indranilbanerjee/digital-marketing-pro / skills/funnel-architect
funnel-architect
VerifiedDesign a funnel for the business model: stages, KPIs, journey. Existing data → funnel-audit. "design a funnel for our SaaS"
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 15.6 KB |
| attribution-models.md | markdown | 6.0 KB |
| funnel-templates.md | markdown | 5.6 KB |
| gap-analysis.md | markdown | 5.1 KB |
| journey-mapping.md | markdown | 5.4 KB |
| sales-marketing-alignment.md | markdown | 17.4 KB |
attribution-models.md
View on GitHub# Attribution Models — Comparison & Implementation Guide
## Model Comparison
| Model | How It Works | Best For | Limitation |
|-------|-------------|----------|-----------|
| **Last-Click** | 100% credit to last touchpoint | Short sales cycles, direct response | Ignores awareness/consideration |
| **First-Click** | 100% credit to first touchpoint | Brand awareness campaigns | Ignores nurture/conversion steps |
| **Linear** | Equal credit to all touchpoints | Balanced overview | Over-credits low-impact touches |
| **Time-Decay** | More credit to touches closer to conversion | Long sales cycles, B2B | Under-credits awareness |
| **Position-Based (U-shape)** | 40% first, 40% last, 20% split among middle | Balanced with emphasis on intro/close | Somewhat arbitrary splits |
| **Data-Driven** | ML model allocates based on actual impact | Accounts with steady conversion volume (Data volume decides how stable the model is; Google publishes no fixed minimum for data-driven attribution, so neither do we) | Unstable on thin data |
| **Marketing Mix Modeling** | Econometric model using aggregate data | Budget allocation across channels | Slow, requires historical data |
---
## Model Selection Decision Tree
```
START: Where is attribution computed?
├── In GA4 or Google Ads → only data-driven and last click exist there
│ ├── Steady conversion volume? → Data-driven, compared with last click
│ └── Thin or lumpy volume? → Last click, and read data-driven as directional
├── In your own warehouse or CRM → any model, including position-based and time-decay
│ ├── Short sales cycle (< 7 days)? → Last-click or linear
│ └── Long sales cycle (> 30 days)? → Time-decay or W-shaped (needs CRM stages)
└── Budget planning across channels → add MMM; no click-path model answers that alone
BUSINESS MODEL OVERRIDES:
- eCommerce / DTC → Last-Click baseline, upgrade to Data-Driven when possible
- B2B SaaS → Position-Based or Time-Decay (long cycles, many touches)
- Local Business → Last-Click (simple journeys)
- Marketplace → Separate attribution for supply and demand sides
```
---
## Platform Implementation Guides
### Google Analytics 4 (GA4)
- **Default**: Data-driven attribution (cross-channel)
- **Configurable models**: Since 2023, GA4's Admin → Attribution Settings exposes **only `data-driven` and `last-click` (paid & organic channels)** — the old menu of linear / time-decay / position-based / first-click was removed. To apply linear, time-decay, position-based, or any custom credit rule, model it in your **warehouse / BI layer** (BigQuery export + SQL, or a BI tool) — not in GA4's UI.
- **Lookback windows**: 30 days for acquisition, 90 days for other conversions
- **Reports**: Advertising → Attribution → Model comparison (compares the two available models)
- **AI Assistant channel**: GA4's default channel grouping now includes an **"AI Assistant"** channel that isolates referral traffic from AI assistants (ChatGPT, Gemini, Copilot, Perplexity, etc.). Include it in channel breakdowns so AI-sourced conversions are not misfiled under "Referral" or "Direct".
- **Limitation**: Only tracks Google-visible touchpoints, cannot see all walled garden data
### Meta (Facebook) Attribution
- **Default**: 7-day click, 1-day view attribution
- **Conversions API (CAPI)**: Server-side event tracking for better match rates
- **Aggregated Event Measurement**: For iOS 14+ tracking limitations
- **Recommended**: Configure CAPI + browser pixel for maximum data coverage
- **Compare**: Meta's self-reported conversions vs GA4's cross-channel view
### Google Ads
- **Default**: Data-driven for most conversion actions ("The Data-driven model is the default attribution model for most conversion actions", support.google.com/google-ads/answer/6259715, undated, read 2026-10-27). Last click is still available; first click, linear, time decay and position-based are no longer supported
- **Compare**: put data-driven beside last click to see which campaigns assist conversions
- **Cross-campaign**: Attribution applies across Search, Display, YouTube, Shopping
- **Recommendation**: Enable data-driven attribution, set appropriate conversion windows
---
## Cross-Device Attribution
### Challenges
- Same user, different devices appears as multiple users
- Cookie-based tracking breaks across devices
- Privacy regulations limit cross-device linking
### Solutions
1. **Deterministic matching**: Logged-in user IDs across devices (most accurate, requires auth)
2. **Probabilistic matching**: Statistical models linking device patterns (less accurate)
3. **Google Signals**: Cross-device data from logged-in Google users in GA4
4. **CRM integration**: Match conversions to known contacts across touchpoints
---
## Custom Model Design Framework
When standard models don't fit:
1. **Define touchpoint categories**: Awareness, Engagement, Conversion Assist, Conversion
2. **Assign category weights** based on business model:
- B2B SaaS: Awareness 20%, Engagement 30%, Assist 20%, Conversion 30%
- eCommerce: Awareness 15%, Engagement 15%, Assist 20%, Conversion 50%
3. **Test against actual outcomes**: Compare model predictions to observed patterns
4. **Iterate quarterly**: Adjust weights based on incrementality test results
---
## Attribution Validation
### How to Verify Your Model Is Accurate
1. **Holdout tests**: Pause a channel, measure true impact vs model-predicted impact
2. **Incrementality tests**: Geo-split or user-split experiments per channel
3. **Cross-model comparison**: Run 2-3 models in parallel, compare conclusions
4. **Sanity checks**: Does the attribution match what you intuitively know about channel performance?
5. **Revenue reconciliation**: Do attributed conversions sum to actual revenue?
### Red Flags
- A single channel claims >80% of conversions (likely measurement bias)
- Model suddenly changes attribution dramatically (check for tracking issues)
- Branded search gets majority credit (it's capturing demand, not creating it)
- Model ignores channels you know drive awareness (review touchpoint capture)
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |