New IPs to full volume,
without the blocks.

Every mailbox provider keeps a reputation score for each IP that sends to it, built from sending patterns, engagement, complaints and bounces. A brand-new IP has no score at all — it’s a blank slate. And from a provider’s side, a blank-slate IP that suddenly fires 10,000 messages on day one is indistinguishable from a spammer’s freshly provisioned server doing the same thing. So it gets treated as one: deferred, throttled, or dropped straight into spam. That’s not a glitch; it’s the system working as intended on an unknown sender.

Warm-up is how you fill that blank slate with positive data before you ask for trust. By starting small and raising volume gradually, you give each provider time to watch real recipients open, read and reply, and to conclude you’re a legitimate sender rather than a burst of spam. There’s no shortcut and no global reputation database to inherit — each provider builds its own view of you, which is why the ramp is run per provider and why patience is the whole technique.

Most warm-ups that fail don’t fail on the schedule — they fail on the list. Send to a dirty list and you collect bounces and spam-trap hits in week one, which is the fastest way to tell a provider your brand-new IP belongs to a spammer. So warm-up genuinely starts before email number one: the list is verified, the dead and risky addresses removed, and the bounce rate kept under roughly two percent throughout. If bounces climb above that in the first week, the right move is to stop and fix the data, not push on with the ramp.

Two more things have to be true before the first send. Authentication — SPF, DKIM and DMARC, with the From, DKIM and Return-Path domains aligned — must already be live, because misaligned headers are a trust signal providers flag immediately. And the ramp should open on your most engaged subscribers: the people who opened or clicked in the last 30 days, and nobody else, for the first weeks. Engaged recipients generate the positive signals that build reputation; sending to the disengaged early just generates indifference the providers notice.

A full warm-up takes roughly four to eight weeks, and where you land in that range depends mostly on your target volume and list quality. A clean, engaged list aiming for modest volume — under about 10,000 a day — can stabilise in three to four weeks. A large or mixed-engagement list, or a target of 100,000-plus a day, should expect six to eight weeks before providers grant full trust. Anyone promising a fixed two-week number is selling a schedule the inboxes won’t honour.

The most common self-inflicted failure is declaring victory early. Two weeks of good-looking metrics is not a warmed IP; reputation takes the full window to establish, and jumping to full volume on the strength of an encouraging first fortnight undoes the work. The ramp ends when the provider signals say it’s ready — steady acceptance at target volume across receivers — not when the calendar or your patience says so.

Warm-up only makes sense if a dedicated IP makes sense, and for plenty of senders it doesn’t. If you send under roughly 50,000 messages a month, or your volume is irregular, a shared IP from a reputable provider is usually the smarter choice: its reputation is pooled across many senders, which is something a low-volume sender simply can’t build alone. A dedicated IP you can’t keep consistently busy doesn’t earn trust — it just sits there looking suspicious, a problem you manufactured by insisting on one. We’d rather tell you that than warm an IP you’d be better off without.

Warm-up earns its place once you’re past that threshold: a genuine dedicated IP or pool, consistent volume to keep it active, and a real list of people who chose to hear from you. The clean-lane line matters here too. A legitimate IP is warmed with a real volume ramp to your own engaged recipients — actual opens and replies — not with the seed-inbox engagement bots aimed at cold-email senders. Manufactured signals are a different game played for a different, lower purpose; we run the real ramp, and only for opt-in mail.

Building a sensible day-by-day ramp is something you can do yourself, and our free warm-up scheduler will lay one out for you. A schedule on paper, though, is the easy half. The hard half is that a static plan can’t see what’s happening: it doesn’t know that Gmail started deferring on day nine while Microsoft stayed happy, and it can’t slow one provider while pressing on with another.

That live adjustment is what you’re paying for. The service enforces the ramp in your PowerMTA, watches each provider’s response every day, and re-paces per provider the moment one pushes back — reading deferrals as the feedback they are rather than barrelling through them. It also carries the discipline not to declare the IP warm a fortnight early. Over four to eight weeks of daily attention, that judgement is the difference between an IP that reaches full volume cleanly and one that gets throttled in week three and has to start over. If your situation is simple and steady, the free scheduler may be all you need — and we’ll say so.

The failure mode is vivid and common: a team provisions a fresh dedicated IP, sends its first real campaign to ten thousand contacts, and watches inbox placement crater — often into the teens — while the bounce rate spikes. The IP’s reputation is damaged before it ever did anything useful, and damaged reputation isn’t undone by easing off. It has to be rebuilt slowly, which means the warm-up you skipped now has to happen anyway, from a worse starting point, with a provider that has already filed you under suspicious.

That’s the real economics of warm-up: it isn’t an optional polish, it’s the cheapest path to volume. A patient ramp that takes a few weeks is far less costly than a cold blast that burns the IP and then forces a longer, harder recovery on top of the campaign you already wasted. Doing it right the first time is slower for a month and faster forever after; doing it wrong is quicker for a day and expensive for a quarter.

Reputation attaches to two separate things: the IP that sends the mail and the domain in the From address. They’re warmed on the same principle but they’re not the same job. A new IP under an established domain leans partly on the domain’s existing standing; a genuinely new domain or subdomain has to earn its own trust regardless of how warm the IP is. When you’re moving to a new sending platform or launching a fresh identity, domain warm-up is always part of the picture — IP warm-up alone won’t cover a brand-new from-domain.

One practical rule that catches people out: warm shared-pool and dedicated traffic separately, and don’t mix streams across a warming IP. Mixing confuses the signals each is trying to build, so a transactional stream and a marketing stream coming online together are kept apart rather than blended. Where both an IP and a domain are new at once, we run the two ramps in step so neither gets ahead of the other — because a warm IP under a cold domain, or the reverse, still reads as a half-trusted sender to the providers weighing both signals together before they decide where your mail lands.