How Many Inboxes Per Domain for Cold Email? The Answer

Stack 20 inboxes on one domain and your reply rates collapse. Every mailbox on a domain inherits the same reputation score with each ESP, and once the cluster looks like a sending farm to filters, every inbox on that domain gets downgraded together.

The real answer to how many inboxes per domain for cold email depends on which provider you're sending through. Google Workspace, Microsoft 365, and SMTP each ship with different ceilings, and the right number balances per-domain volume against the reputation surface you've actually built.

This post breaks down the working numbers, why the ceilings exist, and how to scale cold email volume without burning the domains you spent money on.

Key Takeaways

The right number depends on your provider: Google Workspace runs 2 inboxes per domain. Microsoft 365 self-managed runs 2-3. SMTP runs 2-3. Microsoft 365 configured the way ScaledMail provisions it runs up to 25. Pick by stack, not by guess.
Never send cold email from your primary business domain, and skip subdomains too: Buy separate lookalike domains for outbound. Subdomains share too much reputation with the parent, so a burned subdomain drags your main domain down with it.
Scale by adding domains, not by stuffing more inboxes onto a domain you already have: Spread sending across more reputation surface, keep each inbox at a normal-business 5-25 cold per day, and rotate copy structurally before reply rates degrade.

How Many Inboxes Should You Use Per Domain?

The answer changes based on the provider you send through. ESPs do not all weigh "how many mailboxes on this domain" the same way, and they do not all permit the same per-inbox sending ceiling.

Working numbers from running outbound across hundreds of clients:

Google Workspace: 2 inboxes per domain, 15-25 cold sends per inbox per day.
Microsoft 365 (self-managed): 2-3 inboxes per domain, 15-25 cold sends per inbox per day.
Microsoft 365 (ScaledMail-provisioned): up to 25 inboxes per domain, 5-10 cold sends per inbox per day.
SMTP: 2-3 inboxes per domain, 5-10 cold sends per inbox per day.

Two equations matter for capacity planning. Total per-day per inbox includes warmup, not only cold sends. And the inboxes-per-domain ceiling is set by what each provider tolerates before it starts treating the domain as a sending farm.

The Default for Most Operators: 2-3

Running cold email yourself on Google Workspace or self-managed Microsoft 365? 2-3 inboxes per domain is the right setup. It pushes meaningful volume from each domain and keeps the sending pattern looking like a normal small business with a couple of working mailboxes.

5 inboxes on a brand-new domain reads as a sending operation to ESP filters. 20 reads as obvious cold outbound. The whole point of the multi-domain strategy is spreading sending across many domains running 2-3 inboxes each, not stuffing one.

Why Microsoft 365 with ScaledMail Goes Up to 25

ScaledMail-provisioned Microsoft 365 tenants support up to 25 inboxes per domain because of how the tenants are configured and how IPs rotate. Each inbox draws less individual reputation pressure, so per-inbox cold sends are lower (5-10 vs 15-25 on Google) but total volume per domain ends up similar, distributed across more sending identities.

Don't try this configuration on a vanilla Microsoft 365 tenant you spun up yourself. The ceiling there is 2-3, same as Google Workspace.

What Drives the Right Number for You

Three variables decide it:

The provider you're sending through. See the numbers above.
Domain age and reputation history. A domain you bought yesterday has zero reputation. Start with one inbox running warmup-only volume for 14 days, then add the second inbox.
Total daily cold volume you need. Need 1,000 cold per day on Google Workspace? That isn't "more inboxes per domain." That's about 50 inboxes spread across 25 domains, plus a 20-30% buffer, so you're buying 30-32 domains.

When you need more volume, the move is more domains, not more inboxes per domain.

The Risk of Too Many Inboxes on One Domain

ESP filters cluster sending behavior at the domain level. Every mailbox on a domain shares the domain's reputation score with each ESP. Stack 20 mailboxes on a brand-new domain, and the cluster looks like a sending farm regardless of how clean any individual mailbox is.

Deliverability Falls First, Then the Domain Burns

Two things go wrong when you over-stack a domain.

The first is gradual. Reply rate slowly drops over a few weeks. Inboxes start landing in Promotions, then in spam. That's domain reputation burn from cold-volume accumulation.

The second is fast. Reply rate halves overnight, or drops 90% in a day or two. That's copy fingerprinting. The cluster of inboxes is running similar sequences, ESPs fingerprint the structural signature, and the whole batch downgrades together. The fix for fingerprinting is new copy. The fix for the gradual burn is new domains.

Either way, the more inboxes on the domain, the bigger the surface for filters to fingerprint and downgrade.

One Burned Mailbox Burns the Domain

A domain has a single reputation score with each ESP. Every mailbox on the domain inherits it. So one mailbox sending burned copy contaminates every other mailbox on the same domain.

The diagnostic move: pick one mailbox per domain, run a placement test, and if it lands in spam, treat the entire domain as done. Don't try to save individual mailboxes. Cancel the mailbox subscription first, then delete from the sequencer. Replace from your warm-up pool.

Sender Names and Signatures Matter as Much as Inbox Count

Even at 2-3 inboxes per domain, two common mistakes burn the domain anyway:

Same sender name across multiple mailboxes. Filters cluster these instantly. Each mailbox gets a unique first-name plus last-name combination. 500 mailboxes means 500 unique sender names. No exceptions.
Hyperlinked company name in the signature. Don't put your company URL in the signature at all. Gmail, Outlook, and Apple Mail auto-link any .com string and treat it as a tracking link, which is a known spam trigger at volume. Use the brand name as a plain word, drop the .com.

For more on what kills domains and how to avoid it, see Cold Email Blacklist Prevention: A Step-by-Step Guide.

How Domain Age Changes Your Setup

A domain bought last week has no reputation. ESPs have no history to weigh, so a brand-new domain sending 50 emails on day one looks identical to a throwaway sending farm.

A domain that's been sending for 90+ days at consistent volume with positive engagement has built reputation, and ESPs give it more leeway.

New Domains: 1 Inbox, Warmup Only, for 14 Days

Buy the domain. Configure DNS authentication (SPF, DKIM, DMARC). Provision one inbox. Run warmup at 1:1 ratio inside your sequencer for 14 days minimum before any cold send.

Warmup is not a phase that ends. It runs continuously alongside cold sends for the active life of every mailbox at a 1:1 ratio (one warmup send for every cold send). The 14-day pre-cold window is the floor, not the entire plan. For a deeper breakdown, see our roundup of the 6 best cold email account warm-up services.

After 14 days, add the second inbox if you're on Google Workspace or self-managed M365. Same warmup process, same 1:1 ratio.

Established Domains: Run the Provider Ceiling

Past 60-90 days of clean sending, run the full provider ceiling: 2 inboxes on Google Workspace, 2-3 on self-managed M365, up to 25 on ScaledMail-provisioned M365.

Don't push past the ceiling because the domain is "working." That's how operators end up at 5+ inboxes on a Google domain and burn the entire batch in a month.

The Reputation Curve Plateaus

A domain's reputation curve is not linear. It builds quickly in the first 30-60 days, plateaus, and then degrades from there as cold volume accumulates. Most working domains last 12-18 months at high engagement. Low-engagement clients replace domains every 5 months. Replacing every 1-2 months means something else is broken, usually copy or list quality, not the domain itself.

Set Up Your Inboxes for Maximum Deliverability

The number of inboxes per domain is one variable. The technical setup decides whether any of them work at all.

Use Separate Domains, Not Subdomains

The advice you'll see in older guides ("use a subdomain like connect.yourcompany.com to protect your main domain") does not hold up at meaningful volume. Subdomains share too much reputation with the parent domain. A burned subdomain drags the parent's deliverability down with it.

The actual move: buy separate lookalike domains. If your main domain is acmecorp.com, your sending domains are things like acme-co.com, tryacme.io, acmehq.com, acmemail.co. Each one is its own reputation surface, fully insulated from your main business domain.

Naming rules: no hyphens, no numbers, short, brandable. .com is the default, .co and .net are solid alternatives. Avoid .ai, .xyz, country codes, and anything that looks low-trust to a recipient.

Buy each domain for one year. You'll burn and replace within 12 months anyway.

Configure SPF, DKIM, and DMARC on Every Domain

Three DNS records, all required, all non-negotiable for cold email deliverability:

SPF (Sender Policy Framework) lists which servers are allowed to send mail on the domain's behalf. Without it, every receiving server treats the domain as suspicious.
DKIM (DomainKeys Identified Mail) adds a cryptographic signature to each outbound email so the receiver can verify it actually came from the domain and was not modified in transit.
DMARC tells receivers what to do when SPF or DKIM fails (reject, quarantine, or none) and provides reporting on who's sending mail in the domain's name.

Google Workspace and Microsoft 365 ship with most of this configured at the tenant level. SMTP setups have to do it all manually. Plan for 24-48 hours of DNS propagation before testing.

Run Warmup Continuously at 1:1

For every cold send, your inbox should also send a warmup send. The 1:1 ratio is the safe default. Defer to your warmup provider's recommended ratio if it specifies a different one, since Smartlead, Instantly, EmailBison, and PlusVibe each tune their warmup networks differently.

Warmup runs inside your sequencer. It is not a separate tool. The point is positive engagement at the ESP level (opens, replies, marks-as-not-spam from real seed mailboxes) burying the negative signals from cold outbound.

Warmup is preventative. It cannot rescue a mailbox that is already in spam. Run it on a brand-new mailbox before first send, and keep it running as a continuous cooling-off layer alongside campaigns.

Set Per-Inbox Daily Limits

The numbers: 15-25 cold per inbox per day on Google Workspace and self-managed M365, 5-10 on ScaledMail-provisioned M365 and SMTP. That's the cold ceiling. Total daily volume per inbox is roughly double once warmup is layered in.

Don't push past the ceiling because the inbox seems healthy. The cost of going over is the inbox burns 3-6 months sooner. The benefit is you sent a few extra emails this week. Bad trade.

Best Practices for High Cold Email Deliverability

The technical setup is the floor. What you actually send through it is the ceiling.

Personalize Beyond First-Name Tokens

Generic personalization ("Hi {first_name}, I saw you're VP of Marketing at {company_name}") is the same as no personalization. Filters and recipients both notice. The bar is one piece of research per prospect that ties to a real business problem they have.

Useful personalization: a recent funding round, a job posting that signals the gap your offer fills, a stack indicator from BuiltWith, a LinkedIn post about a specific operational pain. The unifying rule: personalization has to connect to why your offer is relevant to them. If it doesn't, skip personalization for that send entirely. Irrelevant personalization is worse than none.

Turn Off Open Tracking. Don't Just Ignore the Data.

Open tracking is broken. Apple Mail Privacy Protection auto-opens every email and downloads the tracking pixel regardless of whether the recipient looked at the message. You can run an 80% open-rate campaign that gets zero replies because nobody actually saw it.

The bigger problem: the tracking pixel itself is a deliverability liability. ESPs detect the pixel and downgrade emails that carry it. Click tracking is even worse, since it requires the same redirect mechanism that gets domains blacklisted at scale.

Turn open tracking and click tracking off in your sequencer settings. Measure reply rate instead. A reply means the email reached the inbox and the recipient cared enough to respond. That's the only metric that combines deliverability and conversion in one signal.

For a fuller deliverability checklist, see 8 Steps for Cold Email Deliverability Improvement.

Realistic Reply Rate Targets

1.5-3% overall reply rate is healthy.
4%+ is exceptional.
Below 1% means something is broken. Fix deliverability first, then targeting, then copy.
Of replies, 10-30% should be positive interest. Below 10% means targeting or copy is off. Above 30% means scale up.

Anyone promising 15% reply rates is selling something. Ignore.

Keep Bounce Rate Under 1.5%

A high bounce rate signals to ESPs that your list quality is poor, which downgrades sending reputation. Run every list through validation before sending. Pair LeadMagic with BounceBan or Scrubby to catch invalid emails and dead catch-all domains. Under 1.5% bounce rate is the working target.

Stay Compliant

CAN-SPAM in the US, GDPR in Europe, plus equivalents in other jurisdictions. The non-negotiable bits: accurate subject lines, a real physical address in the email, a clear unsubscribe link, and you honor unsubscribes. None of this is optional, regardless of what your sequencer's default settings say.

How to Scale Cold Email Volume Safely

The wrong way to scale: stuff more inboxes onto your existing domains, push each inbox closer to the ceiling, send the same copy at higher volume.

The right way: add more domains, keep each inbox running the same per-day cold volume, rotate copy structurally before reply rates degrade.

Plan Volume by Domain Count, Not Inbox Count

Need 500 cold sends per day on Google Workspace? The math:

500 cold/day divided by 20 per inbox per day = 25 inboxes.
25 inboxes divided by 2 inboxes per domain = 13 domains.
Plus 20-30% buffer for warmup and rotation = 16-18 domains.

Buying 18 domains feels like a lot. Stuffing 12-13 inboxes onto 2 domains feels efficient. The first one works for 12 months. The second one is in spam by month two.

For a more detailed scaling playbook, see How to Master Automated Cold Email Scaling.

Rotate Domain Batches in 3-Week Windows

Best practice at scale: buy 3x the infrastructure you need at peak volume. One batch is active sending, one is in warmup, one is cooling off or replacing. Run the active batch for about 3 weeks, swap it into warmup, rotate the next batch in.

Most operators wait too long to rotate. They keep sending until reply rates have already collapsed, and by then the mailbox is dead. Rotate before the burn, not after.

Watch Reply Rate by Mailbox Cohort, Weekly

Reply rate dropping more than 25% week-over-week with the same ICP and copy is your earliest signal that something is breaking. Pause, swap copy structurally (not just spintax), and resume.

Don't wait to confirm a "trend." By the time you can confirm it, the mailboxes are gone.

Watch the Spam Complaint Rate

Keep complaint rate below 0.1%. Going over is a major red flag with all major ESPs, and recovery is slow. The lever is targeting plus copy: emails so off-target or so generic that recipients click "report spam" instead of just deleting are the actual problem.

If you need an infrastructure spec for higher volume, see 4 Best Infrastructures for Sending 10k Emails Daily.

Tools You'll Use to Run This

The full stack at meaningful volume:

Sequencer

The platform that runs your campaigns. Smartlead, Instantly, EmailBison, and PlusVibe are the four worth considering. Each ships with its own warmup network, reporting layer, and inbox management. Pick one and learn it well rather than splitting volume across multiple.

List Building and Validation

Apollo for list pulls. LeadMagic for email validation, including employment-status checks. BounceBan or Scrubby to validate catch-all domains so you don't throw away about 20% of any list. IcyPeas and Prospeo as upstream finders in a waterfall before LeadMagic.

Cold Email Infrastructure

Domains, DNS, IPs, mailbox provisioning, monitoring. You can build this layer yourself across registrars, Google Workspace, Microsoft 365, and DNS providers. Or you can buy it as a managed service.

Mistakes That Kill Cold Email Deliverability

A clean checklist of what to avoid:

Stacking Too Many Inboxes on a Domain

The whole point of this post. Past the provider ceiling (2 on Google, 2-3 on self-managed M365, 25 on ScaledMail-provisioned M365), the domain burns faster than the volume justifies.

Skipping DNS Authentication

SPF, DKIM, and DMARC are not optional. A domain without all three lands in spam by default at most major ESPs. Misconfigured DKIM (the most common error) is worse than missing DKIM, since it actively fails verification.

Tracking Opens and Clicks

Both downgrade deliverability via the pixel and redirect mechanisms. Both produce broken data thanks to Apple Mail Privacy Protection and click prefetching by major ESPs. Disable both. Measure reply rate.

Sending the Same Copy Forever

Copy fingerprinting at the ESP level is the #1 cause of mailbox death past the first month. ESPs cluster on structural signature (sentence pattern, opening line, CTA shape, signature format), not just keywords. Spintax does not change the structural fingerprint.

The fix: 5+ structurally different variants of your base offer, each with its own spintax, each with its own subject pool. Rotate variants every ~5,000 sends per mailbox.

Hyperlinking Your Company Name in the Signature

Auto-detected as a clickable link by Gmail, Outlook, and Apple Mail. Triggers spam filters at volume. Use the brand name as plain text, no .com.

Following Up Too Aggressively

Maximum 1-2 email touches per prospect, then go dark for 90 days. Long sequences spike spam complaints and burn domains. Most positive replies come from email #1 anyway.

Where ScaledMail Fits

At ScaledMail, we provision and manage the infrastructure layer end to end: secondary sending domains separate from your main business domain, real Google Workspace and Microsoft 365 inboxes, authentication configured correctly (SPF/DKIM/DMARC on every domain), IP rotation, and continuous reputation monitoring.

Warmup runs inside your sequencer (Smartlead, Instantly, EmailBison, PlusVibe), where the engagement signals live.

If you want the foundation built right so the only variables you're working with are your copy and your list, see the setup.

Frequently Asked Questions

Why can't I send cold email from my main business domain? One bad campaign tanks deliverability for the entire domain: invoicing, customer service, CEO communications, all of it. Once a domain is spam-flagged, recovery takes months if it's possible at all. Buy separate lookalike domains for cold outbound. Your main domain is too valuable to risk.

How long does it take to warm up a new inbox? Plan for 14 days of warmup-only volume before the first cold send. After that, warmup runs continuously at 1:1 ratio alongside cold sends for the active life of the inbox. Warmup is preventative and never ends. It runs as long as the mailbox is sending cold.

Is it better to have more inboxes on one domain or spread them across multiple domains? Spread them across multiple. The provider ceiling per domain is 2 (Google Workspace), 2-3 (self-managed M365 and SMTP), or 25 (ScaledMail-provisioned M365). Past the ceiling, the domain burns. To send more volume, add domains, not inboxes.

What's the first sign I'm scaling too fast? Reply rate dropping more than 25% week-over-week with the same ICP and copy. That's your earliest warning. Pause, rotate copy structurally, and resume on the same infrastructure if the dip is small. If it's already a 90% overnight drop, that's copy fingerprinting and you need new copy and new domains together.

If I follow these rules, am I guaranteed to stay out of spam? No, and anyone telling you otherwise is selling something. These rules drastically reduce risk. Deliverability still depends on list quality, copy structure, recipient ESP behavior, and dozens of other variables. The rules are the floor, not a guarantee.

How Many Inboxes Per Domain for Cold Email? Best Practices