How to Warm Up a New Sending IP or Domain Without Burning Your Reputation
You’ve signed the contract, migrated the list, and pointed your DNS at the new provider. The first campaign goes out tomorrow. Within 48 hours, open rates collapse, the ESP’s dashboard shows a spike in deferrals, and your sales team is asking why half the pipeline stopped receiving email. This is the most common failure mode in email migration, and it almost never comes from bad copy or a broken template. It comes from sending too much, too fast, from infrastructure that has no history.
Warmup is the process of teaching mailbox providers — Gmail, Microsoft 365, Yahoo, and the smaller filters in between — that your new IPs and domain deserve to be trusted. It is not a formality. A fresh IP has no reputation, which is not the same as a good reputation. Gmail’s spam filter treats unknown senders with suspicion until engagement data accumulates. Microsoft’s filtering is famously unforgiving of new IPs that suddenly push volume, and a single bad first week can leave your domain throttled for months.
This guide gives you a day-by-day schedule, the volume math behind it, and the specific signals that tell you whether to proceed or hold. It assumes you’re either migrating from one ESP to another or starting from a brand-new domain. The two scenarios share most of the playbook but diverge in a few important places, which I’ll flag.
Understand what the mailbox providers are actually measuring
Before you touch a schedule, you need to know what the filters care about. Gmail’s Postmaster Tools exposes four signals: spam rate, IP reputation, domain reputation, and feedback loop performance. Microsoft’s SNDS and JMRP give you similar telemetry. Yahoo’s Complaint Feedback Loop is still worth enrolling in. These are the instruments you’ll read for the next 30 days.
The signal that matters most early is engagement on delivered mail. Providers infer trust from whether recipients open, click, reply, mark as important, or move your message out of spam. A new IP that sends 5,000 messages and gets 40% opens looks far better than one that sends 50,000 and gets 8%. Volume without engagement is the fastest way to get filtered.
The second signal is complaint rate, and the thresholds are stricter than most marketers realize. Google’s published guidance puts the spam-complaint rate ceiling at 0.1% (one complaint per 1,000 delivered), with 0.3% cited as the level at which filtering becomes aggressive. Microsoft’s JMRP guidance is similar. Track complaints against delivered volume, not sent volume, or you’ll underestimate the rate.
A third, underappreciated signal is infrastructure consistency. Providers watch whether your sending IPs, HELO domains, and From domains stay stable. Migrating mid-warmup, changing your Return-Path, or sending from a shared IP pool one day and a dedicated IP the next resets the clock. Pick your infrastructure and don’t move it for at least 60 days.
Set your baseline before day one
Warmup fails when it starts from a dirty list. Before you send a single message from the new infrastructure, do three things.
Segment by engagement recency. Pull your last 90-day openers and clickers into one segment. Pull 90–180 day openers into a second. Everything older than 180 days goes into a suppression or re-permission bucket and does not touch the new IPs during warmup. If you’re on a new domain with no history, you have no engagement data — in that case, segment by acquisition source and only warm with the sources that historically produced the highest engagement.
Clean the list. Run it through a real-time verification service and remove hard bounces, known spam traps, and role addresses (info@, sales@, admin@). Spam traps are the single most destructive thing you can hit during warmup because they signal list acquisition problems, and providers weight them heavily on new IPs. Expect to lose 5–15% of a list that hasn’t been cleaned in a year.
Fix authentication. SPF, DKIM, and DMARC need to be in place and passing before warmup starts. DMARC at p=none is acceptable for the first two weeks; move to p=quarantine once you’ve confirmed alignment. Gmail and Yahoo both require DMARC for bulk senders as of 2024, and Microsoft has been tightening enforcement. If your DKIM selector is new, publish it at least 48 hours before the first send so DNS propagation settles.
One detail that trips people up: if you’re migrating ESPs, your DKIM keys change. That means your domain’s authentication history resets too, even if the domain itself is old. Treat a domain with new DKIM keys as closer to a new domain than an established one for the first two weeks.
The volume math: how much, how fast
The industry rule of thumb is to double daily volume every 24–48 hours, starting from a few hundred messages. That’s directionally right but too crude for real planning. Here’s the math I use.
Start with your target steady-state daily volume. Call it V. Your warmup should take somewhere between 21 and 30 days to reach V, with the first seven days capped at roughly 2–5% of V per day. That means if you plan to send 200,000 messages per day at steady state, day one is 4,000–10,000 messages, not 500. Sending 500 messages a day for a week on a dedicated IP is a waste of time — the providers won’t see enough volume to form an opinion, and you’ll extend the warmup unnecessarily.
| Day | Daily volume (% of steady state V) | Audience |
|---|---|---|
| 1–3 | 2–3% | 90-day openers only, most recent first |
| 4–7 | 5–8% | 90-day openers, all |
| 8–14 | 10–20% | 90-day openers + 90–180 day openers |
| 15–21 | 25–50% | All 180-day openers, plus recent non-openers |
| 22–30 | 60–100% | Full mailable list |
Two caveats. First, this table assumes a dedicated IP. If you’re on a shared IP pool, the provider’s reputation is pooled across senders, so warmup is less about your IP and more about your domain and your sending patterns. You can move faster, but you’re also exposed to other senders’ behavior — a shared pool is a real trade-off, not a free lunch. Second, if V is under 20,000 per day, you can compress the schedule to 14–18 days. Below about 2,000 per day, dedicated IP warmup is largely pointless; stay on shared infrastructure.
Don’t send the same volume every day. Providers read sudden flatlines as suspicious. Vary volume by 10–20% day to day and avoid the temptation to send everything at 9 a.m. — spread sends across the day, ideally aligned with when your recipients historically engage.
Where to host the site behind the list
None of the tactics in this article matter if the infrastructure behind them is slow,
poorly configured, or capped at a mailbox size you outgrow in a month. If you are setting up
a site, a landing page, or a sending domain from scratch, Hostinger is a sensible starting point: entry plans are cheap enough to
experiment with, the control panel exposes DNS records for SPF, DKIM and DMARC without a support
ticket, and the VPS tier is there when you outgrow shared hosting. What matters more than the
headline price is that it scales predictably as the list grows.
Day-by-day: the first two weeks
Days 1–3: prove you’re real
Send only to your most engaged 90-day segment. Use a plain, transactional-style template — no heavy imagery, no promotional language in the subject line. The goal is not conversions. It’s establishing opens and, ideally, replies. If you have a transactional or notification stream (order confirmations, password resets), route a portion of it through the new IPs during this window. Transactional mail is opened at 60–80% and generates almost no complaints, which is the fastest reputation signal available.
Watch for: delivery rate above 98%, bounce rate under 2%, and any deferrals. Microsoft’s 451 4.7.650 temporary block (“suspicious activity”) is common on new IPs and usually resolves within 24–72 hours if you don’t increase volume.
Days 4–7: extend the engaged segment
Expand to all 90-day openers. If your first three days showed complaint rates above 0.05%, hold at the day-3 volume for another 48 hours before expanding. This is where I see most teams get impatient and blow up the warmup. A complaint rate of 0.1% on day five is a warning, not a green light.
Watch for: Gmail Postmaster Tools domain reputation moving from “Medium” toward “High.” If it drops to “Low,” stop expanding and investigate. Check your seed list results across Gmail, Outlook, Yahoo, and a corporate filter like Proofpoint or Barracuda — corporate filters lag consumer providers by a few days and are often the first to show problems.
Days 8–14: introduce older engagement
Add 90–180 day openers. This is the riskiest expansion because these recipients are less likely to remember you, which raises complaint risk. Send them a re-engagement-flavored message rather than your standard promotional template — it lowers complaints and gives you a clean signal on whether they still want to hear from you.
Watch for: complaint rate crossing 0.08%, hard bounce rate above 3%, or any appearance of your domain in a blocklist like Spamhaus, Barracuda, or SpamCop. A single Spamhaus listing during warmup is recoverable if you act within 24 hours; two in a month is a structural problem.
Days 15–30: scaling to full volume and what to watch
From day 15 you can scale toward full volume, but scale on engagement, not on calendar. The rule I use: only increase volume if the previous 72 hours held complaint rate under 0.08%, bounce rate under 2%, and delivery rate above 97%. If any of those slip, hold volume flat for 72 hours before trying again.
By day 21 you should have enough data to check whether Gmail domain reputation is “High” and Microsoft SNDS shows your IPs in “green” or at worst low-yellow. If Gmail reputation is still “Medium” at day 21, don’t push to 100% volume — extend the ramp by a week and focus on engagement quality (subject line testing, send-time optimization) rather than raw volume.
Three signals deserve daily attention through day 30:
- Gmail Postmaster Tools spam rate. Keep it under 0.1%. The tool reports on a 24-hour rolling basis; a one-day spike is often noise, but two consecutive days above 0.1% requires a hold.
- Microsoft SNDS complaint and trap rates. Microsoft’s trap hits are the most reliable early warning of list quality problems, and they show up faster than Gmail’s.
- Inbox placement from a seed tool. Aim for 90%+ inbox placement at Gmail and Outlook by day 21. If you’re seeing 70% at Outlook while Gmail is at 95%, the problem is usually sending pattern or authentication alignment, not content.
Scenario differences: new domain vs. ESP migration
These two cases are treated as identical in most guides, and they aren’t.
ESP migration keeps your domain’s reputation but resets your IP reputation and your DKIM keys. You can move faster on the domain side because Gmail and Microsoft already have domain-level history. The main risk is that the old ESP is still sending — even a small transactional stream — which creates inconsistent authentication and confuses filters. Shut down the old sending completely on day one, or you’ll spend the warmup chasing phantom deliverability issues.
New domain has no history at all, and you should assume the first 30 days are a probation period. Use a subdomain for marketing (e.g., mail.yourdomain.com) so a warmup problem doesn’t contaminate your transactional or corporate sending. Register the domain at least 30 days before first send if possible — domain age is a weak but real signal, and a domain registered last week sending 10,000 emails looks like a throwaway. If you can’t wait, start at 1% of V, not 3%.
A hybrid case worth naming: migrating to a new ESP and a new domain simultaneously. Don’t. Move one variable at a time. If you must, treat it as a new domain and extend warmup to 45 days.
What to do when it goes wrong
You will hit a block at some point. The response depends on which one.
Microsoft 451 temporary blocks are the most common. They resolve on their own if you reduce volume by 50% and hold for 48–72 hours. Don’t retry aggressively; repeated retries extend the block. Check that your HELO domain matches your sending domain and that your Return-Path is aligned with your DKIM domain.
Gmail spam folder placement with a healthy complaint rate usually points to content or list quality. Run a seed test, check for spammy subject-line patterns, and confirm your DMARC alignment. If placement doesn’t recover within 72 hours at reduced volume, remove your least-engaged 20% of recipients from the sending pool and try again.
Spamhaus listings require a delisting request through their lookup tool. Fix the underlying cause first — usually a spam trap hit from an old list segment — or you’ll be relisted within a week. Document what you found and what you changed; Spamhaus reviewers respond better to specific remediation than to a generic appeal.
One more thing: don’t buy a “warmup service” that sends fake engagement to your IPs. Some providers offer this, and it does move reputation metrics temporarily. It also creates engagement patterns that don’t match your real list, and when the fake traffic stops, the reputation collapses faster than it would have without it. Warm up with real recipients or don’t warm up at all.
Related reading
Conclusion
Warmup is a 21–30 day process with a clear logic: start small with your most engaged recipients, expand only when complaint, bounce, and delivery metrics hold, and read the provider telemetry daily rather than weekly. The volume math matters more than the calendar — 2–3% of steady-state volume on day one, doubling roughly every 48 hours, gated by real signals rather than a fixed schedule. Treat authentication, list hygiene, and infrastructure stability as preconditions, not afterthoughts. And resist the urge to compress the timeline because a campaign is due. A burned IP or domain takes three to six months to recover, which is a far worse outcome than a two-week delay.
Disclosure: this article contains an affiliate link. If you buy through it we may earn a commission at no extra cost to you. We only link to services we would recommend on merit.
Pingback: Moosend vs ConvertKit: Which Is Better for Segmentation Features in 2026? – emailversus.com