ABM Campaign

Build and execute account-based marketing campaigns across three tiers of personalization. From fully bespoke 1:1 plays for strategic accounts to programmatic 1:many sequences at scale.

Research substrate (Exa): primary tools company_research_exa, web_search_exa for account-level intel + trigger events. Citations per .claude/rules/exa-protocol.md ([VERIFIED: exa_search, {url}, accessed YYYY-MM-DD]). Quality gate: ≥3 sources per major claim, ≥50% verified, date filter for any "recent" claim.

---

Claude Code Triggers

Invoke this skill when user says:

• "build an ABM campaign"
• "account-based marketing for [client]"
• "target account list"
• "tiered outreach campaign"
• "reactivation campaign for [segment]"
• "re-engage churned accounts"
• "1:1 account plays"
• "programmatic ABM"

Do NOT invoke when:

• User wants broad lifecycle campaigns → use lifecycle-marketing
• User wants a single cold email → use outreach-emails
• User wants lead nurture sequences → use lifecycle-marketing
• User wants content strategy without account targeting → use content-strategy

---

Three tiers of ABM (voice-locked)

ABM isn't one-size-fits-all. The tier determines how much time, research, and personalization goes into each account. Pick the right tier based on deal size, strategic value, and available bandwidth.

Tier 1: 1:1 (strategic accounts)

Dimension	Specification
Account count	5-15 accounts
Personalization	Fully bespoke. Custom research, messaging, and assets per account
Research depth	Deep dive: org chart, trigger events, stakeholder mapping, competitive displacement
Engagement model	Multi-threaded. Target 3-5 stakeholders per account across roles
Channel mix	Email + LinkedIn + direct mail + personalized content + events/webinars
Cadence	12-20 touches over 8-12 weeks per stakeholder
When to use	High ACV ($50K+), strategic logos, competitive displacement, expansion plays

Tier 2: 1:few (segment plays)

Dimension	Specification
Account count	50-100 accounts grouped into 3-5 segments
Personalization	Segment-level. Same pain points, same industry, same use case
Research depth	Segment research: common challenges, industry trends, shared triggers
Engagement model	Single-threaded initially, multi-thread on engagement signal
Channel mix	Email + LinkedIn + retargeting ads
Cadence	8-12 touches over 6-8 weeks
When to use	Mid-market ACV ($10-50K), clear segment patterns, semi-automated execution

Tier 3: 1:many (programmatic)

Dimension	Specification
Account count	200-500+ accounts
Personalization	Dynamic. Merge fields for name, company, pain point, industry
Research depth	Enrichment-level: firmographics, technographics, intent data
Engagement model	Single-threaded, automated sequences
Channel mix	Email + LinkedIn automation + programmatic ads
Cadence	6-8 touches over 4-6 weeks
When to use	Lower ACV ($5-15K), volume plays, market awareness, pipeline generation

---

Inputs

Input	Description	Source
Campaign goal	Specific outcome with success metric	User specifies
Target market/segment	Who we're going after, why now	User + ICP context
Account list or criteria	Existing list OR rules to build one	User-provided / Clay / Apollo
Channels available	Email, LinkedIn, ads, direct mail	User specifies
Tooling available	CRM, enrichment, sequence platform	User specifies

---

Process

Six-phase flow: Account selection & tiering → Account intel → Play design → Content creation → Execution → Measurement. Each phase has tier-specific outputs. Full step-by-step + fit scoring + cadence templates + MCP integration in the premium reference.

Reactivation play (sub-workflow for cold/churned/dormant accounts) also documented in the premium reference.

---

Anti-hallucination guardrails

1. Never invent account data. Only use verified firmographics from Clay, Apollo, or user-provided lists.
2. Never fabricate engagement history. Check CRM or Gmail for actual touchpoint history.
3. Never assume stakeholder names. Use enrichment tools or mark as [PLACEHOLDER: champion name].
4. Never invent intent signals. Only reference verified intent data or observable behavior.
5. Mark missing data explicitly. Use [PLACEHOLDER: description] for anything unconfirmed.
6. No fake urgency. Only use urgency when there's a real trigger event or deadline.

Per Apollo credit gate (.claude/rules/apollo-credits.md): search is free; enrichment costs credits. Always confirm spend before bulk enrichment.

---

Quality

Pre-delivery checks cover campaign structure (tiering rationale, cadence realism), content quality (Tier 1 hand-written feel, "would they forward?" test), and data quality (no invented stakeholders or engagement history). Worked example + anti-examples (over-tiering, generic Tier 1, Tier 3 cadence pretending to be Tier 2) + post-launch failure-mode triage in the premium reference.

---

Integration with other skills

Skill	Relationship	Usage
icp-research	depends_on	ICP criteria for account scoring and segmentation
lead-scoring	depends_on	Tier assignments use the scored account list
company-context	feeds_into	Deep account research for Tier 1 intelligence briefs
product-messaging	feeds_into	Value props and messaging for email and content creation
outreach-emails	feeds_into	Individual email drafts within ABM cadences
lifecycle-marketing	feeds_into	Post-conversion nurture for won ABM accounts
sales-enablement	feeds_into	Handoff materials when accounts hit engagement threshold
competitor-research	enhances	Competitive context for Tier 1 displacement plays

---

Final ship gate

Run /premortem --output before ship. See /premortem skill for the 5 execution domains (will-it-resonate / will-it-convert / will-it-stay-on-brand / will-stakeholder-push-back / will-it-degrade-over-time) and output template.

Trivial-case escape: ## Premortem\nNo failure modes — trivial change satisfies the contract for genuinely trivial outputs.

---