Warming up a sending domain: a deliverability guide

A new domain starts with zero reputation

Warming up a sending domain means gradually building its reputation with mailbox providers (Gmail, Outlook, Yahoo…) before sending campaigns at full load. A brand-new domain has no history: to spam filters, it is indistinguishable from a domain bought yesterday by a spammer. Sending 20,000 emails on day one from such a domain is the most reliable way to land in spam — sometimes for a long time.

The logic is easy to state, hard to apply: providers judge a sender on its past behavior. No history means no trust. Warming up therefore means producing, week after week, a clean history — low volumes, recipients who open and reply, almost no complaints — until the domain is recognized as legitimate.

The trap in generic guides is that they hand you a fixed schedule ("week 1: 50 emails, week 2: 100…") disconnected from what the domain is actually experiencing. A warm-up is steered by signals, not by a frozen spreadsheet. This article gives the decision framework to read those signals and adjust the pace, plus the prerequisites you cannot skip.

What must be in place BEFORE you warm up

Warming up a misconfigured domain is like accelerating with the handbrake on: you build bad reputation faster. Before the first send, check these foundations.

• Aligned authentication: SPF, DKIM and a published, coherent DMARC record. Without them, some of your messages will be rejected or filed as spam regardless of the pace. The full setup is covered in our SPF, DKIM and DMARC guide.
• A dedicated sending subdomain: send your campaigns from a subdomain (e.g. news.yourdomain.com) separate from your business mailbox. That way a reputation incident on campaigns does not affect your work email.
• A genuinely clean list: opt-in, recent, valid contacts only. A purchased or three-year-old list generates the bounces and complaints that kill a warm-up.
• Ready, non-promotional content to start: the first sends should look like expected messages (welcome, service information), not an aggressive promo stuffed with links.
• A measurement loop: access to your tool's stats (opens, clicks, bounces, complaints) and ideally to the major providers' feedback loops.

These prerequisites belong to the same technical foundation as deliverability in general, detailed in our email deliverability guide. If one of them is missing, fix it before starting the ramp-up.

An example ramp-up (framework, not dogma)

Here is an example of a cautious progression over roughly four to six weeks. Read it as an illustration of the logic, not an official standard: the right volume depends on your list size, your sector and above all the signals you observe (next section).

Two principles govern this table, more important than the numbers themselves:

1. Start with your best recipients. Contacts who open and click send providers the signal "this sender is wanted." Keep dormant segments for the end, once reputation is established.
2. Increase in steps, not continuously. Double the volume, observe for several days, then decide. A sudden jump (×10 overnight) reads as spammer behavior, even on a domain already warming up.

Segment by provider too if you can: Gmail and Outlook do not share the same tolerance thresholds, and a domain can be welcomed by one while watched by the other.

Reading the signals to decide: accelerate, hold, or pull back

This is where a successful warm-up is won. Each step ends with a decision: raise, hold, or lower the volume? That decision is made on signals, not on the calendar. Here is how to read them.

A few benchmarks to make the reading objective:

• Bounces and complaints outrank volume. One rule: never raise the volume while bounces or complaints are rising. Fix first, accelerate later.
• Watch placement, not just opens. A good open rate can hide placement in the "Promotions" tab or in spam at a specific provider. Placement tests (seed lists) remove the doubt.
• A bad step is not a failure. If a segment triggers complaints, go back to the previous volume, let the domain stabilize for a few days, then resume more gently. Pulling back costs less than burning the domain.

This signal-driven steering is what separates a warm-up that succeeds from a schedule applied blindly that ends up in a spam filter.

The mistakes that burn a new domain

Most failed warm-ups come down to a handful of avoidable mistakes:

• Sending too much, too soon. Mistake number one. The initial volume should look ridiculous next to your target.
• Starting on a questionable list. Purchased base, never-confirmed contacts, addresses collected without consent: guaranteed bounces and complaints, reputation sunk from the start. It is also a compliance issue — see our GDPR guide for email.
• Opening with heavy promotional content. Too many links, images, trigger words and buttons in the first sends: filters react badly on a domain with no history.
• Ignoring a provider that drops off. If Outlook sends you to spam while Gmail accepts you, continuing at full load toward Outlook deepens the problem.
• Reusing the main domain for everything. Mixing marketing campaigns and business email on the same domain exposes your critical mailbox to a reputation incident.
• Stopping sends after the warm-up. Reputation is kept through consistency. A warmed domain left silent for several weeks becomes a stranger again.

How long, and how to know it is done

There is no universal duration. In practice a cautious warm-up often spans several weeks, more if your target volume is high or if signals force you to slow down. Duration is a consequence of the signals, not a target to hit.

You can consider the warm-up complete when several conditions hold over time:

• You send at your target volume without any degradation in opens.
• Bounces and complaints stay low and stable.
• Inbox placement is confirmed at your main providers, not just assumed.
• A one-off volume spike (seasonal peak) no longer causes a collapse.

After that, the goal is no longer to warm up but to maintain: a steady pace, ongoing list hygiene, complaint monitoring. Your choice of tool also matters to keep that consistency and to exploit feedback loops — a point we detail in our email marketing software comparison. A successful warm-up is an asset; it is squandered fast if you stop maintaining it.

FAQ

Do I still need to warm up a domain if I use a shared IP?
A domain is worth warming up even on a shared IP: domain reputation is built independently of the IP. On a shared IP you partly inherit the IP reputation managed by your provider, but your domain is still judged on its own behavior.

Can I warm up a domain automatically with a tool?
Some services automate the ramp-up by simulating exchanges. They can help, but nothing replaces real engaged recipients and genuinely expected content: artificial signals have limits and can be detected.

What should I do if my domain is already burned?
If a domain has accumulated complaints and bounces, stop the mass sends, fix authentication and the list, then restart a slow warm-up from your most loyal contacts. In severe cases, switching to a fresh sending subdomain is sometimes faster than rehabilitating the existing one.

How long before launching a big campaign?
Only launch a full-volume campaign once the warm-up is confirmed by the signals (stable opens, low bounces and complaints, confirmed placement). Rushing a large campaign on a still-fragile domain undoes the warm-up work.