How to Warm Up a Cold Email Domain in 14 Days

Most warm-up schedules you will find are a table of numbers with no explanation of what the numbers are for. They tell you to send 10 on Monday and 20 on Tuesday, and they are silent on the only question that matters when something goes wrong: how do I know whether to continue?

This is a schedule, but the schedule is the least important part of it. What determines whether a new domain ends up with a healthy reputation is not which numbers you followed — it is whether you were willing to stop following them when the evidence said to.

What warm-up is actually building

A new domain has no history. To a receiving filter, that absence is not neutral; it is a risk. Warm-up is the process of replacing "we know nothing about this sender" with "we have seen this sender behave well", and the evidence that does that work is engagement, not volume.

This matters because it changes what a good warm-up looks like. Twenty messages a day to people who reply builds reputation faster than two hundred a day to people who ignore you. If you take one thing from the schedule below, take this: the ramp is a way of buying time for engagement signals to accumulate, and volume that produces no engagement is not buying anything.

It also explains why the artificial warm-up services are a poor deal. They generate traffic between accounts they control, opening and replying to their own mail to simulate a healthy sender. Providers have had years to learn what that pattern looks like, and the reputation it produces is a claim about behaviour that never happened. A slower ramp on real mail to real prospects builds something that holds up.

The schedule

The table in this article gives the day-by-day volumes. Rather than repeat them here, it is worth explaining the shape.

Days 1–2 are a test, not a campaign. Five to ten messages, sent to the contacts most likely to write back — existing acquaintances, warm introductions, people who have interacted with your company before. You are not prospecting yet. You are confirming that authentication resolves, that nothing bounces, and that replies happen.

Days 3–7 introduce real prospects, still at low volume and still chosen for likelihood of reply rather than list order. This is the point where most people go wrong, because it feels slow and the temptation to jump ahead is strongest when nothing has visibly broken.

Days 8–14 increase steadily toward roughly 40–50 a day. Note the final row of the schedule: do not spike on day 14. Ending the two weeks with an unusually large send undoes the pattern you spent two weeks establishing, and it is a common way to finish a careful ramp with a reputation hit.

Day 15 onward is not "done". A domain at 50 a day is a domain that can begin real campaigns, not one that can send five hundred. Continue increasing by roughly 20% a week under live campaigns, watching the same numbers.

The numbers that override the schedule

Three metrics decide whether you continue, hold, or stop.

Bounce rate should stay below 3%, and in the first few days should be zero. Bounces at day 1 do not mean your domain is bad; they mean your list is. Fix the list before continuing, because every bounce during warm-up is a disproportionately expensive signal against a domain with no history to absorb it.

Complaint rate should be effectively zero at these volumes, and must stay under 0.1% as you scale. One complaint in your first week at 10 messages a day is a 1% complaint rate, which is catastrophic in context. Treat any complaint during warm-up as a reason to stop and re-examine the targeting.

Reply rate is the one people forget to watch, and it is the most informative. If bounces are low and complaints are zero but nobody is replying, the domain is technically healthy and the campaign is not working. Continuing to ramp in that state does not build reputation — it accumulates a history of unengaging mail, which is worse than no history. Fix the message or the targeting before adding volume.

Warming addresses, not just domains

A subtlety that catches teams in week three: reputation attaches to sending addresses as well as the domain. A carefully warmed domain does not confer instant credibility on a mailbox created yesterday.

If you plan to send from several addresses, introduce them gradually too — one at a time, each starting low, rather than adding three at full volume once the domain "is warm". The domain's history helps, but a new address sending at the domain's current ceiling on its first day is its own spike.

The same logic applies to teams. Adding four colleagues to a workspace that shares one sending domain does not multiply what that domain can safely send; it divides it. The domain's reputation is the shared resource, and everybody's volume draws on the same pool.

RepMail applies a progressive daily limit to new senders automatically — 50 a day to begin with, rising to 100 and then 200 as the sending history builds — so the ramp is enforced by the platform rather than left as something you have to remember. A campaign that exceeds the day's remaining allowance is scheduled to continue when the window reopens rather than failing, and the dashboard shows which step you are currently on, so "why is my limit this number" is answerable at a glance. The numbers above still apply to the judgement calls the platform cannot make for you: who you send to, and whether they reply.