how to set up email infrastructure for cold outreach

The first rule of cold outreach infrastructure: your primary company domain (the one on your website and transactional emails) should never be used for cold outreach. If it gets flagged or blacklisted, your entire business email operation goes down with it.

**Buy dedicated sending domains.** Use slight variations of your brand — e.g., if your company is `acme.com`, register `getacme.com`, `useacme.com`, `tryacme.com`, or `acmehq.com`. Buy from [Namecheap](https://www.namecheap.com) or Google Domains, typically $10-15/year each.

**How many domains do you need?** A practical rule: each domain should support no more than 2-3 mailboxes, and each mailbox should send no more than 30-50 emails/day during steady-state operation. For a 500 email/day program, you need roughly 4-5 domains and 10-15 mailboxes.

**Domain age matters.** Newly registered domains are inherently lower-trust. Buy domains at least 2-3 weeks before you plan to start warming — ideally a month ahead. Some practitioners buy domains in bulk and age them for 30-60 days before activating.

**Forward your sending domains to your main site.** Set up a simple redirect from `getacme.com` → `acme.com`. This way, if a prospect types the domain into a browser, they land somewhere real — a signal that improves domain trust with mailbox providers.

**Use Google Workspace or Microsoft 365.** Avoid cheap or obscure hosting providers. Google Workspace ($6/user/month) and Microsoft 365 ($6/user/month) carry built-in sender reputation and are far more trusted by receiving mail servers than Zoho or random SMTP hosts. Alternate between Google and Microsoft accounts across your domain portfolio to diversify risk.

Proper DNS authentication is non-negotiable. Without it, your emails will either land in spam or be rejected outright. These three records prove to receiving mail servers that you are who you say you are.

**SPF (Sender Policy Framework):** A TXT record that lists which mail servers are authorized to send on behalf of your domain. For Google Workspace it looks like: `v=spf1 include:_spf.google.com ~all`. For Microsoft 365: `v=spf1 include:spf.protection.outlook.com ~all`. Add this in your domain registrar's DNS settings.

**DKIM (DomainKeys Identified Mail):** A cryptographic signature attached to outgoing emails. Google Workspace and Microsoft 365 generate DKIM keys in their admin consoles — you copy the public key and add it as a TXT record in DNS. Enable and activate DKIM in your email provider's admin panel before sending anything.

**DMARC (Domain-based Message Authentication):** Tells receiving servers what to do if SPF or DKIM fails. Start with a monitoring-only policy: `v=DMARC1; p=none; rua=mailto:dmarc@yourdomain.com`. After a few weeks of reviewing reports, graduate to `p=quarantine` and eventually `p=reject`. Tools like [MxToolbox](https://mxtoolbox.com) let you validate all three records are configured correctly.

**Custom Tracking Domains:** If you're tracking opens or clicks, use a custom tracking domain (e.g., `track.getacme.com`) instead of your sending platform's shared tracking domain. Shared tracking domains get flagged as spam triggers. Configure this in Instantly, Smartlead, or Apollo under their tracking settings.

**Check your records with:** MxToolbox, [Mail-Tester](https://www.mail-tester.com), or [GlockApps](https://glockapps.com) before you send a single email. A perfect authentication setup gives you a clean starting baseline.

A brand-new mailbox has zero reputation. Sending volume immediately signals spam behavior to Google and Microsoft's algorithms. Warmup is the process of gradually establishing a sending history that looks human.

**Automated warmup tools** make this largely hands-off. [Instantly](https://instantly.ai) and [Smartlead](https://smartlead.ai) both include built-in warmup networks — pools of real inboxes that exchange emails with your new mailboxes, open them, reply, and move them out of spam. This mimics organic engagement signals. [Lemwarm](https://lemwarm.com) (by Lemlist) is a standalone warmup tool if your sending platform doesn't include it.

**Warmup schedule:** - Week 1: 5-10 emails/day per mailbox (warmup traffic only, no cold outreach) - Week 2: 15-20 emails/day - Week 3: 25-30 emails/day - Week 4+: Introduce cold outreach at 20-30 emails/day, keeping warmup running in the background

**Never fully turn off warmup.** Keep warmup running continuously at 10-15 emails/day even when you're in full sending mode. It acts as a reputation buffer — if your cold emails start generating spam complaints, the warmup engagement offsets the damage.

**Inbox placement testing:** Before scaling, run inbox placement tests with [GlockApps](https://glockapps.com) or [MailReach](https://mailreach.co). These tools show you whether your emails are landing in Primary, Promotions, or Spam across major providers. If you're hitting spam on a new mailbox, don't start sending — investigate DNS, content, or domain age issues first.

**Realistic timeline:** From domain purchase to safe cold outreach takes a minimum of 3-4 weeks. Anyone telling you to skip warmup or do it in a week is giving you advice that will burn your infrastructure.

Your sending platform connects your mailboxes, manages sequences, rotates sending across inboxes, and handles replies. The choice of platform and how you configure it materially affects deliverability.

**Top platforms for cold outreach infrastructure:** - **[Instantly](https://instantly.ai):** Best for teams running high-volume outreach with many mailboxes. Excellent inbox rotation, built-in warmup, and unibox for reply management. - **[Smartlead](https://smartlead.ai):** Strong deliverability focus, good API access, better for technical teams that want custom workflows. - **[Apollo.io](https://apollo.io):** More of an all-in-one (data + sequencing), but deliverability controls are less granular. Good for teams that want simplicity over optimization. - **[Lemlist](https://lemlist.com):** Strong personalization features (images, landing pages), includes Lemwarm. - **[Outreach](https://outreach.io) / [Salesloft](https://salesloft.com):** Enterprise-grade, better for AE-driven outbound with SDR teams — not optimized for high-volume cold infrastructure.

**Inbox rotation:** Configure your sending platform to distribute sends across all mailboxes. If you have 10 mailboxes and want to send 300 emails/day, that's 30/mailbox — well within safe limits. Instantly and Smartlead do this automatically.

**Sending windows:** Restrict sends to business hours in your target timezone (8am-5pm Mon-Fri). Emails sent at 3am or on weekends have worse engagement rates and can trigger spam filters.

**Reply handling:** Use a unibox (unified inbox) so replies across all sending accounts route to one place. Both Instantly and Smartlead offer this. For teams at scale, consider routing replies through a CRM like HubSpot or Salesforce via Zapier or native integrations.

The cleanest infrastructure in the world won't save you if you're sending to bad data. High bounce rates (above 3-5%) are one of the fastest ways to tank domain reputation — mail servers interpret bounces as evidence you're blasting purchased or scraped lists.

**Verify every list before sending.** Run contacts through a real-time email verification tool before they ever enter a sequence: - **[ZeroBounce](https://zerobounce.net):** Industry standard, catches invalid, disposable, and catch-all addresses - **[NeverBounce](https://neverbounce.com):** Good bulk verification, integrates with most sending platforms - **[Millionverifier](https://millionverifier.com):** Cost-effective for high-volume verification - **[Bouncer](https://usebouncer.com):** Strong catch-all detection

**Handle catch-all domains carefully.** Catch-all domains accept all emails at the server level, so verification tools can't confirm individual addresses. These typically have 40-70% real deliverability. Either skip them entirely or send to them from separate, lower-reputation mailboxes so bounces don't damage your primary infrastructure.

**Data enrichment and sourcing:** Tools like [Clay](https://clay.com), [Apollo](https://apollo.io), [Cognism](https://cognism.com), and [ZoomInfo](https://zoominfo.com) provide contact data. Clay is particularly powerful for enriching leads with waterfalling across multiple data providers to find verified emails before they hit your sequences.

**Maintain a suppression list.** Anyone who unsubscribes, bounces, or marks you as spam should be added to a global suppression list that syncs across all sending accounts. This is both a legal requirement under CAN-SPAM/GDPR and a deliverability protection measure.

**Target bounce rate:** Keep hard bounces below 2% and total bounce rate below 5%. If you're consistently above these thresholds, pause sending and audit your data sources.

Cold email infrastructure degrades over time without active maintenance. Domains accumulate complaints, mailboxes age, and Google/Microsoft update their filtering algorithms. Build monitoring into your weekly RevOps workflow.

**Key metrics to track weekly:** - **Bounce rate** per domain and mailbox (target <2% hard bounces) - **Spam complaint rate** (Google Postmaster Tools shows this for Gmail recipients — target <0.1%) - **Inbox placement rate** (run monthly GlockApps tests) - **Open rates** — a sudden drop often signals deliverability degradation before bounce rates spike - **Reply rates** — the ultimate health metric

**Google Postmaster Tools:** Set up [Google Postmaster Tools](https://postmaster.google.com) for every sending domain. It's free and shows domain reputation, IP reputation, spam rates, and delivery errors directly from Google. This is the most authoritative signal you have for Gmail deliverability.

**Domain rotation strategy:** Retire domains after 6-12 months of active use or after any serious spam complaint spike. Buy replacement domains proactively — always have 1-2 aged, warmed domains in reserve. This is standard practice for high-volume outreach teams.

**Scaling the infrastructure:** Adding sending volume means adding domains and mailboxes in parallel — don't simply increase per-mailbox send limits. A team going from 500 to 1,500 emails/day needs to triple their domain/mailbox count, not triple the sends per mailbox.

**Content matters too:** Even perfect infrastructure fails with spammy content. Avoid spam trigger words, keep plain text or minimal HTML, use personalization ({{first_name}}, company-specific details via Clay), and test copy with [Mail-Tester](https://www.mail-tester.com) regularly.