marketing-analytics

# Marketing Analytics Performance measurement, attribution modeling, and data-driven optimization. ## SEOSONA Data Sources | Connector | Data Available | Key Metrics | |-----------|---------------|-------------| | `gsc_connector` | Organic search | Impressions, clicks, CTR, position | | `ga4_connector` | Website traffic | Sessions, conversions, revenue, LTV | | `rank_tracker` | Keyword rankings | Position changes, quick wins | | `backlink_connector` | Link profile | DR, dofollow count, anchor text | | `technical_seo_scanner` | Site health | Issues, speed, Core Web Vitals | --- ## Core Marketing KPI Framework ### Acquisition KPIs | KPI | Formula | Target | |-----|---------|--------| | CAC (Customer Acquisition Cost) | Total marketing spend ÷ New customers | Varies by model | | CPL (Cost Per Lead) | Total spend ÷ Leads generated | < LTV × conversion rate | | Organic CTR | (Clicks ÷ Impressions) × 100 | >3% for positions 1-3 | | Traffic growth MoM | (This month - Last month) ÷ Last month | +5-15% MoM | ### Engagement KPIs | KPI | Formula | Benchmark | |-----|---------|-----------| | Bounce rate | Single-page sessions ÷ Total sessions | <60% ideal | | Session duration | Total time ÷ Sessions | >2 min for content | | Pages per session | Total pageviews ÷ Sessions | >2 pages | | Email open rate | Opens ÷ Delivered | >25% | | Email CTR | Clicks ÷ Delivered | >3% | ### Conversion KPIs | KPI | Formula | Target | |-----|---------|--------| | CVR (Conversion Rate) | Conversions ÷ Visitors | >2-5% (varies) | | ROAS | Revenue ÷ Ad spend | >3:1 minimum | | MQL-to-SQL rate | SQLs ÷ MQLs | >30% | | Trial-to-paid | Paid users ÷ Trial users | >25% | ### Retention KPIs | KPI | Formula | Target | |-----|---------|--------| | LTV (Lifetime Value) | ARPU × Average customer lifespan | >3× CAC | | Monthly churn | Churned ÷ Total at start of month | <3% MoM | | NPS | % Promoters - % Detractors | >50 | | DAU/MAU ratio | Daily active users ÷ Monthly active | >20% | --- ## Attribution Models | Model | How It Works | Use When | |-------|-------------|---------| | **Last-click** | 100% credit to last touchpoint | Direct response, short cycles | | **First-click** | 100% credit to first touchpoint | Brand awareness evaluation | | **Linear** | Equal credit to all touchpoints | Long complex journeys | | **Time-decay** | More credit to recent touchpoints | Short sales cycles | | **Position-based** | 40% first + 40% last + 20% middle | Most B2B | | **Data-driven** (GA4) | ML model based on actual paths | Best when enough data | **Attribution rule of thumb:** - Platform data is inflated (each claims credit) - Use UTM parameters on ALL external links - Compare platform data vs GA4 for truth - Blended CAC (total spend ÷ total customers) is more honest than channel CPA ### UTM Parameter Template ``` utm_source=[where traffic comes from: google, facebook, newsletter] utm_medium=[channel type: cpc, email, social, organic] utm_campaign=[campaign name: spring-launch, brand] utm_content=[ad variation: headline-a, hero-image] utm_term=[keyword for paid: seo-audit-tool] ``` Example: `?utm_source=newsletter&utm_medium=email&utm_campaign=weekly&utm_content=cta-top` --- ## Reporting Cadence | Cadence | What to Review | Who | |---------|---------------|-----| | **Daily** | Paid spend, conversion count, critical alerts | Performance team | | **Weekly** | Channel performance vs. targets, quick wins | Marketing team | | **Monthly** | Full funnel, trends, ROI by channel, optimizations | Leadership | | **Quarterly** | Strategic review, attribution, LTV cohorts, testing roadmap | All stakeholders | --- ## Campaign Performance Analysis Workflow ### Step 1: Define Success Metrics (before campaign) - Primary goal: [specific metric + target] - Time period: [start] → [end] - Baseline: [current state for comparison] ### Step 2: During Campaign — Weekly Check ``` [ ] Spend pacing (on track vs. budget?) [ ] CPA/ROAS vs. targets [ ] Top and bottom performing ads/content [ ] Audience performance breakdown [ ] Landing page conversion rate [ ] Any technical issues? ``` ### Step 3: Post-Campaign Analysis **Data to pull from SEOSONA connectors:** ```python # From 3_MEMORY/seo_exports/<domain>/ gsc_report_*.csv → Organic traffic change during campaign ga4_report_*.csv → Conversion data, traffic sources rank_tracking_*.csv → Keyword position changes ``` **Report structure:** ```markdown ## Campaign: [Name] | [Date range] ### Goal vs. Result | Metric | Goal | Result | Delta | |--------|------|--------|-------| | Organic traffic | +20% | +34% | +14% | | Conversions | 50 | 67 | +34% | ### What Worked [Specific channels + data] ### What Didn't Work [Specific issues + hypothesis] ### Next Steps [Actionable recommendations] ``` --- ## A/B Test Analysis Framework 1. **Check sample size** — Did you reach pre-determined sample? If no → preliminary only 2. **Statistical significance** — p-value < 0.05? (95% confidence) 3. **Effect size** — Is the difference meaningful for business (≥ MDE)? 4. **Secondary metrics** — Do they support primary? 5. **Guardrails** — Did anything get worse? 6. **Segment differences** — Mobile vs. desktop, new vs. returning? --- ## Funnel Analysis (SEOSONA Integration) Map Google Analytics data to funnel stages: ``` TOFU (Awareness): GSC: impressions + clicks by keyword intent GA4: traffic by source, new visitors % MOFU (Consideration): GA4: landing page conversion rate GA4: pages per session, session duration GSC: clicks on feature/solution pages BOFU (Decision): GA4: pricing page visits, trial signups GA4: checkout conversion rate GA4: demo requests RETENTION: GA4: returning visitors, LTV proxy metrics ``` **Content Gap → Traffic Gap analysis:** Run `scripts/run_full_audit.py` → Check `content_gap_*.csv` for keywords without content. --- ## Analytics Best Practices 1. **Track leading indicators, not just lagging** — Content published → Rankings → Traffic → Leads → Revenue 2. **Apples to apples** — Compare same periods (avoid holiday weeks, seasonal spikes) 3. **Statistical significance before conclusions** — Don't optimize based on 5 conversions 4. **Attribution ≠ causation** — Multiple channels deserve credit 5. **Report insights, not just numbers** — "Traffic up 20%" → "Traffic up 20% due to [cause], recommend [action]" 6. **Automate recurring reports** — Manual reports = inconsistency ## Agent Integration **Primary:** Use for all analytics and measurement tasks **Related skills:** `funnel`, `ab_testing`, `paid_ads` **Data sources:** `ga4_connector`, `gsc_connector`, `rank_tracker`