A new sending domain has no history with Gmail, Yahoo, Microsoft, or other mailbox providers. That is the real problem warmup is trying to solve. There is no Google-approved schedule that says a new mailbox is safe at 20 messages on Monday, 40 on Tuesday, and 80 on Wednesday. Google’s published guidance is deliberately less mechanical: authenticate the domain, begin with low volume to people who are likely to want the mail, increase slowly, avoid bursts, and watch server responses, spam rate, and domain reputation as volume changes.
Day zero: make every sender visible before you send
List every system that can emit mail using the domain. Include the human mailbox, CRM, sequencing tool, newsletter platform, website form, invoicing service, support desk, and any transactional provider. For each one, record its From domain, return-path or envelope-from domain if exposed, DKIM selector, and the type of mail it sends.
Then send test messages through every path. In a received Gmail message, open “Show original” and look for SPF, DKIM, and DMARC results. You want the production path to pass—not merely a DNS checker to say that a TXT record exists. A common failure is an SPF record that names Google Workspace but omits the CRM; another is a DKIM record published in DNS while signing is still disabled inside the sending platform.
For Gmail recipients, all senders need SPF or DKIM; bulk senders have stricter requirements including SPF, DKIM, DMARC, alignment, and one-click unsubscribe for qualifying marketing traffic. Yahoo publishes similar authentication and complaint-rate expectations. A small business should configure all three authentication layers even when its current volume is below a bulk-sender threshold.
A four-week ramp is a test plan, not a provider limit
Because mailbox providers do not publish a universal safe quota for a new domain, any numeric ramp has to be treated as an internal experiment rather than a compliance threshold. For a small outbound mailbox, a conservative test plan can begin in the low tens and add volume only after the previous cohort behaves normally:
| Phase | Illustrative daily range per mailbox | Audience | | --- | ---: | --- | | Week 1 | 5–20 | existing contacts and recipients likely to recognize you | | Week 2 | 15–35 | mostly warm contacts; a small tested prospect segment | | Week 3 | 25–50 | controlled prospecting if prior signals remain clean | | Week 4 | 40–75 | move toward normal operating volume only if delivery is stable |
Treat those figures as a laboratory starting range, not a promise of inbox placement. If your normal business mail is only 25 messages a day, you do not need to manufacture 75. If you eventually need much higher volume, a longer ramp is usually easier to diagnose than a steep one.
Do not add every variable at once. Keep copy, recipient source, sending system, and daily timing relatively stable while you change volume. Otherwise a sudden rise in bounces could come from a poor list, a DNS change, or the ramp itself and you will not know which one caused it.
What to watch every morning
Start with SMTP outcomes. Separate hard failures from temporary deferrals. A 5xx response such as an invalid recipient should not be retried forever. A 4xx response can indicate a temporary condition or rate limiting and deserves a different response. Save the full diagnostic text because provider-specific language is often more useful than a dashboard’s generic “bounce” label.
Next, inspect authentication on real received messages. If DMARC fails after a platform change, pause the ramp until the path is understood. Authentication problems are infrastructure problems; sending more mail does not cure them.
For senders with enough Gmail traffic, Postmaster Tools can show spam rate, domain or IP reputation, authentication data, and delivery errors. Low-volume senders may see sparse or missing data, so an empty chart is not a clean bill of health. Combine Postmaster with your own bounce log and campaign data.
Hold the ramp when the evidence changes
Stop increasing volume after a sudden hard-bounce increase, a cluster of rate-limit responses, new policy rejections, authentication failures, or visible complaint trouble. Keep volume flat—or reduce it—while you identify the cause. The wrong response is to continue scaling because a spreadsheet says “week three.”
Also pause when a new list source is introduced. Test that source as its own cohort. If 80% of your historical contacts behave normally but one purchased or scraped segment creates most of the invalid recipients, the domain is not the first thing to fix.
When the domain is actually “warmed”
Warmup is finished when the business has a repeatable sending pattern, not when a calendar expires. You should be able to name every sending service, show passing authentication on real messages, identify where bounces came from, suppress bad addresses, and explain why the next volume increase is reasonable. Continue to avoid sudden bursts after the first month; reputation is ongoing behavior, not a one-time ritual.
For a full domain warmup and deliverability playbook, see: {{BACKLINK_1}}