There is a quiet truth about cold outreach that most playbooks skip: the difference between a campaign that books meetings and a campaign that lands in spam often has nothing to do with copy. It has to do with how the sending mailbox was warmed up before the first prospect ever heard from you.
A poorly warmed mailbox kills a great campaign. A well-warmed mailbox can rescue a mediocre one. This post is the operational playbook we use internally to warm 30 plus mailboxes across 8 to 12 sending domains without triggering throttling, blacklists, or sudden inbox-placement collapses.
Most warmup explanations start with "you need to build sender reputation," then trail off into vague language. The mechanism deserves a more concrete description.
Mailbox providers (Gmail, Microsoft 365, Yahoo, Proofpoint, Mimecast) score every sender against a set of behavioral signals: open rate, reply rate, archive-without-read rate, spam-flag rate, deletion velocity, and folder placement actions taken by recipients. New domains start neutral. Every send updates that score.
Warmup is the process of generating positive engagement signals on a new mailbox before you start sending real cold outreach. The goal is to teach the providers that your mailbox sends emails that humans engage with positively. Once that pattern is established, real cold sends inherit that reputation.
Skip warmup and your first 200 cold sends train the providers that your mailbox sends unread, unengaged, low-quality mail. From that point you are climbing out of a hole that takes weeks to escape, if you can escape at all.
Our cold email deliverability guide covers the SPF, DKIM, and DMARC fundamentals that warmup assumes are already in place. If those are not configured, warmup will not save you.
This is the day-by-day volume schedule we use for new mailboxes on freshly aged domains. It assumes the domain is at least 14 days old (we recommend 21 to 30 days) and that DNS authentication is fully set up before day 1.
Days 1 to 3: 8 to 12 emails per day per mailbox, all warmup-pool traffic, 100% inbound and outbound between trusted warmup peers.
Days 4 to 7: 15 to 22 emails per day. Still 100% warmup pool. Reply rates within the pool should be 30% or higher; if not, your warmup pool quality is poor and needs replacement.
Days 8 to 14: 25 to 40 emails per day. Begin introducing your real cold outreach at 5% to 10% of total daily volume. So if the mailbox sends 35 emails per day, 3 to 4 are cold, the rest are warmup pool.
Days 15 to 21: 40 to 55 emails per day. Cold ratio rises to 20% to 30%. Watch open rates on your cold sends; they should be at or above 35% during this phase. If they drop below 28%, pause and extend warmup by 5 days.
Days 22 to 28: 55 to 75 emails per day. Cold ratio at 50% to 60%. By the end of this phase, your mailbox should be sustainably handling its full target volume.
Day 29 onward: 75 to 110 emails per day for sustained operations. We do not recommend exceeding 100 per mailbox per day for cold outreach regardless of how warmed the mailbox is. Scaling happens through additional mailboxes, not heroic per-mailbox volume.
This schedule is conservative. Aggressive teams compress this to 18 days. We have seen the data on what happens to those mailboxes at week six. The compression does not pay off.
The volume schedule above only works if your warmup pool is healthy. A warmup pool is the network of other mailboxes that your warming mailbox exchanges traffic with during the warmup window.
Here is what most teams miss: not all warmup pools are equal. A pool full of other freshly warmed cold-outreach mailboxes is essentially a closed loop of low-reputation traffic. The providers see through this faster than you would expect.
A high-quality warmup pool has three characteristics:
1. Diversity of providers: At least 35% of pool participants should be on Microsoft 365, at least 25% on Google Workspace, with the remainder split across Yahoo, Proton, Fastmail, and self-hosted servers. A pool that is 90% Gmail-to-Gmail trains a narrow signal that does not generalize.
2. Mailbox age distribution: A healthy pool mixes new mailboxes (in warmup) with established mailboxes (in steady state). If every pool member is brand new, you are not getting reputation transfer; you are getting noise.
3. Realistic reply patterns: The pool should not show suspiciously perfect reply rates. Real human inbox behavior includes some deletes, some archives without reading, some delayed replies. Warmup pools that auto-reply at 100% are an obvious red flag to any modern filter.
OnyxSend's warmup engine maintains a managed pool with these characteristics so customers do not have to manage pool quality manually. If you are running warmup yourself, audit the pool quarterly.
A common mistake is putting too many sending mailboxes on the same domain. There is a per-domain reputation ceiling, and stacking mailboxes accelerates the rate at which you hit it.
Our recommended topology for teams sending 1,500 to 4,000 cold emails per day:
- 6 to 10 sending domains, each aged 21+ days before activation - 3 to 4 mailboxes per domain - 70 to 90 cold sends per mailbox per day at steady state - One primary corporate domain (your main brand) that you do NOT use for cold outreach
Keep the cold-outreach domains separate from your corporate domain. Your team's day-to-day email reputation should not be tied to outbound prospecting reputation. If outbound goes wrong, you do not want sales calls and customer support emails landing in spam too.
For naming the cold-outreach domains, use variants that are clearly related to your brand but distinct. If your corporate domain is acmesoftware.com, use try-acme.com or get-acmesoftware.com. Do not register acmesoftvvare.com or other lookalike domains. Lookalikes trigger filter heuristics and get flagged as phishing.
During warmup and early operations, monitor these signals. Any of them should trigger an immediate volume reduction or pause.
Bounce rate above 3%: A single bounce here and there is normal. A sudden spike means your list quality is bad or your authentication is breaking. Pause immediately.
Spam complaint rate above 0.08%: This is the threshold where Microsoft and Google start applying reputation penalties. Even one complaint per 1,200 sends is too high.
Open rate dropping more than 12 percentage points week-over-week: Inbox placement is shifting. Pause and investigate before it cascades.
Reply rate dropping by more than 30% week-over-week without copy changes: Same diagnosis. Something is changing in how providers route your mail.
Sudden increase in archive-without-read: This is a leading indicator that a placement shift has happened, even if open rate has not yet caught up. Recipients are seeing your subject line in the promotions or updates folder and clearing it without engaging.
Blacklist hits on Spamhaus, SORBS, or Barracuda: Pause sending from the affected IP or domain immediately and investigate. Do not try to send through it.
A good outreach platform monitors these signals automatically and throttles. If you are running warmup manually, you need a daily dashboard that surfaces them. The cost of a missed signal compounds quickly.
Sometimes warmup goes wrong. Maybe a poorly written first campaign generated 4 spam complaints on day 8. Maybe a list import had a 9% bounce rate. The mailbox is now hurting your reputation.
The recovery protocol:
1. Pause cold sending entirely for 14 days. 2. Reduce warmup pool traffic to 8 to 10 emails per day for the first 7 days, then 15 to 20 per day for the next 7 days. 3. Send a small volume (10 to 15 per day) of high-engagement mail to known warm contacts (your team, friends, opt-in subscribers) for 5 days. 4. After day 21, resume warmup pool only at 25 to 35 per day for another 7 days. 5. After day 28, attempt cold sending at 5 to 8 emails per day, gradually scaling.
Recovery is slower than initial warmup. If the damage is severe (bounce rate over 8% on a single day, blacklist hit, multiple complaint signals), the mailbox is rarely worth recovering. Retire it, register a fresh domain, and start fresh.
If you are running 30 plus mailboxes manually, the operational cost of warmup management exceeds the cost of automation. The math is straightforward.
Our platform handles the entire warmup lifecycle: pool participation, daily volume scheduling, provider distribution, reply pattern simulation, and the daily monitoring of bounce, complaint, and engagement signals. When a mailbox enters a degraded state, the system throttles automatically and surfaces an alert before damage compounds.
This integrates with the rest of the OnyxSend stack: ICP scoring, sequence execution, reply classification, and the sequence framework we recommend for new senders. Mailboxes flow from warmup into production sending without manual handoff.
For teams looking to scale beyond 1,000 cold sends per day, automated warmup is no longer optional. Manual warmup at that scale fails predictably. The teams that figure this out early run smoother operations and book more meetings; the teams that learn it the hard way burn through 3 to 4 sets of domains before they accept the discipline.
Warmup is not glamorous work. It does not show up in pipeline dashboards. But it is the foundation that everything else stands on. A team running well-warmed mailboxes against a good ICP with a disciplined sequence will outperform a team with better copy and worse infrastructure every single time.
If you want to see how OnyxSend handles warmup, sequencing, and deliverability monitoring as a single integrated workflow, start a 14-day trial. The first warmup batch starts within 30 minutes of signup.
<!-- onyxsend:related-reading -->
- Deliverability Monitoring for Cold Email: The 8 Signals to Track Daily - Multichannel Outbound: The Email + LinkedIn Cadence That Actually Works - The B2B Automated Prospecting Stack: From Cold List to Booked Meeting - Top 7 Instantly Alternatives for Cold Email in 2026 - Cold Email Automation: The Modern SDR Replacement - How to Write Cold Emails That Actually Get Replies - B2B Cold Email Subject Lines: 35 Templates That Drive Opens in 2026 - OnyxSend cold outreach services - OnyxSend pricing - OnyxSend case studies
<!-- /onyxsend:related-reading -->