Cold Email Domains: How Many You Need, How to Buy and Set Up

Most teams need three to five sending domains to start. The math is simple: two mailboxes per domain, roughly 20 cold emails per mailbox per day once warmed, which puts four domains at about 160 sends a day. Work backwards from your target volume rather than picking a number first.

The reason the answer is a range and not a rule is that daily volume per mailbox is the variable that actually matters, and it depends on how good your list and copy are. A mailbox sending 20 well targeted emails a day to people who reply is safe indefinitely. The same mailbox sending 20 a day to a scraped list with a 0.5% reply rate and rising bounces will be in trouble inside a month. Volume caps are a proxy for engagement, and engagement is what filters actually score.

Here is the sizing that most practitioners converge on. Note that the per mailbox figures sit far below the provider hard limits (Google Workspace allows 2,000 messages a day per user), because provider limits are anti abuse ceilings, not deliverability advice.

Two things fall out of that table. First, the mailbox count is what scales, not the domain count on its own, which is why the question is really how many email accounts you need for cold email. Second, at 500 a day you are running 27 paid seats, and if your sending tool charges per mailbox or per email sent, the tool starts costing more than the infrastructure. ColdMailer is $49 a month flat whether you connect four mailboxes or fifty one. For the full picture of what sits underneath, see our guide to cold email infrastructure.

Your primary domain carries mail that the business cannot afford to lose: invoices, contracts, password resets, support tickets, replies to warm leads. Cold email, by definition, goes to people who did not ask for it, and a meaningful share of them will mark it as spam. That complaint rate attaches to the domain, not to the campaign.

What happens next is quiet and expensive. Nobody sends you an alert. Your invoices start landing in promotions folders. A support reply to an angry customer sits in their junk for two days. Your sales rep's follow up to a deal in procurement never arrives, and you find out when the deal goes cold. By the time you diagnose it, you have weeks of reputation repair ahead of you and no way to stop sending from the domain, because it is the domain your company runs on.

A secondary sending domain makes that failure survivable. If tryacme.co gets flagged, you pause it, retire it, spin up a replacement and keep operating. The cost of that insurance is about $11 a year per domain plus the mailbox seats. There is no scenario where that is a bad trade. This is the same logic behind a dedicated sending domain for any high volume mail stream, and it is the first thing to fix if you are working through cold email deliverability problems.

Use a separate domain if you are sending more than about 50 cold emails a day. Filters treat a subdomain as part of its parent registrable domain, so reputation damage on mail.yourbrand.com can reach yourbrand.com. Subdomains cost nothing and inherit some parent trust, which makes them reasonable for low volume or highly targeted outreach only.

The subdomain case is stronger than most cold email vendors admit. A subdomain launches immediately, needs no redirect, and puts your actual brand name in the from address, which helps recognition and reply rates. It also inherits the parent domain's registration age, and domain age is a real signal that new lookalike domains simply do not have on day one.

The separate domain case wins on the thing that matters at scale: containment. Once you are sending hundreds of cold emails a day, the probability of a reputation incident is not small, it is eventual. Separate domains mean the incident costs you $11 and a week of setup. A subdomain incident costs you your company mail. Many teams run both, keeping a subdomain for semi warm outreach to people who have touched the brand, and lookalike domains for genuine cold lists.

Warm a brand new cold email domain for at least three to four weeks before it touches a real prospect, and keep warm up running underneath campaigns permanently after that. Newly registered domains have zero sending history, so providers apply extra scrutiny. A domain migrated from existing use with prior good history can often ramp in one to two weeks.

The mechanism is worth understanding, because it changes how you run it. Warm up works by generating positive engagement signals at a volume that looks organic: messages sent, opened, replied to, moved out of spam and marked important, on a curve that rises gradually rather than jumping. Filters build a model of what normal looks like for your domain. Warm up is how you tell them that normal includes outbound mail that people answer.

A workable schedule per mailbox looks like this. Week one, warm up only, five to ten messages a day. Week two, 15 to 20 warm up messages a day, still no campaigns. Week three, 25 to 30 warm up, and you can begin five real sends a day. Week four onward, hold warm up steady and climb real sends by about five a day each week until you reach 20. The failure mode is always the same: someone finishes two weeks and immediately sends 200. The step change is exactly what the filters are watching for. Our email warmup software automates the curve, and how to warm up an email domain walks through the full schedule.

The goal is a domain a prospect reads as legitimately yours. Good patterns, roughly in order of how well they age:

• Prefix verbs: get, try, use, join. getacme.com, tryacme.com, useacme.com. These read as normal SaaS marketing domains because they are.
• Suffix nouns: hq, app, team, group, io. acmehq.com, acmeteam.com. Slightly more corporate, works well for services businesses.
• Alternate TLDs on the exact brand: acme.co, acme.io, acme.net. Cleanest option if the names are available and not owned by someone else in your space.
• Hyphenated brand plus category: acme-sales.com. Functional, slightly weaker visually, fine at volume.

What to avoid: novelty and bulk registered TLDs (.xyz, .top, .click, .buzz) which carry poor default reputation because of how they are abused; anything that could be read as a typo of a well known brand, which invites both filtering and legal trouble; numbers and doubled letters (acmee.com, acme2.com) which read as phishing; and generic words with no connection to your company, which give a curious prospect nothing to verify.

Buy them all at the same registrar, on the same account, with WHOIS privacy on, and set auto renew. An outreach domain that lapses mid campaign takes its warm up history with it. Register a couple more than you need at the outset, because the price of a spare domain is trivial next to the delay of warming a replacement from scratch when one goes bad.

Domains are the cheap part and mailboxes are the recurring part. Here is a real 200 a day setup: five domains, two mailboxes each, ten seats total. Prices below are the ones we could verify on the vendors' own pages, and your registrar and plan may differ.

Microsoft 365 Business Basic lists at the same $7.00 per user per month on an annual commitment, so the mailbox line is comparable either way. Microsoft does not publish a stable monthly (non annual) per user figure, so we do not quote one. The tradeoffs between the two providers for outreach specifically are covered in Google Workspace vs Microsoft 365 for cold email.

The number that moves when you scale is the mailbox line. Doubling to 400 a day means 20 seats, so mailboxes go to $1,680 a year while domains go to about $111 and ColdMailer stays at $588. Tools that charge per mailbox connected or per email sent add a third line that grows with the second one, which is the specific cost curve the flat plan exists to flatten.

Every domain in the pool needs the same records, and the most common cause of a new domain underperforming is that one of the five got skipped during setup. Work through this per domain, not per campaign.

• SPF: one TXT record, one SPF record per domain, listing only the services that actually send. Stay under the 10 DNS lookup limit or the record fails open and you lose the check entirely.
• DKIM: generate the key in your mail provider's admin console and publish the TXT record at the selector host. Then verify it is actually signing outbound mail, because a published key with signing disabled passes nothing.
• DMARC: start at p=none with rua reporting so you can see what is authenticating. Read the reports for a couple of weeks, fix what fails, then move to p=quarantine. Google and Microsoft's bulk sender rules require a DMARC record for anyone sending 5,000 or more a day to Gmail.
• MX: pointed at your mail provider, correct priorities, no leftovers from a previous host.
• Custom tracking domain: a CNAME so open and click links use your domain rather than a shared tracking host that thousands of other senders have already burned.
• Redirect: 301 from the sending domain to your main site, so the domain resolves to something real.

Once the records are live, send a test to a seed address and read the raw headers: SPF pass, DKIM pass, DMARC pass, aligned. Anything short of that is a setup bug, not a deliverability mystery. The full walkthrough with record examples is in SPF, DKIM and DMARC setup for cold email.

Owning five domains does nothing on its own. The value shows up when a campaign's volume is spread across every mailbox on every domain so that no address carries an unnatural load. That is inbox rotation, and it is the operational payoff for all the setup work above.

Done well, rotation also varies the send timing. Ten mailboxes firing 20 messages each at exactly 9:00am is ten mailboxes with an identical, machine shaped pattern. Spread across business hours with randomized gaps, the same volume reads as ten people working through their morning. ColdMailer handles the distribution and pacing across every connected SMTP account, and because there is no per mailbox charge, adding a sixth domain to relieve pressure on the other five costs you the domain registration and the seats, nothing more.

Watch the pool, not just the campaign. Reply rate per domain is the earliest honest signal you have. If four domains hold at 6% and the fifth drops to 1%, that fifth domain has a problem no aggregate metric will show you. Pause it, run warm up only for two weeks, check its authentication, and bring it back slowly. If it does not recover, retire it. That is what the spares are for. Connecting the mailboxes themselves is standard SMTP email sending, so any provider that gives you credentials will work.