus and run the domain through MXToolbox's blacklist check. A freshly warmed domain should be on zero lists. Any listing during warmup means pausing, finding the cause, fixing it, and adding 7 days. Ongoing blacklist monitoring should be running before launch, not after the first incident.
5. Postmaster Tools reputation at Medium or better. Unknown is the default for a domain with too little Gmail volume, and it is acceptable during the first 2 weeks. By day 28 it should have moved. Low is a stop sign. Mailgun's guide to reading the Postmaster dashboards explains what each band actually reflects.

Those 5 are the readiness gate. Treat them as a checklist a human signs off on, not as dashboards someone glances at. Continuous deliverability monitoring is the version of this that runs forever once the campaign is live.

What burns a new domain in the first 60 days?

Failed warmups almost never fail because of a setting inside the warmup tool. They fail because of operational choices made around the warmup that compound quietly for weeks. The same patterns show up across the agencies and in house teams that come to us after burning a set of domains in the first 2 months.

The second and third rows account for most of what we see. Teams tell us they warmed for 30 days and still cannot reach the inbox, and the first thing we check is DNS and registration date. Usually DMARC is missing, DKIM was published but never switched on, or the SPF record blew past the 10 lookup limit and has been returning permerror the entire time. The warmup produced no reputation because the engagement was never attributable. Fixing the records and warming again from day 1 is the only path, and the previous 30 days do not carry over. Bounce rate causes and fixes covers the list side of the same problem.

The first cold batch, and how fast you can scale after it

The day 22 batch is the most fragile send the domain will ever make. The reputation exists but is thin, and the warmup data is positive but untested against cold traffic, which behaves nothing like a warmup network. The send has to be small, verified, and aimed at the warmest segment of the list available.

Send 10 to 20 messages per mailbox to verified addresses only, at confirmed decision makers inside fit tier accounts, with copy that reads conversational rather than pushy. The goal of the first batch is not booked meetings. The goal is confirming that cold traffic does not move the reputation. Reply rate should land at 3 percent or better and bounce rate should stay under 2 percent. Numbers outside those bands point at the list, not the domain. Reply rate benchmarks by segment gives the comparison set, and the positive reply rate definition explains why the raw reply number alone is a weak read.

Run the first 3 batches small before scaling. By day 30 to 35, if replies are holding, bounces are low, and seed placement is still above 95 percent, move to the full 30 per mailbox cap and stay there through day 120. Litmus's deliverability research and Validity's 2026 benchmark report both track how much of long run placement is set by early sending behavior. Conservative volume in the first 90 days is the cheapest reputation insurance available, and nobody who skipped it has ever told us it was worth the time saved.

For anyone running the invite motion specifically, the ramp is the same but the copy load is lighter, which helps. Domains and warmup for podcast invites covers the deployment we run, and podcast invite deliverability covers why an invite clears filters that a sales pitch does not.

When to stop warming and start over

Not every domain warms cleanly. Some were burned by a previous owner before you bought them, and no amount of warmup fixes a domain that arrives with history. The tell is a domain that sits at Low reputation past day 21 with clean authentication, zero complaints, and a warmup network that is engaging normally. That combination means the problem predates you.

Check the domain's history before buying. If the name previously hosted a site, it may carry spam associations that follow the root domain. Registering a fresh name is cheaper than a month of diagnosing one that will never clear, and a month is what a stubborn domain costs. The rule we use: a domain that has not reached Medium reputation by day 28 with everything else clean gets replaced rather than extended.

The same call applies mid campaign. A domain whose placement degrades for 2 consecutive weeks despite a clean list and stable volume is telling you something the dashboards cannot name. Rotate it out, keep the mailbox set balanced, and let the replacement run its own ramp. The spam folder diagnosis tree covers how to tell a domain problem from a list problem from a copy problem before you retire anything, and the deliverability overview ties the full measurement stack together.

The practitioner take: the cheapest month in outbound is the one before you send

Warmup is not a ritual. It is a reputation building exercise scored on a fixed set of inputs: authenticated sends, positive engagement, low complaint rates, low bounce rates, consistent volume patterns, and how long the domain has existed. 30 days is the floor for a brand new name. Some domains take 45. A few never clear and should be replaced instead of nursed.

The shift worth naming for 2026 is that mailbox providers got noticeably better at recognizing warmup traffic for what it is. Engagement that does not resemble real correspondence gets discounted, so identical templates bouncing between bot inboxes are worth less every quarter. Network diversity and engagement realism now matter more than which logo is on the tool. Pick the one with the largest, most varied network across Google, Microsoft, and consumer providers, and read its placement reporting rather than its summary score.

The sequence, start to finish: register a secondary domain 30 to 60 days before you need it, publish MX and SPF and DKIM and DMARC and verify every one, stand up 3 mailboxes, verify the domain in Postmaster Tools, run the 30 day ramp without shortcutting, confirm the 5 readiness signals, send the first cold batch on day 22 to a verified segment, hold at 30 per mailbox through day 120, and keep warmup running forever. Teams that follow that order rarely have a deliverability problem past month 1. Teams that skip steps usually reach out 90 days later asking how to recover a set of burned domains, and the honest answer is that the month they saved at the start costs 3 on the back end.