Outreach Specialist

Key rule: Diagnostic offers outsell direct pitches 3:1 in cold outreach because they lower commitment and reframe the seller as advisor. Default to diagnostic unless the user has a specific reason not to.

2. What's the market segment?

Scaling rule: As company size increases, personalization depth must increase and volume must decrease. A 500-email SMB campaign can work at Level 2 personalization. A 500-email enterprise campaign at Level 2 will get 0 replies.

3. What's the volume tier?

Critical: High-volume only works with established infrastructure (30+ days warmed). Never launch high-volume from new domains -- deliverability will collapse within 48 hours.

4. What's the infrastructure state?

Deliverability-First Rules

These are the decisions where mistakes cost weeks or months to recover from. Infrastructure before copy, always.

1. Authentication is non-negotiable. SPF + DKIM + DMARC must all pass before sending a single cold email. Missing any one cuts inbox placement by 30-50%. See references/infrastructure-setup.md for configuration.
2. Warm-up cannot be skipped or rushed. The 30-day ramp exists because mailbox providers build sender reputation over time. Shortcuts trigger spam filters that take longer to clear than the warm-up itself.
3. 50 emails/day/account is the ceiling for Google Workspace. At 51+, Google's rate limiting activates unpredictably. Microsoft 365 allows ~80/day but flags patterns faster.
4. Bounce rate above 4% is an emergency. Stop all sends immediately. Audit the list. Every bounce degrades domain reputation and the damage compounds -- a 6% bounce day can take 2 weeks to recover from.
5. Spam complaint threshold is 0.1%. One complaint per 1,000 emails. Above this, inbox placement drops across ALL campaigns on that domain, not just the offending one.
6. Plain text outperforms HTML in cold email. No images, no HTML formatting, no multiple links. One link maximum. Tracking pixels are acceptable but reduce deliverability by ~5%.
7. Never send cold email from your primary business domain. Use a dedicated outreach domain (e.g., company-reach.com instead of company.com). If the outreach domain gets burned, your business email is unaffected.

Targeting Precision Framework

Why micro-campaigns beat blast campaigns: a 50-contact campaign with Level 3 personalization produces 2.76x the reply rate of a 500-contact campaign with Level 1 personalization. The math always favors precision.

Personalization Depth Levels

Decision rule: Spend personalization time where deal size justifies it. $500 ACV = Level 2 max. $5K ACV = Level 3. $50K+ ACV = Level 4.

ICP-to-Campaign Flow

1. Define ICP: industry + company size + role + geography + qualifying signal
2. Build niche list: one industry-role combination per micro-campaign (never mix)
3. Score leads: verified email + decision-maker confirmed + enrichment data available
4. Disqualify ruthlessly: no generic emails (info@, contact@), no unverifiable addresses, no companies outside ICP
5. Cap campaign size: 25-50 for micro, 100-200 for scaled, split into batches for high-volume

Email Copy Architecture

These are structural rules, not templates. Templates are in references/campaign-templates.md.

The 4 rules that separate 3% reply-rate emails from 15% reply-rate emails:

1. 60-120 words maximum. Every word above 120 reduces reply probability. Cold emails are not the place to educate -- they earn the right to a conversation.
2. 3rd-5th grade reading level. Short sentences. Common words. No jargon unless the recipient uses that exact jargon daily. Tools: Hemingway App or readability-score.com.
3. One CTA, stated once. "Would a 15-minute call this week make sense?" not "Book a demo OR reply OR visit our site OR download our whitepaper." Multiple CTAs produce zero action.
4. Subject lines that look internal. 2-4 words that could be an internal memo subject: "quick question", "scheduling idea", "[first name] - timing". Never salesy subjects ("Exclusive offer!", "Don't miss out!").

Follow-up architecture: See references/campaign-templates.md for full sequence frameworks. The key principle: each follow-up must introduce a genuinely new angle (proof, data, different pain point). Never reference the previous email ("Just following up on my last email...").

Anti-Pattern Decision Table

Rationalization Table

NEVER List

1. NEVER send cold email without SPF + DKIM + DMARC all passing. No exceptions, no "we'll set it up later."
2. NEVER exceed 50 emails/day per Google Workspace account or 80 per Microsoft 365 account.
3. NEVER launch campaigns from a domain less than 14 days into warm-up. Micro-campaigns only during warm-up.
4. NEVER send to unverified email addresses. Every address must pass verification before entering the send queue.
5. NEVER include more than one external link in a cold email. Each additional link increases spam score.
6. NEVER reference previous emails in follow-ups ("per my last email", "just circling back"). Each email stands alone.
7. NEVER use the prospect's first name more than once in an email. Repeated name insertion is an obvious automation tell.
8. NEVER send identical copy to more than 50 recipients. Mailbox providers detect duplicate content and flag it.
9. NEVER continue sending when bounce rate exceeds 4%. Stop immediately, audit, fix, then resume.
10. NEVER pitch a solution before diagnosing the problem. The first email earns a conversation, not a sale.

Performance Diagnostic Framework

When campaign metrics are below benchmarks (see references/outreach-benchmarks.md), diagnose in this exact order. Each layer must be healthy before the next matters.

Layer 1: DELIVERABILITY (are emails reaching inboxes?)
  Check: inbox placement rate, bounce rate, spam folder rate
  Fix: authentication, warm-up status, sending volume, list quality
  |
Layer 2: TARGETING (are the right people receiving emails?)
  Check: ICP match rate, title accuracy, company fit
  Fix: list source, enrichment quality, qualification criteria
  |
Layer 3: COPY (is the message compelling?)
  Check: open rate (subject line), reply rate (body + CTA)
  Fix: subject line testing, body length, CTA clarity, reading level
  |
Layer 4: TIMING (is the send time optimal?)
  Check: open rate by day/time, industry work patterns
  Fix: send window adjustment, timezone targeting, seasonal factors

Critical rule: If Layer 1 is broken, do NOT touch Layers 2-4. Rewriting copy when emails land in spam is wasted effort. Fix infrastructure first, measure again, then move up the stack.