What Is Cold Email Sending Infrastructure?

Cold email sending infrastructure is everything between your list and the recipient's inbox: sending domains, mailboxes, DNS records and the tool that connects them. Get it wrong and good copy never gets read. This page covers the parts, how they work together and the one input people forget.

Key takeaways

• Infrastructure has four layers: domains, mailboxes, DNS authentication and a sending tool. Each one can fail on its own.
• Most teams send from secondary domains, not the main company domain, so a reputation problem stays contained.
• Volume is capped per mailbox and ramped slowly. More mailboxes give more total volume, not faster ramps.
• The list is the fifth layer. Bounces and spam complaints damage a domain faster than any DNS mistake.

What it is

Cold email sending infrastructure is the set of domains, mailboxes, DNS records and sending software that together deliver outbound email and build the sender reputation that decides inbox placement.

A domain is the name you send from. A mailbox is one account on that domain, such as jane@acme.com (a placeholder). DNS records prove to receivers that the mailbox is allowed to send for the domain.

The sending tool, often called a sequencer, logs in to the mailboxes and sends on a schedule. It is the only layer most buyers compare. It is also the layer that fixes none of the others.

How it works

The layers stack in a fixed order. Build them bottom to top.

1. Register the domains. Buy separate sending domains that look like your brand but are not your main domain. For example, getacme.com next to acme.com. A reputation hit then lands on the spare, not on the domain your staff use for normal mail.
2. Create mailboxes. Add a small number of mailboxes per domain on Google Workspace, Microsoft 365 or another host. Each mailbox has its own daily send limit and its own reputation signals.
3. Publish authentication records. Set SPF (RFC 7208), DKIM (RFC 6376) and DMARC (RFC 7489) on every sending domain. See what SPF, DKIM and DMARC are for cold email.
4. Warm up. Raise volume in small steps so providers see steady, wanted mail first. Tools compared in best cold email warm-up tools.
5. Connect the sending tool. Link each mailbox, set a daily cap and add an unsubscribe route. Best cold email campaigns compares the options.
6. Feed it a clean list. Load only addresses you have checked. Watch bounces and spam complaints from the first day.

A worked example

Say you plan 180 cold emails a day. Assume you cap each mailbox at 30 a day. That cap is your own choice, not a standard.

• 180 divided by 30 gives 6 mailboxes.
• Put 2 mailboxes on each of 3 secondary domains: getacme.com, tryacme.com and acme-mail.com (all placeholders).
• Publish SPF, DKIM and DMARC on all 3 domains before the first send.
• Warm all 6 mailboxes, then raise the caps over several weeks.

A 2,000-contact list at 180 a day takes about 11 days to send once. If one domain starts to bounce, you pause its 2 mailboxes and the other 4 keep sending.

Sending infrastructure vs a sending tool

People use the two terms as if they mean the same thing. A sequencer is one part of the infrastructure, not all of it.

A good sequencer on a bad domain still lands in spam. A good domain with a bad sequencer setup still burns reputation. You need both.

When it matters

Before your first campaign

Build the stack before you write the first email. A new domain has no reputation, and Gmail treats unknown senders cautiously. Google lists SPF or DKIM as a baseline for all senders and asks for all three records at 5,000 or more messages a day (Google sender guidelines, September 2026).

When you scale volume

Add mailboxes and domains, not volume per mailbox. Raising one mailbox from 30 to 300 a day looks like a spike to a receiver. Adding nine more mailboxes at 30 a day each looks like normal growth.

Keep spam complaints low. Google asks senders to stay under a 0.10% spam rate in Postmaster Tools and never reach 0.30% (Google sender guidelines, September 2026).

When replies drop

Check the layers in order. Read DMARC reports for failed sources, then check bounce rates by domain, then check the list. Rewrite copy last. Most placement drops trace to one domain or one bad list segment.

When a domain gets burned

Stop sending from it. Do not move the same list to a new domain and repeat the pattern. Find what caused the bounces or complaints first. A fresh domain with the same list is burned in days.

How LeadOcean handles it

LeadOcean does not register domains, host mailboxes or send email. It has no sequencer, dialer or CRM. It supplies the layer people skip: a list you can filter by email status before it touches your infrastructure.

The field is email_status and the filter is emailStatus. LeadOcean tracks 13 email status values, including verified, catch_all_valid, catch_all, risky and invalid. A default LeadOcean count covers the mailable set: verified, catch_all_valid and catch_all.

Sizing is free. Send count=true as a query parameter with limit set to 1. The call returns meta.total, capped at 100,000, and spends no records.

curl -X POST "https://api.leadocean.io/v1/people/search?count=true" \
  -H "x-api-key: $LEADOCEAN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "jobLevel": ["VP"],
    "jobFunction": ["Sales & Business Development"],
    "country": ["CA"],
    "emailStatus": ["verified"],
    "limit": 1
  }'

Passing emailStatus replaces the default, so this count is people with a verified email, not the wider mailable set. Drop count=true and raise limit to pull rows. Each person returned counts as one record. The app at app.leadocean.io has an Exports page with a filter builder, a column picker and a CSV download.

LeadOcean gives no refund or credit-back for bounced emails. Read email_status on every row before you load a sequencer. The free plan is 1,000 records, one-off, with no card. Pro is $499 a month, flat. See pricing, and the cold email deliverability guide for the full picture.

FAQ