Advisor Marketing Attribution Analyst

An autonomous agent that designs and implements marketing attribution models — tracking customer journeys across touchpoints, attributing conversions to channels, and optimizing marketing spend with data-driven ROI analysis.

When to Use This Agent

Choose Advisor Marketing Attribution Analyst when:

• You need to understand which marketing channels drive conversions
• You want to implement multi-touch attribution beyond last-click
• Marketing spend allocation needs data-backed optimization
• You need to track customer journeys across online and offline touchpoints

Consider alternatives when:

• You need campaign creative development (use a content marketer agent)
• You need analytics dashboard setup (use a data analyst agent)
• You need general marketing strategy without attribution focus

Quick Start

Copy
# .claude/agents/attribution-analyst.yml
name: advisor-marketing-attribution-analyst
description: Design marketing attribution models and optimize spend
agent_prompt: |
  You are a Marketing Attribution Analyst. Help with:

  1. Design multi-touch attribution models
  2. Implement tracking across channels and touchpoints
  3. Build customer journey maps from touchpoint data
  4. Attribute conversions to marketing channels
  5. Calculate channel-level ROI and ROAS
  6. Recommend budget reallocation based on attribution data

  Attribution principles:
  - Multi-touch > single-touch (last-click misattributes)
  - Track the full journey, not just the conversion
  - Attribution is a model, not truth — use multiple models
  - Incremental lift testing validates attribution models

Example invocation:

Copy
claude "Design an attribution model for our B2B SaaS with a 90-day sales cycle across paid ads, content marketing, email, and events"

Sample attribution output:

Attribution Model Design — B2B SaaS (90-day cycle)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Recommended Model: Time-Decay Multi-Touch

Typical Customer Journey (5-8 touchpoints):
  Day 1:  Google Ad click (blog post)           [5%]
  Day 7:  Organic search (product page)          [8%]
  Day 14: Email nurture (case study)             [10%]
  Day 30: LinkedIn ad (demo video)               [12%]
  Day 45: Webinar attendance                     [15%]
  Day 60: Direct visit (pricing page)            [18%]
  Day 75: Sales email (proposal)                 [15%]
  Day 90: Direct visit (sign-up) ← conversion   [17%]

Channel Attribution Summary (last quarter):
  Channel           Revenue Attributed  ROAS   Recommendation
  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  Google Ads        $245K               3.2x   Maintain
  Content/SEO       $380K               8.5x   Increase 25%
  Email Nurture     $290K               12.1x  Increase 15%
  LinkedIn Ads      $180K               2.1x   Optimize targeting
  Events/Webinars   $320K               4.0x   Maintain
  Direct/Brand      $185K               N/A    Brand investment

Recommended Budget Shift:
  Move 20% of LinkedIn budget → Content/SEO
  Expected impact: +$85K attributed revenue per quarter

Core Concepts

Attribution Model Comparison

Model	How It Works	Best For	Limitation
Last-Click	100% to final touchpoint	Simple reporting	Ignores awareness
First-Click	100% to initial touchpoint	Awareness measurement	Ignores nurture
Linear	Equal credit to all touchpoints	Even distribution	No time weighting
Time-Decay	More credit to recent touchpoints	Long sales cycles	Undervalues awareness
Position-Based	40% first + 40% last + 20% middle	Balanced view	Arbitrary weights
Data-Driven	ML-based credit allocation	High-volume B2C	Requires large datasets

UTM Tracking Implementation

Copy
// Consistent UTM parameter framework
const UTM_STANDARDS = {
  source: {
    // Where the traffic comes from
    values: ['google', 'linkedin', 'twitter', 'email', 'direct'],
    required: true
  },
  medium: {
    // Marketing medium
    values: ['cpc', 'organic', 'social', 'email', 'referral', 'event'],
    required: true
  },
  campaign: {
    // Specific campaign name
    format: '{year}-{quarter}-{campaign-name}',
    example: '2025-q2-developer-productivity',
    required: true
  },
  content: {
    // Ad variant or content piece
    format: '{type}-{variant}',
    example: 'blog-post-a',
    required: false
  }
};

// First-touch attribution capture
function captureFirstTouch() {
  if (!localStorage.getItem('firstTouch')) {
    localStorage.setItem('firstTouch', JSON.stringify({
      source: getUTM('source'),
      medium: getUTM('medium'),
      campaign: getUTM('campaign'),
      timestamp: new Date().toISOString(),
      landingPage: window.location.pathname
    }));
  }
}

ROI Calculation Framework

Channel ROI = (Attributed Revenue - Channel Cost) / Channel Cost × 100

Example:
  Content Marketing:
    Annual spend: $50,000 (writers, tools, distribution)
    Attributed revenue: $425,000 (time-decay model)
    ROI: ($425K - $50K) / $50K × 100 = 750%

  Google Ads:
    Annual spend: $200,000
    Attributed revenue: $640,000 (time-decay model)
    ROI: ($640K - $200K) / $200K × 100 = 220%

  Despite higher absolute revenue, Google Ads has
  lower ROI than Content Marketing per dollar spent.

Configuration

Option	Type	Default	Description
model	string	"time-decay"	Attribution model to apply
lookbackWindow	number	90	Days to look back for touchpoints
channels	string[]	["paid", "organic", "email", "social"]	Marketing channels to track
conversionEvents	string[]	["signup", "demo", "purchase"]	Events to attribute
utmEnforcement	boolean	true	Enforce UTM standards
reportFrequency	string	"monthly"	Attribution report cadence

Best Practices

1. Use multiple attribution models and compare — No single model tells the complete truth. Run time-decay, position-based, and linear models on the same data. If all three agree that content marketing deserves more budget, the signal is strong. If they disagree significantly, investigate why before making budget changes.

2. Implement first-touch AND last-touch tracking from day one — Even if you plan to use multi-touch attribution later, capture first-touch data immediately. First-touch data cannot be reconstructed retroactively. Store the initial UTM parameters and landing page for every visitor, and the last touchpoint before conversion.

3. Validate attribution with incrementality testing — Attribution models estimate channel value, but incrementality tests prove it. Run holdout experiments: turn off a channel for a geographic region and measure the lift in conversions in the control region. This confirms whether the attributed conversions are truly incremental.

4. Account for the full sales cycle in B2B — A B2B customer who converts today may have been influenced by a blog post 6 months ago. Set lookback windows to match your average sales cycle length (90 days minimum for B2B SaaS). Short lookback windows over-credit bottom-funnel channels and under-credit awareness.

5. Standardize UTM parameters across teams — If paid ads use source=google_ads while the content team uses source=google, attribution breaks. Create a UTM naming convention document, enforce it with URL builders, and audit UTM usage quarterly. Inconsistent UTMs are the most common cause of attribution data quality issues.

Common Issues

Last-click attribution over-credits brand and direct traffic — Customers who already know your product visit your site directly to convert. Last-click attributes 100% of the conversion to "direct," ignoring the Google Ad, email nurture, and webinar that built awareness. Multi-touch attribution and first-touch tracking solve this, but require implementation effort.

Cross-device journeys break attribution tracking — A user clicks a Google Ad on mobile, reads emails on desktop, and converts on their work laptop. Each device creates a separate tracking session. Use authenticated user tracking (log in to your product ties devices together) and Google Analytics cross-device reports to stitch journeys across devices.

Attribution data conflicts with platform-reported conversions — Google Ads reports 100 conversions, Facebook reports 80, and your attribution model attributes 120 total — but you only had 90 actual conversions. Each platform takes full credit for any conversion it touched. Your attribution model is the source of truth because it deduplicates across platforms and applies fractional credit.