How long to warm up an email domain before a campaign?

Plan for three to six weeks of warmup before a campaign that needs meaningful daily volume. M3AAWG, the cross-industry anti-abuse body whose members include the mailbox providers doing the filtering, says to "consider 6 weeks of warm-up as an average," and the nine published ESP schedules below run from 21 to 45+ days. The single biggest variable is your target volume — a domain that will send 100 emails a day can be ready in days, while ramps to millions per day take 40 or more. And one number is non-negotiable: a domain registered today should send nothing for at least 24 hours, because Spamhaus lists every newly registered domain for its first day of existence.

What nine published warmup schedules actually say

Nine senders publish concrete warmup guidance. Here is what each one says, with the fixed automated ramps listed separately where a vendor runs one:

Source	Published schedule	Time to full volume	Model
SendGrid (generic table)	50/day on day 1 → 1,000,000/day on day 21	21 days	Fixed daily-volume table
SendGrid (automated warmup)	Starts at 20 emails/hour on day 0	41 days, fixed	Hourly caps on a clock
Amazon SES (per-provider guide)	150 per provider on day 1 → 21M/day total on day 40	40 days	Separate ramp per mailbox provider
Amazon SES (automatic)	Time-based ramp to a 50,000/day ceiling	45 days, fixed	Calendar-driven, ignores actual volume
Litmus	1% of your engaged list, doubling weekly to 100%	4–8 weeks, "8 standard"	Percentage of list
Klaviyo	30-day-engaged segment daily, then widening engagement windows	No fixed end date	Engagement-gated phases
Braze	Three tracks, all starting at 50/day	21–30+ days	Conservative / Moderate / Aggressive
Postmark	50–100 per major provider on days 1–2, then +20–50%/day	3–6 weeks	Growth-rate rule
Resend	150–400/day → ~2,000/day for a new domain	7-day baseline	Compact table + metric gates
Mailgun	15 stages, from 1,000/day to 250,000/day	15+ stages; skipped days reset progress	Staged daily + hourly caps
MailReach (cold outreach)	5–10/day per mailbox → 75–100/day cap	~3 weeks	Per-mailbox, plain-text sends

Three things converge across every table:

• Day one is tiny everywhere. 50 to 400 emails for bulk senders; 5–10 per mailbox for cold outreach. Nobody publishes a schedule that starts in the thousands for a new domain.
• Time-to-target clusters at three to six weeks. The two fully automated clocks — SendGrid's 41 days and SES's 45 — both land at roughly six weeks. SendGrid's customer guidance says 30 days is typical, up to 60, with 1–2 weeks possible only for small targets. Resend's own launch of new.email took six weeks.
• The growth rate decays as volume grows. The SendGrid table doubles daily only while volumes are trivially small (up to 16,000/day on day 8), then slows to roughly +21–33% per day. The AWS table has the same shape. "Double your volume every day" is only true at the start of every schedule and wrong at the end of all of them.

Note also what the schedules disagree on: SendGrid and AWS are calendar tables, while Klaviyo and Braze reject fixed calendars entirely and gate each step on engagement — Klaviyo tells senders to roll back a step if open rates drop below 20%, and Braze's automation pauses the ramp when delivered rate falls under 90%, bounces exceed 5%, or complaints exceed 0.04%.

Six weeks: the closest thing to an industry answer

The only non-vendor number in this topic comes from M3AAWG, which frames reputation-building as two distinct phases: a warm-up phase that "brings the sending domain from unknown to noticed," then a ramp-up phase toward planned long-term volume — and suggests six weeks as the average for the whole exercise.

Independent practitioner experience runs slightly longer. Al Iverson, who writes the Spam Resource blog and has run his own mailbox provider, reported in February 2025: "it might not be only a four week or thirty day endeavor. I've seen it sometimes take five to eight weeks." He also notes that some spam-folder placement during the first week or two is normal, not a sign the warmup has failed.

So if you are planning a campaign launch backwards from a date: six weeks of warmup is the defensible default, and anything under three weeks is a bet that only pays off at low volume.

Day one is special: your domain is already on a blocklist

Warmup duration questions usually assume the clock starts when you send the first email. For a newly registered domain, there is a gate before that.

Spamhaus operates the Zero Reputation Domain (ZRD) list, which lists every newly registered domain for its first 24 hours — automatically, with no exceptions for legitimate businesses and no delisting process: "Unfortunately, we do not accept requests for removal from this blocklist." The listing simply expires after 24 hours. The return code even encodes the domain's age in hours, so receivers can reject a 2-hour-old domain outright while merely scoring a 20-hour-old one. As Iverson puts it, any provider subscribing to the full Spamhaus suite "is going to reject mail from your domain, if your domain is less than 24 hours old" — including your test messages.

The scrutiny does not end at hour 25. Spamhaus states that "for the first 30 days after initial registration, most companies specializing in threat intelligence will flag a newly registered domain", and Iverson independently recommends aging a domain at least 30 days before commercial sending. The underlying logic is stated in the Spamhaus DBL FAQ: "unknown reputations begin as 'poor' by default."

Practical consequence for campaign planning: register the domain first, well before you need it. Set up SPF, DKIM, DMARC, and a real website immediately; wait at least a full day before any send; begin low-volume warmup inside the 30-day aging window so the domain accumulates history while the threat-intelligence flags decay. If your domain is already live and mail is bouncing, see what to do when a new domain is blocklisted — and you can check any domain's age and blocklist status with our domain checker.

What stretches or shrinks the timeline

• Target volume — the dominant factor. Reaching a steady 100/day takes days. Reaching 50,000/day takes SES 45 days on its automatic ramp. Reaching 21M/day takes the full 40-day AWS table. Work backwards from the volume your campaign actually needs, not from the biggest table you found.
• Provider mix. Microsoft is widely reported as the strictest and slowest to open up to new senders; Iverson says spam-folder placement at Microsoft in the first week is not weird at all. Validity's 2026 benchmark (a deliverability vendor, but large-sample) measured 2025 inbox placement at 89.8% for Gmail versus 77.4% for Microsoft. A Microsoft-heavy list warms slower than a Gmail-heavy one, and AWS advises ramping per provider, not just in aggregate.
• Engagement. Warmup is an engagement game, not just a volume game. Sending to people who open and reply lets you progress on Klaviyo- and Braze-style gates; sending to a cold or stale list stalls the ramp or reverses it.
• Consistency. Mailgun's staged plan explicitly warns that skipping days resets progress, and Google's sender guidelines say to send at a consistent rate and avoid bursts. A 21-day schedule executed with gaps is not a 21-day schedule.
• New domain vs. subdomain. A new subdomain of an established domain still needs its own ramp — Litmus notes it is treated as an unknown sender — but it skips the 24-hour ZRD listing and the 30-day new-registration scrutiny that a fresh domain faces.
• Dedicated IP. If you are also warming a dedicated IP, the IP is usually the tighter constraint and the schedules above apply to it in full. On a shared ESP IP, only the domain needs warming.

The mechanics of executing the ramp — what to send, to whom, and in what order — are covered in how to warm up an email domain. WarmEnvelopes automates the sending side of this using the same decaying-growth logic the published tables share; the ramp-up schedule page shows exactly what it sends on which day.

What rushing actually costs

Google's sender guidelines are explicit: "avoid introducing sudden volume spikes if you do not have a history of sending large volumes." A domain whose entire observed history is zero that suddenly emits a campaign is, statistically, indistinguishable from a compromised or throwaway domain — that is exactly the signature filters are trained on. Google's bulk-sender FAQ adds that enforcement for new domains runs on an accelerated timetable: new domains get less benefit of the doubt, not more.

The failure modes are concrete:

• Spam-foldering that outlasts the mistake. Early spam placement is normal; a volume spike converts it from a temporary state into a formed negative opinion.
• Ramp resets. Metric breaches roll you backwards — Braze's automation holds the ramp on bounce or complaint spikes, Klaviyo sends you back a phase, and a bad enough week means restarting weeks of work.
• A permanent classification. Sending 5,000+ messages to personal Gmail accounts in a rolling 24 hours even once makes you a bulk sender permanently, with all of Gmail's bulk-sender requirements attached from then on.
• Decay if you stop. Warmup is not a bank deposit. SendGrid notes that most reputation systems only store data for about 30 days, so a domain warmed in March and left idle until June effectively starts over.

The honest caveat: no schedule is validated

Every number in the comparison table above is vendor guidance. As of August 2026, no mailbox provider publishes a warmup schedule of its own, and no independent study validates any specific curve — 21 days versus 40 days versus six weeks has never been tested head-to-head in public. SendGrid says of its own famous table that "using it does not guarantee success." Laura Atkins of Word to the Wise, one of the field's most cited independent practitioners, is blunter still: "There's no right way to warm up an IP nor is there a specific formula that everyone should follow… Each IP warms differently."

What the schedules are good for is the shape, which everything published agrees on: start near zero, grow at a rate that shrinks as your volume grows, hold or roll back when bounces, complaints, or engagement move the wrong way. The calendar answer — three to six weeks, six if you want the defensible default — is a planning estimate. The real gate for launching your campaign is your metrics: bounce rate under a few percent, complaint rate near zero, and engagement holding steady at each volume tier. When those survive several consecutive increases at something near campaign volume, you are as warmed up as anyone can honestly claim to be.