One of the most common questions in cold email communities, and the one people get wrong most often: how many mailboxes should you put on one domain? I’ve set up cold email infrastructure for our clients at KomsGro, watched a Reddit thread where an agency owner lost three Google Workspace accounts in a single month from doing this wrong, and re-built our own setup twice. Here’s the math that actually holds up in 2026.
The short answer: 2 to 3 mailboxes per domain
The consensus across practitioners and tool guides, and what has survived every test I’ve thrown at it: two to three mailboxes per domain is the sweet spot. One mailbox per domain works too, and is even safer, just more expensive. Beyond three or four on a single fresh domain, you’re concentrating risk: if that domain gets flagged, every mailbox on it burns at once.
The math people miss: the per-mailbox sending cap is the real constraint, not the domain count. The widely accepted ceiling is around 50 emails per mailbox per day once warmed. So if you want to send 1,500 cold emails a day, the honest calculation is:
1,500 sends / 50 per mailbox = 30 mailboxes / 3 per domain = about 10 domains.
That’s the scale where people get burned. The Reddit thread I mentioned above (“how I send 3,700+ cold emails per day”) is the extreme end of this math, and it works only with serious infrastructure discipline. Most teams don’t need that volume; most teams need 100 to 300 emails a day, which is 2 to 6 mailboxes across 2 to 3 domains.
Why the domain cap matters (the part everyone learns the hard way)
Google and Microsoft watch behavior per domain. When one mailbox on a domain starts blasting unverified contacts or getting marked spam, the domain’s reputation drops, and every mailbox on it suffers together. There’s no per-mailbox reputation firewall inside a domain.
That’s why the “burning through domains faster than they can warm them” complaint keeps showing up in agency threads. The usual cause isn’t volume, it’s one of these three:
- Skipping the warmup. A new domain that starts sending 50 cold emails a day from day one is flagged almost immediately. Warmup takes 2 to 3 weeks minimum, longer if you’re cautious.
- Using the main company domain. Your primary domain’s reputation carries your invoices, your support replies, everything. Never point cold email at it. Buy secondary domains (komsgro.com is ours; we send from variations of it) and let them take the risk.
- No list verification. Bounce rates above 2 to 3 percent will torch a domain’s reputation quickly. Every list gets verified before a single send.
When the math changes: what actually burns domains (from the field)
Since the table above is theory until tested, here are the real-world failure cases behind the “burning through domains” complaints, and how the math prevents each:
- The 10-mailbox domain. An agency moved from 5 to 12 mailboxes on one domain to hit a client volume target. Within a week, deliverability dashboards showed spam placement across all 12. The domain was the single point of failure; splitting to 4 domains of 3 would have contained it.
- The warmup skip. A team launched with fresh domains on a Monday deadline and skipped warmup entirely. Google flagged the domain within days. Three weeks of warmup would have prevented a month of rebuilding.
- The shared-fate mistake. Campaign mailboxes and the company newsletter mailbox on the same secondary domain. The newsletter’s complaint rate (normal for marketing email) dragged the cold mail down with it. Separate the purposes: campaigns on their own domains, marketing email on your main one, never married.
- The verification skip. One client imported a list from 2019 “just to start fast”. 12 percent bounces on day one, and the domain spent a month recovering. Verification is a $30 a month step that protects a $100 infrastructure.
Each of these is a reason the safe column (one mailbox per domain) exists. If your program is mission-critical, that column is the right answer despite the cost.
The scaling plan I actually use
Here’s the ramp, which lines up with what practitioners on the cold email subreddits are running in 2026:
- Weeks 1 to 3: warmup only. Mailboxes send internal warmup traffic, slowly increasing. No campaigns.
- Week 4: start campaigns at 20 to 30 emails per mailbox per day. Watch bounces and spam placement daily.
- Week 5 onward: ramp to the 50-per-mailbox ceiling if metrics are clean.
- Adding capacity: when you need more volume, add mailboxes to a NEW domain, warm it for 2 to 3 weeks, then bring it online. Never add volume to an existing warmed domain beyond its ceiling.
A concrete example from our stack: a client sending 300 emails a day runs 10 mailboxes across 4 to 5 secondary domains. Each mailbox holds under the ceiling, each domain carries 2 to 3 mailboxes, and if one domain’s reputation dips, 80 percent of the program keeps sending while we fix it.
How many domains do you actually need? (the planning table)
Work backwards from your meeting target. The math that holds up in practice:
| Daily cold email target | Mailboxes needed (at 50/day) | Domains (at 3 per) | Domains (at 1 per, safest) |
|---|---|---|---|
| 100/day | 2 | 1 to 2 | 2 |
| 300/day | 6 | 2 to 3 | 6 |
| 500/day | 10 | 4 to 5 | 10 |
| 1,000/day | 20 | 7 to 10 | 20 |
The “safest” column is what I’d recommend for anyone who’s been burned before: one mailbox per domain costs more per unit, but a single reputation incident costs one mailbox instead of the whole fleet.
Domain choice details that matter more than people think
- Buy domains that look like your brand, not like spam. komsgro-outbound.com beats k0m5gr0-biz.net every time. Old, typo-squat-looking domains hurt deliverability signals.
- Register for 2+ years where you can. Long registration periods are a (weak but real) trust signal.
- Authenticate everything. SPF, DKIM and DMARC on every sending domain. Not optional in 2026.
- Redirect the secondary domain to your main site. Prospects who check who you are should land on your real website.
- Don’t mix warm mailboxes and campaign mailboxes on the same mailbox. Warmup tools keep a baseline of activity running; campaigns ride on top, never replace it.
What this costs at real scale
Since someone always asks: at 300 emails a day (6 mailboxes, 3 domains), the infrastructure runs roughly $50 to $120 a month for domains, mailboxes and warmup tooling combined. The mailbox providers charge $3 to $10 per mailbox monthly, domains run $10 to $20 a year each, and warmup rides on your sending tool (Instantly and Smartlead both include it). The infrastructure is cheap; the discipline is the expensive part.
The mistake hierarchy (from most to least common)
- Sending campaigns from the main domain.
- Skipping warmup or cutting it to one week.
- One domain carrying ten mailboxes.
- Unverified lists.
- Ramp speeds that look like a spam signal.
If you fix those five, the mailbox-per-domain question answers itself: as many domains as your volume requires, 2 to 3 mailboxes on each, all warmed, all verified, all under the ceiling.
This whole setup is part of the outbound marketing system we run at KomsGro, and the sending limits behind it are covered in detail in how many cold emails to send per day if you want the volume math next. Or skip the trial and error entirely and have us build the plumbing.
Quick answers
Can I use my main company domain if I keep the volume low? You can, and people do, but I don’t recommend it for true cold outreach. The risk isn’t just deliverability: complaint rates on cold campaigns are inherently higher, and your main domain’s reputation affects every business email you send. Secondary domains exist precisely to isolate that risk for $10 to $20 a year.
Do warmup mailboxes count against the cap? No. Warmup traffic (the tool’s internal ramp) sits under the campaign ceiling, not on top of it. But keep campaign volume under 50 per mailbox regardless of what warmup is doing.
Is one mailbox per domain overkill for a small team? If you’re sending 100 emails a day, two mailboxes on two domains is the safe answer and costs maybe $40 extra a year. The insurance framing helps: each extra domain is an insurance policy against a shared-fate reputation event.
What about buying pre-warmed accounts? I don’t recommend it. The history is unverifyable, the accounts are frequently recycled from spammed pools, and you inherit whatever the seller did. Fresh domains plus your own warmup take three weeks and give you a known-clean start.
The verification step before any of this math
Every number above assumes one input: a verified list. Unverified lists are the reason the mailbox math fails in practice. A 10 percent bounce rate on day one doesn’t just look bad in the dashboard; it teaches mailbox providers that your domain sends to bad addresses, which is one of the strongest spam signals that exists.
The verification habit that matches the infrastructure discipline: every list verified before it’s loaded, every time, no exceptions for “small test batches”. Verification tools run a few cents per contact at volume, and the cheapest verification is cheaper than one burned domain plus three weeks of re-warming.