Warmup is not a switch you turn on for two weeks and then forget. It is a ramp, and the ramp continues after the warmup tool stops being involved: a mailbox that has been exchanging warmup mail for three weeks is ready to start sending cold, not ready to send 30 a day.
Here are the schedules, and the more useful part, when to hold.
A new mailbox, week by week
| Week | Cold sends/day | Warmup mail | Notes |
|---|---|---|---|
| 1 | 0 | Yes, automated | Authentication verified before anything sends. No cold traffic at all |
| 2 | 5–8 | Yes | Best-fit segment only. You want replies, not reach |
| 3 | 12–15 | Yes, reducing | Watch bounce rate daily, not weekly |
| 4 | 20–25 | Reduced | Approaching steady state |
| 5+ | 25–30 ceiling | Light, ongoing | Hold. More volume means more mailboxes |
Two rules matter more than the numbers. Send your best segment first. Early replies build reputation faster than anything a warmup tool does, and early bounces damage it faster than anything else. Keep some warmup traffic running permanently, it maintains the positive-engagement baseline that cold sending erodes.
A new domain is slower than a new mailbox
Domain age is its own signal, and Microsoft in particular weighs it heavily. A brand-new domain should sit for one to two weeks before its mailboxes send anything cold, even if the mailboxes themselves are warmed. Add mailboxes to a new domain gradually as well: three mailboxes appearing on a domain the day it was registered is a pattern.
For an existing, healthy domain, the mailbox schedule above is the only constraint.
Scaling an established setup
The safe rule for increasing total volume on a working setup is no more than 20 to 30% per week, and only when the metrics below are clean. Doubling volume because last week went well is the most common way a mature setup degrades, and because placement damage lags by days, it usually looks fine right up until it does not.
The three signals that mean hold
| Signal | Hold at | Roll back at |
|---|---|---|
| Bounce rate | Over 2% | Over 4%: stop, fix the list before anything else |
| Reply rate falling with unchanged copy | Any sustained fall | Down by a third: placement problem, not a copy problem |
| Spam complaints | Over 0.1% | Over 0.3%: cut volume by half immediately |
A fourth, less numeric signal: if you cannot explain why reply rate moved, do not add volume that week. Ramping into an unexplained change is how a small problem becomes a domain you have to abandon.
And confirm placement rather than inferring it. Delivery rate will look fine while everything lands in spam, so a placement test at each ramp stage is what turns this from guesswork into a measurement.
What breaks a ramp
- An unverified list. Bounces during weeks two and three do disproportionate damage, because the mailbox has almost no positive history to offset them.
- Sending in bursts. Twenty-five emails between 9:00 and 9:03 is not the same as twenty-five spread across the day, and providers can tell.
- Adding mailboxes mid-ramp and averaging. Each mailbox has its own age and its own ceiling. Ten mailboxes at week one and ten at week four are not twenty mailboxes.
- Pausing and resuming at the old volume. After a two-week gap, restart a step or two back rather than where you left off.
Next: the full warmup guide, volume ramp schedules in detail, or how many mailboxes you need in the first place.
Frequently asked questions
How long does mailbox warmup take?
About four weeks to reach a steady 25 to 30 cold sends a day: no cold traffic in week one, 5 to 8 a day in week two, 12 to 15 in week three, and 20 to 25 in week four.
How fast can I scale sending volume on an established setup?
No more than 20 to 30% a week, and only when bounce rate, reply rate and complaint rate are all clean. Placement damage lags by days, so a fast ramp looks fine until it does not.
Do new domains need longer than new mailboxes?
Yes. A newly registered domain should sit for one to two weeks before its mailboxes send cold mail, and mailboxes should be added to it gradually rather than all on day one.
When should I stop ramping?
Hold at a bounce rate over 2%, any sustained fall in reply rate with unchanged copy, or complaints over 0.1%. Roll back at 4% bounces or 0.3% complaints.
Should warmup traffic ever stop?
No. Keep a light ongoing volume permanently. It maintains the positive-engagement baseline that continuous cold sending erodes.