Domain warming
Ramping a new sending domain from a trickle to full volume over weeks, so providers learn to trust you instead of flagging a sudden blast.
See it
What it is
A brand new sending domain has no reputation, and to a filter, no history plus sudden volume looks exactly like a spammer who just registered a throwaway. Warming is the fix: start far below your target, add volume only while the receiving side keeps accepting cleanly, and reach full rate over however long that takes. The starter numbers you see quoted (a few dozen messages a day, growing something like 30% to 100% every couple of days, done in four to six weeks) are one template, not a rule. A ramp to 5,000 a day and a ramp to a million are different exercises, and a domain with existing sending history starts from a different place than a fresh one.
Order matters more than the numbers, and the provider sets the pace. Send to your most engaged recipients first, because replies and clicks buy trust far faster than raw volume does. Then let the response steer each step: hold or halve the next increase when 4xx deferrals, complaints, or bounces climb, and step up sooner when everything stays clean. Warm each provider separately, since Gmail, Outlook, and Yahoo track you independently and one of them will lag. If you also moved to a dedicated IP, that is a second, parallel ramp (IP warming), and the domain half is the one that follows you forever.
The trap: warming a cold or purchased list is worse than not warming at all, because you are patiently teaching Gmail that you are a spammer, one small batch at a time. Two more details: a domain registered yesterday draws more suspicion than an aged one, so buy the sending domain before you need it rather than blasting on day one, and warmth decays, so a domain that goes quiet for a month or two needs a short re-ramp rather than a straight return to full volume.
Ask AI for it
Build a warming schedule for a new sending subdomain, mail.example.com, whose target is 50,000 marketing sends per week. Treat the ramp as adaptive rather than fixed: propose a starting daily volume and a step size, but make each step conditional on the previous few days of deferrals, complaints, and bounces, and say what holds the ramp flat and what rolls it back. Output a week-by-week table with daily volume, the recipient segment to use each week (most engaged first, defined by clicks and replies in the last 90 days rather than opens), and the split across Gmail, Outlook, and Yahoo. Add send-time pacing so no single hour spikes. Include stop conditions: hold volume flat if the complaint rate passes 0.1%, and pause and investigate if hard bounces pass 2% or any provider starts deferring with 4xx codes. Assume SPF, DKIM, and DMARC are already aligned on the subdomain.