Your sender reputation doesn't come with you when you switch email providers. The trust you've built with Gmail, Yahoo, and Microsoft is attached to the infrastructure you're leaving behind, and the moment your mail arrives from new IPs, mailbox providers treat you with the suspicion they reserve for strangers. Warming is how you rebuild that trust deliberately. We've covered the cutover mechanics of switching providers before; this guide covers the warming half of the job.
TL;DR: If you send serious volume, ramp up gradually. The schedule is below.
What actually needs warming
Mailbox providers score two things separately: the reputation of the IP address your mail arrives from and the reputation of the domain it claims to be from. A migration can leave you cold on one, the other, or both.
Established domain, shared IPs. The most common migration and the gentlest. AhaSend's shared IPs carry an established, actively maintained reputation you inherit on day one, and your domain keeps its history. Expect some initial caution from mailbox providers, nothing more.
New subdomain, shared IPs. A fresh subdomain has no history of its own. It inherits some benefit of the doubt from the parent domain, but it needs to earn its own record. Ramp deliberately.
A cold dedicated IP. A fresh dedicated IP that suddenly sends at volume looks exactly like a spammer who just spun up infrastructure. This is the scenario the day-by-day schedules were written for.
Even the "no warming needed" case benefits from a gradual switch, because the combination of your domain and new infrastructure is new to the mailbox providers. Read about the canary approach here.
Before you move a single message
- Inventory your mail streams. Every kind of email, its sending address, and its daily volume and peaks. Prevents a forgotten cron job from blasting an unwarmed address mid-migration.
- Split streams onto subdomains. Transactional on its own subdomain, separate from marketing. Reputation is scored per domain, so a bad campaign can't drag down your password resets.
- Authentication. Nothing to remember here: AhaSend requires SPF, DKIM, and DMARC before a domain can send at all, and verifies your DNS continuously afterward. See the domain setup guide.
- Bring your suppression list. The most skipped step and the most expensive to skip. Import it before the first real send: how to import a suppression list into AhaSend.
The ramp
The strategy differs per type of mail. Transactional mail has a built-in advantage: it's inherently engaged. A password reset or an order confirmation is mail the recipient is actively waiting for, which produces exactly the signals warming needs. You don't pick recipients there, your users do. Marketing mail is the opposite: you do pick the recipients, so pick well. Start with your most engaged segment, the people who opened or clicked in the last 30 days, then widen to less recent segments as clean days accumulate, and keep the coldest part of your list for last.
What you control on both sides is which mail moves, and how fast:
- Migrate stream by stream. Move one well-defined stream first, say email verification for new signups, then add streams week by week as the identity builds history.
- Use percentage routing. Send 10 percent of a stream through AhaSend and shift the split upward as metrics stay clean, the canary approach.
- Batch what isn't time-sensitive. Resets go immediately, but campaigns, digests, and reminders can be spread across the day. Smooth volume warms faster than bursts.
The pace: double the daily volume until you reach 300,000 per day, then slow down to steps of 150,000 and later 100,000 per day. Only step up when the previous level stayed clean.
| Day | Daily volume (new identity) | Condition to continue |
|---|---|---|
| 1 | 200 | Bounces and complaints in the normal range |
| 2 | 400 | Same, no reputation warnings |
| 3 | 800 | Same |
| 4 | 1,500 | Same |
| 5 | 3,000 | Same |
| 6 | 6,000 | Same |
| 7 | 12,000 | Metrics stable across a full week, including your peak days |
| 8 | 25,000 | Same |
| 9 | 50,000 | Same |
| 10 | 100,000 | Same |
| 11 | 200,000 | Same |
| 12 | 300,000 | Doubling ends here |
| 13 | 450,000 | Steps of +150,000 per day |
| 14 | 600,000 | Same |
| 15+ | +100,000 per day | Continue until you reach your production volume |
Three notes. If your production volume is lower than a row, you're done sooner: at 5,000 emails a day the ramp takes under a week. If your list is heavily concentrated on one mailbox provider, treat the daily number as a cap per provider, Gmail is the strictest. And the table sets the pace, but your bounce and complaint rates hold the veto.
How this fits the percentage schedule in the migration guide: percentages are your control knob, but mailbox providers see absolute volumes. Five percent of a million messages a day is 50,000 messages from an identity nobody knows, which is not a gentle start. At serious volume, pick a starting percentage that puts your first days in the low thousands and let each step roughly double the absolute volume.
Watch the dials
Set up webhooks for delivery, bounce, and complaint events, and use the deliverability reports for aggregate rates.
Decide your pause rules in advance: hard bounces below 3 percent, spam complaints below 0.1 percent. If either climbs, hold or step back a level until it stabilizes. Adding three days to a migration is always cheaper than digging a new identity out of the spam folder. A temporary dip in open rates is normal; rising bounces or complaints are not.
Warming never fully ends
Reputation is maintained, not achieved. Two situations re-open the question after migration: a big volume jump (a launch, a seasonal peak, an acquired user base), which you should ramp into rather than step into, and an identity going quiet, because reputation decays without traffic. The habits that carry you through the migration, steady volume, clean lists, separated streams, metrics you actually read, are the same ones that keep you warm afterward.