How should we warm up a new sending domain?
Our answer
Authenticate the domain and mail only people who clicked or opened in the last 30 days. Then grow volume by about 1.5 times a step over 3 to 8 weeks, while bounces stay under 2% and complaints under 0.1%.
- Confidence
- Medium
- Strongest proof
- 10 official rules
- 3 practitioners, 1 opinion or claim
- Sources
- 8 sources
- 14 statements
- KOGCISMSL
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentStart with your most engaged subscribers. Klaviyo, Omnisend, Customer.io and Sendlane all begin a new domain with their latest engaged subscribers and widen the audience only while engagement holds.
Small steps are safer than doubling. Customer.io caps each step at 1.5 times because providers treat doubling as too aggressive, and Google warns that sudden doubling can bring rate limits. Klaviyo's and MailerSend's schedules double early sends.
Never use a warm-up service that fakes engagement. Spamhaus (email blocklist operator) says services that fake opens and clicks get a domain listed. Customer.io does not support them, and Omnisend says good senders don't need them.
Our judgment
Grow by no more than about 1.5 times a step, holding each step a few days. Customer.io's September 2026 guide caps steps there, Sendlane grows 1.4 times a send and Omnisend's most used plan grows 50% a day, while Klaviyo's and MailerSend's doubling schedules are the faster end. Expect about 3 to 8 weeks, with smaller lists and daily senders at the short end as Google and Customer.io describe. That range comes from Omnisend's plan of 20 or more days, Sendlane's month for metrics to settle, Klaviyo widening its engagement window in 2-week steps to 90 days after 4 weeks, and Customer.io's table to day 60. Hold or step back when opens fall below the 20% floor Klaviyo and Omnisend use, bounces pass 2% (MailerSend, Customer.io) or spam complaints reach 0.1% (Omnisend, MailerSend, Customer.io). Read opens loosely, because Apple privacy opens inflate them, as Klaviyo says.
Our judgment, built from Customer.io, Sendlane, Omnisend, Google, Klaviyo and MailerSend: our reading of these sources, not a finding from any one of them.
What the sources say
14 statements from 8 sourcesKlaviyo says new marketing domains should warm up with under 10,000 recipients first, starting with 30-day engaged contacts and widening only while open rates stay above 20%. Risky winback flows come last.
“For your first campaign, we recommend sending to fewer than 10,000 recipients.”
Klaviyo's warm-up schedule for senders without engagement data starts at 2,000 emails a day and doubles early on, and says to pause increases if engagement falls or complaints rise.
“If engagement declines or negative engagement rises, pause the increase and address the cause.”
Sendlane says to warm a domain by reacting to metrics: first send to 1,500 most engaged subscribers, grow 40% when opens reach 20% and clicks 5%, skip anyone unengaged 90+ days, and add automations later.
“To correctly warm your domain, you must react to each send's metrics.”
Gmail's sender guidelines say raise sending volume slowly, starting with engaged users, avoid sudden spikes such as doubling volume, and cut back if messages bounce or are deferred.
“For example, immediately doubling previously sent volumes suddenly could result in rate limiting or reputation drops.”
Since November 2025 Gmail has been rejecting more mail that fails its sender rules; bulk senders qualify for mitigation once spam rates stay below 0.3% for 7 straight days and all rules are met.
“Bulk senders will be eligible for mitigation when their spam rates remain below 0.3% for 7 consecutive days.”
Spamhaus says keep bounces below 5%, never use warm-up services that fake engagement, avoid sudden volume spikes, and start slowly to the most engaged on any domain idle for 90 days.
“Keep the overall bounce rate below 5%.”
Klaviyo says repeated sends to dormant subscribers hurt sender reputation, suggests suppressing never-engaged people who got 5+ emails in 180 days, and notes Gmail and Yahoo require a branded domain and DMARC.
“In general, any more than 1 email per inbox per day is risky.”
Customer.io says warm up a new sending domain in 2 to 3 day stages rising no more than 1.5x each, starting at 100 a day, since doubling is too aggressive, and pause at a spam rate of 0.10%.
“Recent trends from inbox providers indicate that doubling volume at each step is too aggressive.”
Litmus, which sells deliverability tools, says to clean lists before the holidays, suppress people who ignore a re-engagement campaign, and start the volume ramp in September.
“Start your ramp-up in September, not October”
MailerSend says to warm a new domain by sending 1,000 or fewer emails a day to recent engagers, then doubling to 6,000 to 8,000 by day six or seven, keeping spam complaints under 0.1%.
“Start by sending to people who engaged in the past 14 days, and gradually expand to those who engaged in the last 90.”
Omnisend says warm up new domains or platforms manually over 20+ days, sending first to engaged contacts, with open rate at 20% or more and spam at 0.1% or less before full sends. Lists under 500 can skip it.
“Automatic warm-up has been removed from Omnisend.”
Klaviyo says warming a dedicated IP, used only by managed accounts that qualify, typically takes 30 to 40 days. Clean lists first and treat clicks, not opens, as the main engagement signal.
“it is important to focus on clicks as your primary engagement events”
Klaviyo says new senders moving to it must warm up, and a branded domain needs warming if registered under 30 days ago or never used to send. Dedicated IPs warm separately.
“Any new sender must warm up their sending infrastructure when moving to a new email service provider, like Klaviyo.”
Omnisend says domain warmup means sending to opted-in subscribers, not simulated inbox tools, and gives an example ramp from 250 emails a day in week 1 to 2,000 in week 4.
“Your initial send list should comprise existing customers, recent subscribers, or website visitors with high on-site activity.”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer