How to warm up a new sending domain
How to build sending reputation on a new domain by ramping volume gradually: who needs a warm-up, which recipients and content to send first, two sample ramp schedules, the bounce and complaint signals to watch, and the stop signals that mean pause immediately.
Warming up a sending domain means raising its daily volume gradually so mailbox providers can build a reputation for it before it sends at full scale. A brand-new domain with correct SPF, DKIM, and DMARC can still see half its first large batch deferred or filtered, because correct authentication proves who is sending, not that the sender is trusted. Trust has to be earned through a record of mail that people open and do not report, which is why a warm-up is paced over weeks.
Why warm-up exists
Mailbox providers maintain reputation scores for every sending domain, and a brand-new domain has no score, which is worse than having a mediocre one. An unknown sender behaves exactly like a spammer who just registered a new domain, and providers cannot tell the difference without evidence.
The evidence they need is sending history: messages that are delivered, opened, and not reported as spam. That history has to build over time, because a sudden flood from an unknown domain is itself a spam signal. Fifty emails sent during development all land in the inbox, while fifty thousand from the same new domain are alarming, and overshooting earns deferrals, rate-limiting, or in the worst case a blocklist that takes months to get removed from.
Who needs to warm up
Not every situation needs a warm-up, and the deciding factor is whether the sending IPs already carry a reputation. A warm-up is needed whenever the IPs start at zero:
- A brand-new domain or subdomain that has never sent email before.
- A provider switch where the new provider sends from different IPs. Domain reputation carries over, IP reputation does not.
- A domain coming back into use after months of inactivity.
- A new subdomain for a separate sending stream, such as splitting transactional from marketing.
The warm-up is usually skippable when the reputation already exists or the volume is negligible: a provider switch that keeps the same sending IPs (rare, mostly during dedicated IP migrations), a steady sender adding only a few thousand messages a day, or a very low-volume sender under a few hundred emails a day, where a gradual ramp is indistinguishable from normal sending.
A new domain on Lettr always needs a warm-up, regardless of the previous provider. The authentication records transfer, but the new sending IPs have no established reputation yet, and each new subdomain needs its own warm-up (one reason to separate transactional and marketing email from the start).
What and who to send first
The single most important decision during warm-up is who receives the first sends. Every signal is amplified at this stage. A 2% bounce rate on 100,000 emails is a problem; a 2% bounce rate on the first 500 is an emergency, because it is the only data the provider has. The early sends need to produce strong positive signals: high open rates, low bounce rates, zero complaints.
Sort the recipient list by engagement and start at the top. The first batch should go to the people who opened a recent email, clicked something lately, or actively use the product. They are the most likely to open and the least likely to mark a message as spam. The rest of the list comes in as volume increases.
The newest signups are the wrong place to start. New users have the highest bounce rates (typos and disposable addresses) and the lowest engagement (no habit formed yet). The worst audience of all is a re-engagement campaign aimed at users who have not opened in months.
On the content side, the top-performing email types are the safest. Transactional emails that people expect and open (order confirmations, account notifications) are ideal. A marketing subdomain should warm up on its highest-engagement content, never a cold re-engagement campaign.
Anything that could generate complaints stays off the schedule during warm-up. No re-engagement blasts, no aggressive promotions, nothing aimed at segments left untouched for a while. One bad send in the first week can set the domain back significantly.
Warm-up schedules
There is no universal schedule, because every sender has different volume targets, audience quality, and risk tolerance. The principle is always the same: start low, increase gradually, and watch the signals. The two schedules below are starting points to adjust against the feedback each send produces.
Low-volume senders (target: under 10,000/day)
| Day | Daily volume | Notes |
|---|---|---|
| 1-2 | 100-200 | Most engaged recipients only |
| 3-4 | 500 | Monitor bounce and complaint rates |
| 5-7 | 1,000 | Check Postmaster Tools if sending to Gmail |
| 8-10 | 2,500 | |
| 11-14 | 5,000 | |
| 15+ | Target volume | Maintain steady sending from here |
For low-volume senders, two weeks is usually enough. The ramp feels slow, but it is building a foundation. Reaching target volume three days in saves nothing if Gmail starts deferring everything on day four.
High-volume senders (target: 50,000+/day)
| Week | Daily volume | Notes |
|---|---|---|
| 1 | 1,000-2,000 | Most engaged recipients, monitor everything |
| 2 | 5,000-10,000 | Review Postmaster Tools, check deferral rates |
| 3 | 15,000-25,000 | Expand to moderately engaged recipients |
| 4 | 30,000-50,000 | |
| 5-6 | 50,000-100,000 | Approach target volume |
| 6+ | Target volume |
High-volume warm-ups take four to six weeks, with no way to compress that significantly without risking throttling or blocks. A launch timeline that does not leave a full month for the ramp needs to start earlier. This is easy to plan for and painful to rush.
The signals to watch
Warm-up depends on monitoring three categories of signal and knowing what each one means. Set up Google Postmaster Tools and feedback loops at Yahoo and Microsoft before the first send. Without them, the entire ramp runs blind.
Bounce rate
Track hard bounces (permanent, 5xx) and soft bounces (temporary, 4xx) separately, because they mean very different things. Hard bounces above 2% on any single send indicate poor list quality, and the right response is to stop, clean the list, and resume with verified addresses. During warm-up, anything above 1% is cause for concern.
Soft bounces and deferrals are normal during warm-up, up to a point. Mailbox providers deliberately slow delivery from unknown senders, so a deferral rate of 5-10% in the first few days is unremarkable. A deferral rate above 20%, or one that persists past the first week, means the ramp is too fast.
Complaint rate
This is the most damaging signal. Gmail's published threshold is 0.3% before it takes action, but the warm-up target is zero, because even a single complaint on a small-volume send has an outsized impact.
Complaints in the first week need an immediate diagnosis. The usual causes are recipients who never opted in, content confusing enough to be mistaken for spam, or an unrecognizable "from" name. Fix the cause before sending more.
Open and click rates
High open rates tell providers the messages are wanted, which is why the first sends go to engaged recipients. First sends at 60-70% open rates build a strong reputation foundation; first sends at 15% tell Gmail that most recipients do not care.
Stop signals
Not everything during warm-up is a normal growing pain. Five signals call for an immediate pause, not a push-through:
- Hard bounce rate above 5% on a single send. The list has serious quality issues. Stop sending, remove every bounced address, and consider re-verifying the whole list before resuming. Continuing to send to invalid addresses during warm-up is the fastest way to get blocked.
- Complaint rate above 0.3% on any send. This is past Gmail's threshold. A list-quality cause (recipients who never signed up) needs the list fixed; a content cause (an unrecognized sender) needs the "from" name and template fixed. Do not resume until the root cause is addressed.
- Deferral rate above 30% that is not decreasing. Some deferrals are normal in the first few days, but persistent high deferrals after a week mean the ramp moved too fast. Drop back to a lower volume and hold there for several days before increasing again.
- Postmaster Tools shows a "Bad" reputation. Stop sending to Gmail immediately, reduce volume dramatically, and send only to the most engaged Gmail recipients until reputation recovers. This can take one to two weeks of reduced sending.
- Blocklist appearance. Check whether the domain or sending IP has landed on a major blocklist (Spamhaus, Barracuda, SURBL). If so, pause all sending and work to get delisted first. Sending while blocklisted makes everything worse.
FAQ
How long does it take to warm up a sending domain?
Do I need to warm up a domain if my provider has a good reputation?
What happens if I send full volume from a cold domain?
Who needs to send the first warm-up emails?
Bottom line
One idea sits under every rule here: a warm-up is the sender proving, gradually, that volume from this domain is wanted. A slow ramp is not wasted time. Reputation is the one part of an email stack that cannot be provisioned on demand, which is why it has to be scheduled rather than rushed at launch.
On Lettr, authentication is verified during setup and bounce and complaint events stream through webhooks in real time, so a bad ramp surfaces in minutes instead of days, part of Lettr's deliverability tooling. Create a free account and verify a domain before the next launch needs it.