# Measurement & Analytics Framework — AI US Growth Agency

## Data Architecture

```
Shopify (orders, products, customers)
    │
    ├──→ GA4 (web behavior, conversion events)
    │       │
    │       └──→ Looker Studio / Dashboard
    │
    ├──→ Meta Ads (impressions, clicks, conversions via CAPI)
    │       │
    │       └──→ Meta Attribution
    │
    ├──→ Google Ads (impressions, clicks, conversions via GTag)
    │       │
    │       └──→ Google Attribution
    │
    └──→ Warehouse / Spreadsheet (merged truth)
            │
            └──→ SKU-level P&L
```

## Core Metrics

### Primary (Customer Economics)
| Metric | Formula | Target | Frequency |
|--------|---------|--------|-----------|
| New Customer CAC | Total ad spend / New customers acquired | <$50 | Weekly |
| New Customer Revenue % | New customer revenue / Total revenue | >60% | Weekly |
| MER | Total revenue / Total ad spend | >3.0x | Weekly |
| Contribution Margin (pre-ad) | Revenue - COGS - shipping - transaction fees | Track | Weekly |
| Contribution Margin (post-ad) | CM pre-ad - ad spend | >25% | Weekly |
| First-order Payback | CAC / Monthly contribution per customer | <60 days | Monthly |
| 30/60/90 Day Repurchase | Cohort retention rates | Track | Monthly |

### Secondary (Platform Diagnostics)
| Metric | Platform | Use |
|--------|----------|-----|
| Platform ROAS | Meta, Google | Diagnostic only — NOT primary KPI |
| CTR | Both | Creative quality indicator |
| CPM | Both | Audience/competition indicator |
| Frequency | Meta | Ad fatigue warning |
| Impression Share | Google | Budget constraint indicator |

## Attribution Rules

### Why Platform ROAS is NOT the Primary Metric
- Meta and Google both claim credit for same conversions
- Platform attribution windows differ (7-day click vs 1-day view)
- Neither platform sees organic, email, or direct revenue
- Incrementality ≠ platform-reported conversions

### MER as Truth Anchor
MER (Marketing Efficiency Ratio) = Total Revenue / Total Ad Spend
- Cannot be gamed by attribution windows
- Shows the full picture of paid + organic
- Simple enough for founders to understand

## Data Quality Checklist

### Weekly
- [ ] Conversion API (CAPI) events firing correctly
- [ ] GA4 purchase event matches Shopify order count (±5%)
- [ ] New vs returning customer flag accurate
- [ ] UTM parameters preserved through checkout
- [ ] Product feed updated (Google Merchant Center)

### Monthly
- [ ] SKU cost data refreshed
- [ ] Shipping cost averages updated
- [ ] Transaction fee rates verified
- [ ] Return/refund rate calculated
- [ ] Attribution overlap between Meta and Google estimated

## SKU Contribution Margin Model

```
Revenue per unit: $X
- COGS: ($X)
- Shipping: ($X)
- Transaction fees: ($X)
- Packaging: ($X)
= Contribution Margin (pre-ad): $X

- Ad cost per order: ($X)  ← weighted average CAC
= Contribution Margin (post-ad): $X
```

## Incrementality Test Design

### Geo-Holdout Test (Quarterly)
1. Select 2-3 comparable geographic regions
2. One region: pause all paid ads for 2 weeks
3. Compare organic/direct revenue delta
4. Calculate true incrementality: (test revenue - holdout revenue) / ad spend

### Facebook Conversion Lift (Optional)
- Requires minimum $30K/test budget
- Platform-native A/B test with holdout group
- Valid for Meta-only incrementality

## Data Gaps & Assumptions

| Data Point | Status | Action |
|-----------|--------|--------|
| COGS per SKU | Must be client-provided | Include in onboarding questionnaire |
| Shipping cost avg | Must be client-provided | Include in onboarding questionnaire |
| Return rate | Available from Shopify | Pull via API |
| LTV (3+ months) | Not available for new clients | Estimate after 90 days |
| Cross-device attribution | Not measurable without identity graph | Acknowledge as limitation |