email-deliverability

Email Deliverability

Overview

Deliverability is the foundation on which all outbound email rests. Without it, the best copy, the most sophisticated sequence, and the highest-quality list produce zero pipeline — because the email never reaches the inbox. The inbox placement rate is the single most important metric in cold email, and it is entirely a function of technical configuration plus ongoing reputation management.

This skill produces a deliverability setup and monitoring plan: DNS authentication configuration (SPF, DKIM, DMARC), a 4-week warmup schedule with per-week volume targets, reputation monitoring protocols, bounce and complaint handling procedures, and a deliverability health dashboard definition. It covers the full lifecycle from initial setup through ongoing maintenance.

The non-obvious rule: most deliverability problems are not technical. They are behavioral. Sending too many emails too fast, sending to unverified addresses, sending from a primary domain, or ignoring bounce spikes will defeat even perfect DNS configuration. Deliverability is 30% setup and 70% discipline.

CRITICAL: Never describe validation internals. Email validation is a blackbox subject — this skill covers the strategy of validation (when to validate, how to handle results) but NEVER describes how validation algorithms work.

When to Use

• User says "set up SPF/DKIM/DMARC" or "configure DNS for email" → activate this skill
• User asks "how do I warm up a domain" or "what's the warmup schedule" → activate this skill
• User reports "my emails are going to spam" or "open rates dropped" → activate this skill
• User asks about "bounce rate targets," "complaint handling," or "reputation monitoring" → activate this skill
• User wants to "check domain reputation" or "set up DMARC monitoring" → activate this skill
• User asks "how many emails per day per mailbox" → activate this skill

Do NOT use for:

• Buying and configuring secondary domains and mailboxes → use domain-infrastructure
• Selecting a sending platform → use sending-platforms
• Designing the email sequence → use cold-email-strategy
• Writing email copy → use cold-email-copywriting

Authoritative Foundations

This skill draws from the following established methodologies:

• Google Postmaster Tools — The authoritative source for Gmail deliverability data. Provides domain-level spam rate, IP reputation, domain reputation, feedback loop data, authentication status, and encryption status. Every sending domain must be registered here. Google processes ~45% of B2B email — if your Gmail deliverability is broken, your outbound is broken.

• Microsoft SNDS (Smart Network Data Services) — The equivalent for Outlook/Hotmail/Office 365. Provides IP-level reputation data, spam complaint rates, and trap hits. Microsoft processes ~20% of B2B email. Combined with Google Postmaster Tools, these cover ~65% of all B2B inboxes.

• M³AAWG (Messaging, Malware, Mobile Anti-Abuse Working Group) — Industry best practices for email sending. Their guidelines inform the volume limits, authentication requirements, and complaint thresholds used by all major inbox providers and email service providers.

• Validity / Return Path — Email certification and reputation data. Industry-standard Sender Score (0-100) used by many inbox providers as a reputation input. Sender Score above 90 is good; below 70 is concerning; below 50 will cause widespread blocking.

• RFC 7208 (SPF), RFC 6376 (DKIM), RFC 7489 (DMARC) — The IETF standards that define these protocols. This skill describes compliant configurations, not every possible variation.

• Eric Nowoslawski (Growth Engine X) — Agency-scale deliverability defaults: 2 inboxes/domain, 30 sends/inbox/day baseline, 3-week warmup, 50% backup capacity, 1:1 active-to-backup inbox ratio. → ../cold-email-strategy/references/eric-nowoslawski-outbound.md

• Eric Nowoslawski — Growth Engine X Cold Email Infrastructure. Operational defaults for agency-scale outbound: 2 inboxes per domain, 30 sends/day/inbox baseline (escalate only on winning campaigns with ~1 positive per 50 sends), 3-week warmup minimum, 50% spare capacity always warming, 1:1 backup rule (backup inboxes = active sending capacity for zero-downtime swap). Burned inboxes are not reliably recyclable — provision fresh capacity. Canonical playbook → ../cold-email-strategy/references/eric-nowoslawski-outbound.md.

Prerequisites

• Run domain-infrastructure first — you need domains and mailboxes to configure deliverability for
• Required user inputs: sending domains, mailbox credentials/providers, current sending volume if already active
• Access required: DNS management console for each sending domain, admin access to Google Workspace / M365 / Zoho for mailbox configuration
• Optional: LeadMagic API key for pre-send email verification (verify addresses before adding to sequences to maintain low bounce rates)
• Reference files: references/deliverability-primer.md for detailed warmup schedules and sending limits

Step-by-Step Process

Phase 1: Intake

Gather the following from the user. Ask all questions at once:

1. Sending domains: List all domains being used for cold email. Confirm these are SECONDARY domains — never the primary business domain.

2. Mailbox details: For each domain, how many mailboxes? Which provider (Google Workspace, M365, Zoho, other)? Are they already sending or new?

3. Current volume: How many emails per day per mailbox currently? If already sending, what's the current bounce rate, spam complaint rate, open rate, reply rate?

4. DNS access: Who controls DNS? Do they have access to add SPF, DKIM, DMARC records? Which DNS provider?

5. Current authentication status: Are SPF, DKIM, DMARC already configured? If yes, what are the current records? If not, this is step one.

6. Sending platform: Which platform (Smartlead, Instantly, Salesforge, Apollo, other)? Different platforms have different warmup capabilities and authentication requirements.

7. Deliverability issues: Any known problems? Emails going to spam? Blacklisting? Sudden open rate drops? Specific domains blocking?

8. Compliance requirements: GDPR, CAN-SPAM, CASL, other regulations? This affects list practices and unsubscribe handling.

Phase 2: Research

Before configuring, audit the current state:

1. DNS audit: Check current SPF, DKIM, DMARC records for all sending domains. Use dig TXT [domain] for SPF, dig TXT [selector]._domainkey.[domain] for DKIM, dig TXT _dmarc.[domain] for DMARC. Document what exists and what's missing.

2. Blacklist check: Run all sending IPs and domains through MXToolbox, Spamhaus, Barracuda, and SURBL blacklists. If any domain or IP is listed, that's the first priority to address before sending resumes.

3. Reputation check: If domains have been sending, check Google Postmaster Tools for domain reputation, spam rate, and feedback loop. Check Microsoft SNDS for IP reputation. Check Sender Score (Validity) for overall score. Document the baseline.

4. Bounce analysis: If currently sending, pull bounce data for the last 30 days. Categorize bounces: hard bounce (invalid address), soft bounce (mailbox full, server issue), block (spam filter rejection). Calculate rates per mailbox and per domain.

5. Spam complaint analysis: Pull complaint data from the sending platform and Google Postmaster Tools. Industry target: under 0.1% complaint rate (1 complaint per 1,000 sends). Above 0.3% triggers Google spam filtering regardless of authentication.

6. Inbox placement test: If possible, use GlockApps, MailGenius, or Mail-Tester to send test emails and see where they land (inbox, spam, promotions, missing) across Google, Microsoft, Yahoo, and other providers.

Phase 3: Execution

Step 1: Configure SPF

SPF (Sender Policy Framework) tells receiving servers which IP addresses and hostnames are authorized to send mail on behalf of your domain.

Create/update the TXT record at the domain root:

For dedicated sending IPs (recommended for high volume):

v=spf1 ip4:[your_sending_ip] -all

For sending platforms with shared IPs (most SMB outbound):

v=spf1 include:[platform_spf_host] -all

Common platform SPF includes:

• Google Workspace: include:_spf.google.com
• M365: include:spf.protection.outlook.com
• Smartlead: include:_spf.smartlead.ai
• Instantly: include:_spf.instantly.ai
• Salesforge: include:spf.salesforge.ai

Rules:

• Use -all (hard fail) for cold sending domains, not ~all (soft fail). You want receiving servers to reject unauthorized mail, not mark it as suspicious.
• One SPF record per domain. If you already have an SPF record, merge the includes into a single record.
• No more than 10 DNS lookups per SPF record (the include chain counts). If you exceed this, use subdomain-specific SPF records.
• The + or - qualifier on all must be the last mechanism.

Step 2: Configure DKIM

DKIM (DomainKeys Identified Mail) adds a cryptographic signature to each email, proving the message wasn't altered in transit.

For Google Workspace:

1. In Google Admin console, go to Apps → Google Workspace → Gmail → Authenticate email.
2. Generate a DKIM key for your domain. Google provides the DNS record.
3. Add the TXT record at google._domainkey.[domain] with the value Google provides.
4. After DNS propagation (up to 48 hours), click "Start Authentication" in Google Admin.

For M365:

1. In Microsoft 365 Defender portal, go to Email & Collaboration → Policies & Rules → Threat Policies → Email Authentication Settings → DKIM.
2. Select your domain, enable DKIM, and copy the CNAME records.
3. Add CNAME records: selector1._domainkey.[domain] and selector2._domainkey.[domain] pointing to the Microsoft-provided targets.

For Zoho:

1. In Zoho Mail Admin, go to Email Authentication → DKIM.
2. Generate keys and add the provided TXT record.

Key selection: Use a distinct DKIM selector for each sending platform (e.g., google._domainkey, smartlead._domainkey). This allows independent key rotation without breaking other services.

DKIM key length: 2048-bit minimum. 1024-bit keys are deprecated and increasingly rejected by major inbox providers.

Step 3: Configure DMARC

DMARC (Domain-based Message Authentication, Reporting and Conformance) tells receiving servers what to do when SPF or DKIM fails and provides aggregate reports on authentication results.

Create the TXT record at _dmarc.[domain]:

Starting configuration (monitoring mode):

v=DMARC1; p=none; rua=mailto:dmarc-reports@[yourdomain].com; ruf=mailto:dmarc-forensics@[yourdomain].com; fo=1;

After 2-4 weeks of monitoring with zero issues, move to quarantine:

v=DMARC1; p=quarantine; pct=100; rua=mailto:dmarc-reports@[yourdomain].com;

Eventual target (full protection):

v=DMARC1; p=reject; pct=100; rua=mailto:dmarc-reports@[yourdomain].com;

Rules:

• Start with p=none — you need monitoring data before enforcement.
• Set up rua to a real email address that receives DMARC aggregate reports (XML files). Use services like DMARCIAN, Postmark DMARC, or Valimail to parse these into human-readable dashboards.
• After 2-4 weeks of clean monitoring (no legitimate mail failing auth), move to p=quarantine for another 2-4 weeks.
• p=reject is the goal but only after full confidence. A misconfigured reject policy will silently discard legitimate mail.
• pct=100 means apply the policy to 100% of mail. Use pct=25 or pct=50 to gradually roll out enforcement.

Step 4: Design the Warmup Schedule

New domains and mailboxes need 2-4 weeks of warmup before sending cold email at full volume. Warmup builds a positive reputation by sending small volumes to known-engaging addresses, establishing a pattern of high engagement before introducing colder contacts.

Week 1 — Foundation (Days 1-7):

• Daily volume: 5-10 emails per mailbox
• Recipient composition: 100% warm (internal team, friends, colleagues, test accounts) who will open and reply
• Sending pattern: spread across the day, not all at once
• Goal: establish initial engagement signal

Week 2 — Gradual Increase (Days 8-14):

• Daily volume: 15-25 emails per mailbox
• Recipient composition: 70% warm, 30% lukewarm (known contacts, past conversations, opted-in lists)
• Sending pattern: spread across business hours
• Goal: build volume tolerance while maintaining high engagement

Week 3 — Introducing Cold (Days 15-21):

• Daily volume: 30-40 emails per mailbox
• Recipient composition: 50% warm/lukewarm, 50% cold (verified addresses, Tier 3 personalization, highest-quality leads first)
• Sending pattern: normal business hours
• Goal: transition to cold sending without engagement cliff

Week 4 — Full Volume (Days 22-28):

• Daily volume: 40-50 emails per mailbox
• Recipient composition: 20% warm/lukewarm, 80% cold
• Sending pattern: optimized for recipient time zone
• Goal: sustained reputation at full capacity

Important: If at any point during warmup the bounce rate exceeds 2% or the spam complaint rate exceeds 0.1%, pause the escalation and stay at the current volume until resolved.

Using platform-based warmup (Smartlead, Instantly, Salesforge): Most platforms include automated warmup that sends to their network of warmup addresses. This is a supplement, not a replacement for the manual warmup above. Automated warmup builds base reputation; manual warmup with real contacts builds domain-specific engagement patterns.

NEVER use automated warmup as the sole warmup method. The warmup network addresses are known to inbox providers — engagement with them is weighted lower than engagement with real, diverse contacts.

Step 5: Set Up Reputation Monitoring

Ongoing monitoring is not optional. Set up the following:

1. Google Postmaster Tools: Register every sending domain. Check weekly at minimum. Key metrics to monitor:

   • Spam rate: must stay under 0.1% (target <0.05%)
   • Domain reputation: must stay "high" or "medium-high"
   • Feedback loop: watch for complaint spikes
   • Authentication: all DKIM/SPF should show "pass"

2. Microsoft SNDS: Register every sending IP. Check weekly. Watch for:

   • Spam complaint rate: under 0.1%
   • Trap hits: zero is the only acceptable number
   • Block status: any block requires immediate investigation

3. Blacklist monitoring: Use MXToolbox or similar to check major blacklists weekly. Automate with alerts. Blacklists to monitor:

   • Spamhaus (SBL, XBL, PBL)
   • Barracuda
   • SURBL
   • SpamCop
   • Invaluement
   • UCEPROTECT

4. Sending platform metrics: Monitor in-platform deliverability metrics daily:

   • Delivery rate (target: >98%)
   • Bounce rate (target: <2%, warning at 3%, critical at 5%)
   • Spam complaint rate (target: <0.1%, critical at 0.3%)
   • Open rate trends (a sudden 30%+ drop signals deliverability issue)
   • Inbox placement rate (if platform provides)

5. DMARC aggregate reports: Parse the XML reports from rua to track:

   • Which IPs are sending on your behalf (authorized and unauthorized)
   • SPF and DKIM pass/fail rates
   • Spoofing attempts (someone else sending as your domain)
   • Gradual DKIM alignment failures (key rotation issues)

Step 6: Establish Bounce and Complaint Handling Procedures

Use references/bounce-complaint-procedures.md for the full procedure library. Main output should include bounce thresholds, complaint thresholds, owner, escalation route, and stop conditions.

Phase 4: Delivery

Deliver a complete deliverability setup and monitoring plan:

1. DNS Configuration Spec: Exact SPF, DKIM, DMARC records to add, formatted for copy-paste into the DNS console.

2. Warmup Calendar: Week-by-week volume targets, recipient composition, monitoring checkpoints. Formatted as a calendar with daily targets.

3. Monitoring Dashboard Spec: What to track, where, how often, and what thresholds trigger action.

4. Incident Response Plan: For bounce spikes, complaint spikes, blacklisting, and domain reputation drops. What to do in the first hour, first day, first week.

5. Quarterly Audit Checklist: Deliverability degrades over time. A quarterly audit prevents silent decay.

Output Format

# Deliverability Setup Plan for [Company]

## Output Artifact Details

Use `references/output-artifacts.md` for the full audit, DNS configuration, warmup schedule, monitoring dashboard, incident response plan, and quarterly audit checklist. The main output must summarize current risk, DNS status, ramp schedule, alert thresholds, and owner.

## Quality Check

Before delivering, verify:

- [ ] Are SPF, DKIM, DMARC records provided in copy-paste format?
- [ ] Does SPF use `-all` (hard fail), not `~all`?
- [ ] Is DKIM using 2048-bit keys minimum?
- [ ] Does DMARC start with `p=none` and plan to progress to `p=reject`?
- [ ] Is the warmup schedule 2-4 weeks with per-week volume targets?
- [ ] Does the plan NEVER recommend sending cold from the primary domain?
- [ ] Are bounce thresholds clear (target <2%, warning 2-5%, critical >5%)?
- [ ] Are complaint thresholds clear (target <0.1%, critical >0.3%)?
- [ ] Is Google Postmaster Tools registration included?
- [ ] Is blacklist monitoring included with specific blacklists named?
- [ ] Does the incident response plan have time-based escalation steps?
- [ ] Is there a quarterly audit checklist?
- [ ] Are unsubscribe requirements specified (CAN-SPAM, GDPR)?
- [ ] Is email validation mentioned but NEVER described internally?

## Common Pitfalls

1. **Sending cold email from the primary business domain.** One spam
   complaint, one blacklisting event, and your transactional emails
   (password resets, invoices, customer notifications) are broken.
   This is the #1 deliverability mistake and it is catastrophic when
   it happens. Always use secondary domains for cold outreach.

2. **Skipping warmup.** A new domain sending 50 emails on Day 1 gets
   flagged within 48 hours. Google and Microsoft's algorithms interpret
   sudden volume from a new domain as spam behavior. No amount of
   authentication quality overrides the volume pattern.

3. **Soft DMARC (`p=none`) forever.** Many teams set `p=none` and never
   progress to enforcement. This is better than nothing but leaves the
   domain vulnerable to spoofing. Have a plan to reach `p=reject` within
   8 weeks of clean monitoring.

4. **Ignoring bounce rate creep.** A bounce rate that creeps from 1% to 3%
   over months is easy to miss because it's gradual. But inbox providers
   catch it. Set a weekly bounce rate review and a hard alert at 3%.

5. **No email verification before sending.** Every unverified address in a
   sequence is a potential bounce. Bounces above 5% trigger spam filtering
   at Google and Microsoft, and recovery takes weeks. Always verify before
   sending.

6. **Warmup-only via platform networks.** The automated warmup networks
   (warmup pools) are known to inbox providers. They're a supplement, not
   a replacement. Real engagement from real, diverse contacts builds
   stronger reputation than automated warmup alone.

7. **No DMARC reporting.** Without `rua` reports, you don't know who is
   sending on your behalf. You won't catch spoofing attempts, gradual
   DKIM alignment failures, or misconfigured third-party senders until
   they cause a deliverability crisis.

8. **Too many mailboxes per domain.** When one mailbox on a domain gets
   flagged for spam, it affects the domain reputation — which affects
   all mailboxes on that domain. Limit to 2-3 mailboxes per domain to
   contain blast radius.

## Sending Limits Reference

Use `references/sending-limits-reference.md` for provider-specific sending limits. Keep the primary output focused on safe ramping, monitoring thresholds, DNS auth, and incident response.

## Execution Artifacts

- `references/framework-notes.md` — Framework index and authority routing
- `templates/output-template.md` — Deliverable shell for agent output
- `scripts/check-output.py` — Lightweight deliverable validator
- `../cold-email-strategy/references/eric-nowoslawski-outbound.md` — Infra at scale, backup inbox strategy, volume escalation rules (Eric Nowoslawski)
- `references/bounce-complaint-procedures.md` — Bounce and complaint handling SOP
- `references/deliverability-primer.md` — Deliverability fundamentals
- `references/output-artifacts.md` — Extended output tables and inventories
- `references/sending-limits-reference.md` — Provider sending limits reference

## Related Skills

- `domain-infrastructure` — adjacent workflow to use before or after this skill
- `sending-platforms` — adjacent workflow to use before or after this skill
- `cold-email-strategy` — adjacent workflow to use before or after this skill
- `campaign-analytics` — adjacent workflow to use before or after this skill