mkt-attribution
Attribution Mapping
Communicate Track — Step 4 of 4. Validates the system — every initiative and content piece connects to a KPI, with action items for every gap.
Inputs Required
- KPIs from
.agents/mkt/targets.mdor existing OKRs - Initiatives from
.agents/mkt/prioritized-initiatives.md - Content from
.agents/mkt/imc-plan.mdor.agents/mkt/content/
Output
.agents/mkt/attribution.md
Quality Gate
Before delivering, verify:
- Every KPI has ≥1 initiative mapped or is explicitly flagged with an action item
- Every initiative maps to a KPI or has a "Kill" recommendation with reason
- Every recommendation is a verb (Create / Kill / Rebalance / Test) — not an observation ("this seems low")
- Review date is set (~4 weeks from today)
Chain Position
Previous: mkt-content-create | Next: Monthly review (re-run this skill)
Before Starting
Step 0: Product Context
Check for .agents/mkt/product-context.md. If available, read for product context.
Required Artifacts
None — can build from scratch by interviewing.
Optional Artifacts
| Artifact | Source | Benefit |
|---|---|---|
targets.md |
mkt-funnel-planner | KPI baselines |
prioritized-initiatives.md |
mkt-initiative-prioritize | Initiative list to map |
imc-plan.md |
mkt-imc | Content plan to audit |
Previous attribution.md |
mkt-attribution | Check progress on prior action items |
Read upstream files if they exist: .agents/mkt/targets.md, .agents/mkt/prioritized-initiatives.md, .agents/mkt/imc-plan.md. If previous .agents/mkt/attribution.md exists, read it — check progress on prior action items before building new map.
If no upstream files exist, build from scratch by interviewing:
- "What's your North Star metric?"
- "What are your 3-5 primary KPIs?"
- "What initiatives are you running or planning?"
- "What content are you producing?"
Step 1: KPI Hierarchy
North Star: [metric] — [current] → [target]
├── KPI 1: [metric] — [current] → [target] — Owner: [person]
├── KPI 2: [metric] — [current] → [target] — Owner: [person]
└── KPI 3: [metric] — [current] → [target] — Owner: [person]
Step 2: Map Initiatives → KPIs
| Initiative | Primary KPI | Contribution | Confidence |
|---|---|---|---|
| [Name] | [KPI] | ~X% of target | H/M/L |
Unknown contributions: If you can't estimate contribution, mark "Unknown" and add action item: "Run mkt-experiment to measure actual impact."
Flag Orphans
| Orphan Initiative | Action |
|---|---|
| [Name] | Kill — no KPI connection / Reframe — target [KPI] / Test — run experiment |
Step 3: Map Content → Initiatives
| Content | Initiative | Type | Channel |
|---|---|---|---|
| [Piece] | [Initiative] | Reach / Trust / Conversion | [Channel] |
Flag Orphan Content
| Orphan Content | Action |
|---|---|
| [Piece] | Assign to [initiative] / Repurpose for [initiative] / Sunset |
Step 4: Gap Analysis → Action Items
This is the core deliverable. Every gap produces a specific action.
| Gap | Action | Owner | Deadline |
|---|---|---|---|
| [KPI X uncovered] | Create: Run mkt-initiative-generate targeting this KPI |
[Who] | [When] |
| [Initiative Y orphaned] | Kill: Archive, remove from roster | [Who] | Now |
| [Content Z unassigned] | Rebalance: Redirect to support [initiative] | [Who] | [When] |
| [Contribution unknown] | Test: Design experiment via mkt-experiment |
[Who] | [When] |
| [Stage imbalanced] | Rebalance: Shift X% effort from [over] to [under] | [Who] | [When] |
Re-Run Protocol
Attribution maps should be reviewed and refreshed regularly:
- Monthly review cycle: Set review date ~4 weeks from creation
- On re-run:
- Read previous
.agents/mkt/attribution.md - Check status of each prior action item (Done / In Progress / Blocked / Dropped)
- Rename previous file to
attribution.v[N].md - Rebuild map with current state
- Compare: Did action items move the needle? Which KPIs improved?
- Read previous
- Escalation: If same gaps persist across 2+ reviews, escalate from "Rebalance" to "Kill" or "Create"
Artifact Template
---
skill: mkt-attribution
version: 1
date: [today's date]
status: draft
---
# Attribution Map
**Date:** [today]
**Skill:** mkt-attribution
**Review date:** [today + 4 weeks]
## KPI Hierarchy
```
North Star: [metric] — [current] → [target]
├── KPI 1: [metric] — [current] → [target] — Owner: [person]
├── KPI 2: ...
└── KPI 3: ...
```
## Initiative → KPI Mapping
| Initiative | KPI | Contribution | Confidence |
|------------|-----|-------------|------------|
| ... | ... | ... | ... |
### Orphan Initiatives
| Initiative | Action |
|------------|--------|
| ... | ... |
## Content → Initiative Mapping
| Content | Initiative | Type | Channel |
|---------|-----------|------|---------|
| ... | ... | ... | ... |
### Orphan Content
| Content | Action |
|---------|--------|
| ... | ... |
## Action Items
| # | Gap | Action | Owner | Deadline |
|---|-----|--------|-------|----------|
| 1 | ... | ... | ... | ... |
## Previous Review Progress
[If re-running: status of prior action items]
> On re-run: rename existing artifact to `attribution.v[N].md` and create new with incremented version.
Worked Example
# Attribution Map
**Date:** 2026-03-13
**Skill:** mkt-attribution
**Review date:** 2026-04-10
## KPI Hierarchy
```
North Star: Weekly Signups — 200 → 350
├── KPI 1: Paid signup rate — 1.2% → 3.0% — Owner: Sarah
├── KPI 2: Homepage bounce rate — 52% → 40% — Owner: James
└── KPI 3: Organic signups — 80/wk → 120/wk — Owner: (unassigned)
```
## Initiative → KPI Mapping
| Initiative | KPI | Contribution | Confidence |
|------------|-----|-------------|------------|
| Restore Paid Targeting | Paid signup rate | ~55% of total gap | H |
| Restore Social Proof | Homepage bounce | ~35% of total gap | M |
### Orphan Initiatives
None — both initiatives map to KPIs.
## Content → Initiative Mapping
| Content | Initiative | Type | Channel |
|---------|-----------|------|---------|
| Status Update Carousel | (none — brand awareness) | Reach | LinkedIn |
### Orphan Content
| Content | Action |
|---------|--------|
| Status Update Carousel | **Rebalance:** Add CTA linking to product page, reframe as supporting Paid Targeting initiative |
## Action Items
| # | Gap | Action | Owner | Deadline |
|---|-----|--------|-------|----------|
| 1 | KPI 3 (organic) uncovered | **Create:** Run `mkt-diagnosis` on organic traffic stagnation | James | 2026-03-20 |
| 2 | KPI 3 has no owner | **Assign:** Name an owner for organic signups KPI | Marketing lead | 2026-03-14 |
| 3 | Carousel content orphaned | **Rebalance:** Add product CTA to carousel, track as Paid Targeting support | Content lead | 2026-03-17 |
References
- references/attribution-examples.md — Worked examples across business types
More from hungv47/agent-skills
mkt-copywriting
This skill should be used when the user asks to "write copy", "create a headline", "write a tagline", "improve my copy", "write landing page copy", "create ad copy", "write email subject lines", "write product description", or mentions copywriting, persuasive writing, marketing copy, or conversion copy. Uses Harry Dry's Marketing Examples framework.
19tool-plan-interviewer
This skill should be used when the agent enters plan mode, when the user mentions "plan file", "spec file", "interview me about the plan", "help me think through this", or when starting implementation planning. Conducts in-depth, multi-round interviews using AskUserQuestion to surface non-obvious requirements, edge cases, and tradeoffs before writing specifications.
14design-user-flow
This skill should be used when the user asks to "create a user flow", "map navigation", "design user journey", "create wireflow", "plan screen transitions", "map the app flow", "design navigation", or mentions user flow, wireflow, navigation map, user journey, screen transitions, or pre-interface design research. Creates user flow diagrams and wireflows for digital products.
14eng-system-architecture
This skill should be used when the user asks to "design system architecture", "plan the tech stack", "create database schema", "design API structure", "architect the system", "plan the backend", "design infrastructure", or mentions system architecture, technical architecture, database design, API design, or tech stack selection. Transforms product documentation into comprehensive technical blueprints.
13mkt-diagnosis
This skill should be used when the user asks to "diagnose the problem", "break down the issue", "build a logic tree", "what's causing this", or mentions problem diagnosis, logic trees, MECE decomposition, or structured problem solving. This is the ENTRY POINT for problem analysis — use when no diagnosis exists yet.
13mkt-lp-optimization
This skill should be used when the user asks to "optimize a landing page", "improve conversion rates", "review my landing page", "create a high-converting landing page", "audit my LP", "reduce bounce rate", or mentions landing page design, conversion optimization, A/B testing strategy, or CTA optimization.
13