How to warm up a new sending domain or IP address

A new domain or a new IP address has no sending history, so mailbox providers have nothing to judge it by except what it does from the first message on. Warming up means building that history on purpose: start small, send mail people want, and increase volume only as fast as the signals stay good.

None of this is a published standard. Gmail, Outlook.com and Yahoo describe principles, not schedules, and the right pace depends on your volume, your list and how your recipients react.

Why reputation is built gradually

Providers rate senders by how recipients respond over time: whether messages are read, replied to, marked as spam or bounced. A sudden burst of mail from a sender without history is exactly what compromised servers and spammers look like.

The providers say so directly:

• Google’s sender guidelines ask you to “start with a low sending volume to engaged users, and slowly increase the volume over time”, to send at a consistent rate rather than in bursts, and to avoid sudden spikes if you have no history of large volumes.
• Microsoft notes that IPs not previously used to send email typically have no reputation at Outlook.com, so their mail is more likely to run into delivery problems until they earn one.

Reputation is attached to more than one identifier. Providers track IP addresses and domains; Yahoo, for example, says that each IP and each DKIM domain has a reputation. That gives several warm-up situations:

Situation	What is new	What to expect
New domain on an email service’s shared IPs	the domain	the IPs have history; your domain still has to earn its own
New dedicated IP, established domain	the IP	Microsoft says a domain with a good reputation can shorten the ramp-up
New domain and new IP	both	the slowest case — be the most conservative
Existing list moved to a new email service	usually the IP, sometimes the DKIM domain	treat it as a warm-up, not a switch-over

Step 1: Authentication and infrastructure first

Warming up with broken authentication builds a bad reputation, not a good one. Before the first real message:

• SPF authorizes the server or service that sends — SPF checker.
• DKIM signs with your own domain, not only the service’s — DKIM checker.
• DMARC is published, at least with p=none, and SPF or DKIM aligns with the From domain — DMARC checker; see how to set up DMARC.
• Reverse DNS: your own sending IP needs a PTR record whose name resolves back to the same address — reverse DNS lookup and reverse DNS for mail servers.
• TLS for outgoing connections. Gmail requires TLS and valid forward and reverse DNS from all senders.
• One-click unsubscribe in marketing mail (List-Unsubscribe and List-Unsubscribe-Post), required by Gmail and Yahoo for bulk senders — see sender requirements.

Then send a test to a mailbox you control and paste its headers into the email header analyzer: you want spf=pass, dkim=pass for your domain and dmarc=pass.

Step 2: Start with your most engaged recipients

The first messages set the tone. Send them to the people most likely to read them and least likely to complain:

• people who signed up or bought something recently;
• customers who have opened, clicked or replied in recent months;
• transactional mail such as receipts and password resets, if it moves to the new domain or IP — recipients expect it.

Leave old, unconfirmed or inactive segments out until the reputation is established, and never warm up with a bought list — Google’s guidelines tell senders not to purchase addresses at all. If your list hasn’t been cleaned in a while, do that first: see email list hygiene.

Step 3: Increase volume gradually

Google names three factors that decide how fast you can go:

• how much you send — the larger the volume, the more slowly you should increase it;
• how often you send — daily sending lets you ramp up faster than weekly sending;
• recipient feedback — send only to people who subscribed.

Microsoft says a new IP can usually expect to be fully ramped within a couple of weeks or sooner, depending on volume, list accuracy and junk complaint rates. That describes what typically happens; it is not a guarantee.

The working rule: raise the volume only after the previous step looked healthy, and keep the daily volume steady instead of one large send a week.

Phase (example)	Daily volume (example)	Recipients
First days	a few hundred	the most engaged only
End of week 1	a few thousand	recently active recipients
Week 2	roughly doubling every day or two while metrics hold	add older but confirmed recipients
Afterwards	target volume, at a steady daily rate	the full confirmed list

When the numbers get worse, don’t push through. Google’s advice is to reduce the volume when messages start bouncing or being deferred, until the error rate comes down. Hold, find the cause, then continue.

Step 4: Monitor while you ramp up

Google Postmaster Tools (postmaster.google.com) shows data about mail sent to personal Gmail accounts once you verify your domain: the user-reported spam rate, authentication, encryption, delivery errors and compliance with Gmail’s sender requirements. Google asks senders to keep the spam rate below 0.10% and never reach 0.30%. Data appears only when daily volume is high enough, so an empty dashboard in the first days is normal. Google has announced that the older Domain and IP Reputation dashboards will be retired, so don’t build your process around them.

Microsoft SNDS (Smart Network Data Services) gives data about your sending IPs at Outlook.com, including user complaints, and is where you join the Junk Email Reporting Program (JMRP). It works per IP address, so it is most useful when you control your IPs. SNDS has moved to a new portal; the old address redirects there.

Yahoo offers a Complaint Feedback Loop (CFL) that reports complaints from Yahoo users.

Your own logs matter as much:

• Deferrals (4xx replies) are temporary refusals, often rate limits. A rising share means you are going too fast.
• Bounces (5xx replies) point to list quality problems.
• Complaints from feedback loops: unsubscribe the person right away.
• Placement: check where test messages land in your own accounts at the major providers.

Bounces and complaints during warm-up

A new sender has no history to balance out a bad send, so problems weigh more than they would for an established one.

• Stop mailing an address after a permanent failure such as 550 5.1.1. Microsoft’s policies say senders must not retry a message after a permanent 5xx failure and must stop sending to a recipient after multiple non-delivery responses.
• Don’t keep mailing addresses that bounce again and again, even with temporary errors: Google recommends automatically unsubscribing recipients who have multiple bounced messages.
• Make unsubscribing easier than complaining, and process requests promptly.

For the wider picture, see why emails go to spam.

Dedicated or shared IP?

	Shared IP at an email service	Dedicated IP
Reputation	pooled with other customers	yours alone
Warm-up	the service keeps the IPs in use; you warm up your domain	you warm up both the IP and the domain
Volume	works for low or irregular volume	needs steady volume to build and keep a reputation
Risk	other senders’ behaviour affects you	only your own behaviour counts
Monitoring	the service watches its IPs; you watch your domain	you can register the IP in SNDS and check blocklists yourself

Google puts the shared-IP trade-off plainly: the activity of every sender on a shared IP affects the reputation of all of them. A dedicated IP pays off only when you send enough mail, regularly enough, for providers to form a view of it. With small or irregular volume, a well-run shared pool is usually the better choice — and your domain reputation goes with you either way.

Common mistakes

• Full volume on day one after moving to a new email service or IP.
• A new domain to escape a bad reputation. If the cause — consent, list quality, content — stays, the new domain soon earns the same reputation.
• Warming up with the worst part of the list: old, reactivated or imported addresses, where bounces and spam traps concentrate.
• Irregular bursts instead of a steady daily rate.
• Counting on artificial engagement. Automated “warm-up” traffic between seed mailboxes doesn’t change how your real recipients respond, and their response is what providers measure.

Checklist

1. SPF, DKIM with your own domain and DMARC pass and align; PTR and TLS are in place.
2. Marketing and transactional mail are separated; marketing mail has one-click unsubscribe.
3. The list is clean, and the first sends go to recent, engaged recipients.
4. Volume goes up only after a healthy step, at a steady daily rate.
5. Postmaster Tools and SNDS are set up before the first send; bounces, deferrals and complaints are checked daily.
6. When the metrics worsen: hold or reduce, fix the cause, then resume.