返回技能列表

revenue-attribution

guia-matthieu
更新于 2 days ago
4 次查看
111
20
111
在 GitHub 上查看
其他general

关于

The revenue-attribution skill analyzes marketing and sales touchpoints to attribute revenue credit using multi-touch models like first-touch, last-touch, and linear attribution. It's used for justifying spend, optimizing channel mix, and evaluating campaign ROI by determining which activities drive conversions. This enables smarter budget allocation and resolves credit disputes between marketing and sales.

快速安装

Claude Code

推荐
主要方式
npx skills add guia-matthieu/clawfu-skills -a claude-code
插件命令备选方式
/plugin add https://github.com/guia-matthieu/clawfu-skills
Git 克隆备选方式
git clone https://github.com/guia-matthieu/clawfu-skills.git ~/.claude/skills/revenue-attribution

在 Claude Code 中复制并粘贴此命令以安装该技能

技能文档

Revenue Attribution

Determine which marketing and sales activities drive revenue using multi-touch attribution models, enabling smarter budget allocation and campaign optimization.

When to Use This Skill

  • Justifying marketing spend to leadership
  • Optimizing channel mix allocation
  • Evaluating campaign ROI
  • Resolving marketing/sales credit disputes
  • Building attribution reports

Methodology Foundation

Based on Bizible/Marketo Multi-Touch Attribution and Google Analytics Attribution Models, covering:

  • First-touch attribution (awareness credit)
  • Last-touch attribution (conversion credit)
  • Linear attribution (equal credit)
  • Time-decay attribution (recency-weighted)
  • Position-based (U-shaped, W-shaped)

What Claude Does vs What You Decide

Claude DoesYou Decide
Explains attribution modelsWhich model fits your business
Calculates credit distributionHow to act on insights
Identifies top-performing channelsBudget reallocation amounts
Shows model comparisonFinal attribution policy
Highlights discrepanciesException handling

What This Skill Does

  1. Model education - Explain different attribution approaches
  2. Credit calculation - Apply models to touchpoint data
  3. Channel analysis - Compare performance by source
  4. Model comparison - Show how results differ by model
  5. Optimization recommendations - Where to invest more/less

How to Use

Analyze attribution for this closed-won deal:

Deal: [Company Name]
Value: $[Amount]
Close Date: [Date]
Sales Cycle: [Days]

Touchpoint Journey:
1. [Date] - [Channel] - [Action]
2. [Date] - [Channel] - [Action]
...
[List all touchpoints chronologically]

Questions:
- Which channels deserve credit?
- Compare first-touch vs last-touch
- Recommend budget allocation

Instructions

Step 1: Understand Attribution Models

ModelLogicBest For
First-Touch100% to first interactionAwareness measurement
Last-Touch100% to final conversionDirect response
LinearEqual split across allLong consideration cycles
Time-DecayMore credit to recentSales-assisted journeys
Position-Based40/20/40 (first/middle/last)Balanced view
W-Shaped30/30/30 + 10 remainderInclude MQL moment

Step 2: Map the Customer Journey

Document all touchpoints with:

  • Timestamp - When it occurred
  • Channel - Source (Paid, Organic, Email, Event, etc.)
  • Action - What happened (visit, download, demo, etc.)
  • Stage - Awareness, Consideration, Decision

Step 3: Apply Attribution Model

First-Touch Example:

Journey: Paid Search → Email → Webinar → Demo → Close
Revenue: $50,000

First-Touch Attribution:
- Paid Search: $50,000 (100%)
- All others: $0

Linear Example:

Same journey, 4 touchpoints:
- Paid Search: $12,500 (25%)
- Email: $12,500 (25%)
- Webinar: $12,500 (25%)
- Demo: $12,500 (25%)

Position-Based (40/20/40):

- Paid Search: $20,000 (40% - first)
- Email: $5,000 (10% - middle)
- Webinar: $5,000 (10% - middle)
- Demo: $20,000 (40% - last)

Step 4: Aggregate by Channel

Sum attribution across all deals to see total channel contribution:

Channel Performance (Position-Based):
- Paid Search: $500K attributed (35%)
- Events: $300K attributed (21%)
- Organic: $280K attributed (19%)
- Email: $220K attributed (15%)
- Referral: $150K attributed (10%)

Step 5: Calculate ROI by Channel

Channel ROI = Attributed Revenue / Channel Spend

Example:
- Paid Search: $500K revenue / $100K spend = 5x ROI
- Events: $300K revenue / $200K spend = 1.5x ROI

Examples

Example 1: Deal-Level Attribution

Input:

Analyze attribution for TechStart Inc:

Deal: TechStart Inc
Value: $75,000 ARR
Close Date: March 15
Sales Cycle: 62 days

Touchpoint Journey:
1. Jan 12 - Google Ads - Clicked ad, visited pricing
2. Jan 15 - Organic Search - Blog post read
3. Jan 22 - LinkedIn Ad - Whitepaper download (MQL)
4. Feb 1 - Email nurture - Opened 3 emails
5. Feb 10 - Webinar - Attended "Product Demo Day"
6. Feb 18 - Direct - Requested demo (SQL)
7. Mar 5 - Sales - Discovery call
8. Mar 12 - Sales - Proposal review
9. Mar 15 - Direct - Signed contract

Compare first-touch, last-touch, and position-based.

Output:

## Attribution Analysis: TechStart Inc ($75,000)

### Journey Summary
- **Total Touchpoints**: 9
- **Marketing Touchpoints**: 6 (Jan 12 - Feb 18)
- **Sales Touchpoints**: 3 (Mar 5 - Mar 15)
- **Sales Cycle**: 62 days
- **Key Conversion**: LinkedIn whitepaper (MQL), Demo request (SQL)

### Model Comparison

| Channel | First-Touch | Last-Touch | Position-Based | Linear |
|---------|-------------|------------|----------------|--------|
| Google Ads | $75,000 | $0 | $22,500 | $8,333 |
| Organic Search | $0 | $0 | $3,750 | $8,333 |
| LinkedIn Ads | $0 | $0 | $7,500 | $8,333 |
| Email | $0 | $0 | $3,750 | $8,333 |
| Webinar | $0 | $0 | $3,750 | $8,333 |
| Direct (Demo) | $0 | $0 | $7,500 | $8,333 |
| Sales | $0 | $75,000 | $26,250 | $25,000 |

### Position-Based Breakdown (Recommended)

**First Touch (30%): $22,500**
- Google Ads gets full first-touch credit
- Drove initial awareness and pricing interest

**MQL Touch (30%): $22,500**
- LinkedIn Ads: $15,000 (whitepaper = MQL)
- Webinar: $7,500 (influenced SQL)

**Last Touch (30%): $22,500**
- Demo Request: $7,500
- Sales (Close): $15,000

**Middle Touches (10%): $7,500**
- Organic: $2,500
- Email: $2,500
- Other: $2,500

### Insights

1. **Google Ads** initiated the journey - valuable for awareness
2. **LinkedIn Ads** drove the MQL - critical conversion point
3. **Sales** closed but didn't initiate - last-touch overstates
4. **Email** nurtured but didn't convert alone - assist role

### Channel Recommendations

| Channel | Attribution | Action |
|---------|-------------|--------|
| Google Ads | High first-touch | Maintain/increase for awareness |
| LinkedIn Ads | High MQL driver | Invest more for lead gen |
| Webinar | Solid influence | Continue as mid-funnel |
| Email | Assist role | Optimize, don't cut |

Example 2: Multi-Deal Channel Analysis

Input:

Analyze Q1 attribution across these closed deals:

Deal 1: $50K - Google Ads → Email → Demo → Close
Deal 2: $30K - Event → LinkedIn → Demo → Close
Deal 3: $80K - Referral → Demo → Close
Deal 4: $45K - Organic → Webinar → Email → Demo → Close
Deal 5: $25K - Google Ads → Event → Demo → Close

Total Q1 Revenue: $230K

Channel Spend:
- Google Ads: $15K
- Events: $25K
- LinkedIn: $8K
- Email: $3K
- Webinar: $5K
- Referral: $0

Which channels should we invest more in?

Output:

## Q1 Channel Attribution Analysis

### Total Revenue: $230,000

### Attribution by Model

#### Position-Based Attribution (40/20/40)

| Channel | First | Middle | Last | Total | % of Rev |
|---------|-------|--------|------|-------|----------|
| Google Ads | $30,000 | $2,500 | $0 | $32,500 | 14% |
| Events | $12,000 | $5,000 | $0 | $17,000 | 7% |
| LinkedIn | $0 | $6,000 | $0 | $6,000 | 3% |
| Email | $0 | $11,000 | $0 | $11,000 | 5% |
| Webinar | $0 | $9,000 | $0 | $9,000 | 4% |
| Referral | $32,000 | $0 | $0 | $32,000 | 14% |
| Demo | $0 | $0 | $30,500 | $30,500 | 13% |
| Sales/Close | $0 | $0 | $92,000 | $92,000 | 40% |

### ROI Analysis

| Channel | Attributed Rev | Spend | ROI |
|---------|----------------|-------|-----|
| Referral | $32,000 | $0 | ∞ (Best) |
| Google Ads | $32,500 | $15,000 | 2.2x |
| Webinar | $9,000 | $5,000 | 1.8x |
| LinkedIn | $6,000 | $8,000 | 0.75x |
| Events | $17,000 | $25,000 | 0.68x |
| Email | $11,000 | $3,000 | 3.7x |

### Efficiency Ranking

1. **🥇 Referral** - $0 cost, $32K attributed → Infinite ROI
2. **🥈 Email** - 3.7x ROI → High-value nurture
3. **🥉 Google Ads** - 2.2x ROI → Profitable acquisition
4. **Webinar** - 1.8x ROI → Solid mid-funnel
5. **LinkedIn** - 0.75x ROI → Below break-even
6. **Events** - 0.68x ROI → Expensive for return

### Recommendations

**Increase Investment:**
- **Referral Program**: 14% of revenue at $0 cost
  - Formalize referral rewards
  - Target: 2x referral deals in Q2

- **Email Nurture**: 3.7x ROI
  - Expand sequences
  - Add $2K budget for tools

- **Google Ads**: 2.2x ROI
  - Profitable, test 20% budget increase
  - Focus on high-intent keywords

**Optimize/Test:**
- **LinkedIn**: 0.75x is below target
  - Test new audiences before cutting
  - Could be essential for certain segments

**Reduce/Reallocate:**
- **Events**: $25K for $17K attributed
  - Evaluate which events drive pipeline
  - Consider smaller, targeted events
  - Reallocate $10K to Google Ads

### Proposed Q2 Budget Shift

| Channel | Q1 Spend | Q2 Proposed | Change |
|---------|----------|-------------|--------|
| Google Ads | $15K | $20K | +$5K |
| Events | $25K | $15K | -$10K |
| LinkedIn | $8K | $8K | — |
| Email | $3K | $5K | +$2K |
| Webinar | $5K | $6K | +$1K |
| Referral | $0 | $2K (rewards) | +$2K |
| **Total** | **$56K** | **$56K** | Rebalanced |

Skill Boundaries

What This Skill Does Well

  • Explaining attribution model mechanics
  • Calculating credit across touchpoints
  • Comparing models side-by-side
  • Identifying channel efficiency

What This Skill Cannot Do

  • Access actual CRM/analytics data
  • Track offline touchpoints automatically
  • Account for brand lift effects
  • Prove causation (only correlation)

When to Escalate to Human

  • Choosing official attribution model for company
  • Budget allocation decisions over $50K
  • Complex B2B journeys with multiple stakeholders
  • Reconciling attribution across systems

Iteration Guide

Follow-up Prompts

  • "How would results change with time-decay model?"
  • "What if we excluded sales touchpoints?"
  • "Show me channel performance by deal size."
  • "Build attribution for all Q1 deals (I'll provide data)."

Attribution Maturity

  1. Basic: Last-touch only
  2. Intermediate: First and last comparison
  3. Advanced: Position-based or custom
  4. Expert: ML-based algorithmic attribution

Checklists & Templates

Attribution Report Template

## Attribution Report: [Period]

### Summary
- Total Revenue: $X
- Deals Analyzed: X
- Model Used: [Position-Based]

### Channel Attribution
| Channel | Revenue | % | ROI |
|---------|---------|---|-----|

### Top Insights
1.
2.
3.

### Budget Recommendations
| Channel | Current | Recommended | Rationale |
|---------|---------|-------------|-----------|

Touchpoint Tracking Checklist

  • UTM parameters on all campaigns
  • CRM synced with marketing automation
  • Offline events logged manually
  • Sales activities timestamped
  • Content downloads tracked

References

  • Bizible Multi-Touch Attribution Guide
  • Google Analytics Attribution Modeling
  • Forrester B2B Attribution Research
  • Marketo Revenue Cycle Analytics

Related Skills

  • pipeline-forecasting - Predict revenue by source
  • lead-scoring - Score by attributed channel
  • ad-spend-optimizer - Automate budget shifts

Skill Metadata

  • Domain: RevOps
  • Complexity: Advanced
  • Mode: centaur
  • Time to Value: 30-60 min for analysis
  • Prerequisites: Touchpoint data, deal values, channel spend

GitHub 仓库

guia-matthieu/clawfu-skills
路径: skills/revops/revenue-attribution
0
ai-skillsanthropicclaude-codeclaude-skillsmarketingmcp-server

相关推荐技能

llamaguard

其他

LlamaGuard是Meta推出的7-8B参数内容审核模型,专门用于过滤LLM的输入和输出内容。它能检测六大安全风险类别(暴力/仇恨、性内容、武器、违禁品、自残、犯罪计划),准确率达94-95%。开发者可通过HuggingFace、vLLM或Sagemaker快速部署,并能与NeMo Guardrails集成实现自动化安全防护。

查看技能

cost-optimization

其他

这个Claude Skill帮助开发者优化云成本,通过资源调整、标记策略和预留实例来降低AWS、Azure和GCP的开支。它适用于减少云支出、分析基础设施成本或实施成本治理策略的场景。关键功能包括提供成本可视化、资源规模调整指导和定价模型优化建议。

查看技能

quantizing-models-bitsandbytes

其他

这个Skill使用bitsandbytes库量化大语言模型,能在GPU内存有限时通过8位或4位量化减少50-75%内存占用,同时保持精度损失最小。它支持INT8、NF4、FP4等多种量化格式,可与HuggingFace Transformers无缝集成,适用于需要部署更大模型或加速推理的场景。还提供QLoRA训练和8位优化器支持,让开发者能轻松实现高效模型压缩。

查看技能

dispatching-parallel-agents

其他

该Skill用于并行处理3个以上无依赖关系的独立故障,可为每个问题域分派专属Claude代理同时执行调查修复。它通过并发处理多个独立问题显著提升故障排查效率,特别适用于测试文件、子系统等无共享状态的场景。

查看技能