What is email warm-up and why does it matter?

Email warmup is the practice of sending small, gradually increasing volumes from a new domain or IP so mailbox providers can build a trust score for your sender identity before you send at scale. Done right it takes four to eight weeks; done wrong, your first campaign lands in spam and drags the rest of the program down with it.

Google's Postmaster guidance is blunt about the mechanism: a gradual increase in sending volume over a period of time is the best way to build a positive sender reputation (Google Postmaster Tools, 2024). That is the whole case for warmup in one sentence. The messy part, which is what this article is about, is knowing why the filters care and when you genuinely need to bother.

What email warmup is

Warmup establishes a positive sending reputation for a new domain or IP by starting at very low volume, usually 50 or fewer emails on day one, then scaling up methodically over several weeks. You are not tricking anything. You are giving the filters the data they need to decide you are legitimate.

Think of it like a credit score. Before you have sent anything, Gmail has no record for @yournewdomain.com, and the default posture is skeptical. Warmup generates the sending history, engagement, and authentication pattern that move you from unknown to trusted. The M3AAWG Sender Best Common Practices document describes reputation as the aggregate of authentication, complaint rate, engagement, and sending pattern consistency (M3AAWG, 2015); warmup is what populates that aggregate for the first time.

Two things warmup will not do for you. It will not compensate for bad content, a dirty list, or missing authentication, because broken SPF, DKIM, or DMARC sinks you regardless of ramp schedule. And it is not a one-time checkbox; reputation decays once you stop sending consistently, and a long gap can force a partial re-warm. The full email warmup definition in our glossary covers the terminology if you need the short version.

Why filters punish a cold send

Mailbox providers treat sudden volume from unknown senders the way your bank treats a $10,000 wire to a country you have never visited: suspicious until proven otherwise. Without a ramp you look statistically identical to a spammer on the filter's first impression, and the penalty compounds with every additional send.

Three separate mechanisms do the punishing. The first is throttling, where Gmail and Outlook queue, delay, or deferral-bounce mail from unknown domains with 421 errors. The second is junk folder routing; Outlook.com is aggressive with new senders, and without SNDS data confirming you are legitimate, chunks of a first campaign go straight to Junk (Microsoft SNDS, 2024). The third is self-inflicted: a first campaign to unengaged recipients spikes your bounce rate and spam complaint rate, which poisons the reputation you have not finished building.

The signals filters watch fall into three buckets. There is sender reputation, the long-running score. There is engagement, meaning opens, replies, clicks, and adds-to-contacts. And there are throttling thresholds, meaning volume that stays inside the provider's tolerance for a new sender. Warmup improves all three at once, which is why nothing else substitutes for it.

Domain warmup and IP warmup are different systems

Domain warmup builds reputation for your sending domain, the part after @ in your from address. IP warming builds reputation for the IP address the mail actually leaves from. They are separate systems, and in 2026 domain reputation carries more weight than IP reputation for most senders.

Gmail has been explicit about the shift. Its Postmaster Tools documentation states that domain reputation has a larger impact on deliverability than IP reputation for senders using shared IPs (Google Postmaster Tools, 2024). If you are on a shared pool at any modern ESP, the IP reputation is already established by the thousands of senders sharing it; your marginal contribution does not meaningfully shift it, so your job is mostly to warm the domain.

The portability row is the one people miss. Spend a year building a pristine domain reputation on one platform and then move, and the domain score travels with you; the dedicated IP's reputation stays behind. That asymmetry is a decent argument for treating the domain as the asset and the IP as plumbing.

Shared infrastructure makes warmup easier, never optional. Sending 50,000 emails on day one from a brand-new domain gets filtered no matter which well-behaved IP delivered them.

When you actually need to warm up

You need to warm up whenever the mailbox provider has no recent positive history for your sender identity. That covers more situations than most teams expect.

The obvious ones: a brand-new sending domain, a new marketing subdomain (moving from yourco.com to mail.yourco.com), a migration to a new ESP, or a jump from transactional-only to marketing sending on the same domain.

The less obvious ones are where teams get caught. If you paused sending for 60 days or more, reputation has decayed, and Postmaster Tools will often show a previously High domain drop to Medium. If next week's volume is ten times last week's because of a seasonal peak or a launch, even a warm domain needs a mini-ramp. And if you restructured authentication, rotating DKIM keys or flipping DMARC from p=none to p=reject, partial re-warming is safer than assuming the old reputation transfers cleanly.

You do not need a full warmup if you are staying inside your established volume band on the same domain and IP with clean engagement. Just keep sending consistently; consistency is most of the job.

How long it takes

Most domains need four to eight weeks of deliberately paced sending to reach full volume with clean reputation. The exact length depends on your target daily volume, your engagement rates, and the mailbox provider mix of your list.

As rough ranges: under 1,000 sends a day, two to three weeks. Between 1,000 and 10,000 a day, four to five weeks. Between 10,000 and 100,000, six to eight weeks. Above 100,000 a day, budget eight to twelve.

Engagement is the accelerator. A list that opens at 35% and clicks at 6% warms faster than one that opens at 12% and clicks at 0.4%, because filters reward positive recipient behavior rather than volume itself. That is the whole reason the first week of warmup should go to your most engaged 1%, not a random slice. Track open rate and click rate by week during the ramp; a flat or falling trend is a signal to hold volume rather than push it.

The myths that get domains burned

The honest downside nobody advertises: warmup is boring. You send small volumes to small segments for weeks while the campaign calendar sits idle, and there is no shortcut that does not end with a burned domain. Here are the beliefs that lead teams to try anyway.

Warmup fixes bad lists. It does not; if 30% of your list bounces, a slow ramp exposes that faster than a cold blast would. Clean the list first, and the email bounce rates guide has the playbook.

Transactional mail alone warms you up. Transactional volume helps, but filters read the full sending pattern. If your target state is marketing sends, the warmup has to include marketing-like content at some point.

Seed list tools are enough. Sending to 30 seed inboxes daily does not replicate real engagement; those accounts never reply, click through, or save your address to contacts. Seed tests are monitoring, not a substitute for recipients.

Buying a warmed-up domain lets you skip the process. Resold domains with sending history are mostly a scam, and on the rare occasion the history is real, the reputation is usually already damaged.

The biggest actual mistake I see is doubling volume and content aggressiveness in the same week. If week two is twice the volume of week one and also introduces promotional content for the first time, you have changed two variables at once; when deliverability dips, you will have no idea which one caused it.

How to tell the warmup is finished

Three signals have to line up. Postmaster Tools shows your domain reputation as High for seven consecutive days. Your inbox placement rate holds above 90% across a seed test. And you have sent at target daily volume for three days with no spike in spam rate or bounces. If any of the three looks shaky, extend the ramp; a week of patience is much cheaper than a recovery.

You can run real campaigns during the back half of a warmup, as long as they fit inside the ramp volume and pull from your most engaged segments first. Weeks one and two are the exception; those belong to warmup traffic alone.

For the mechanical schedule, with week-by-week volumes, segment selection, and what to send when, use the step-by-step domain warmup guide. For everything downstream of reputation, the email deliverability guide covers authentication, content signals, and monitoring end to end.