Analyze Campaign Performance with Data Visualization

This skill provides expert marketing campaign performance dashboard design, covering data visualization, KPI tracking, attribution modeling, and actionable insight generation across the full campaign measurement ecosystem from data collection to executive reporting. It organizes metrics into four tiers - primary KPIs (ROI, ROAS, customer acquisition cost, conversion rate), secondary KPIs (CTR, CPC, engagement rate, lead quality score), operational metrics (impressions, reach, frequency, budget utilization), and attribution metrics (first-touch, last-touch, multi-touch values) - and structures dashboards across three audience levels: an executive summary (high-level ROI, spend efficiency, goal progress), a campaign manager view (channel performance, optimization opportunities, budget allocation), and tactical analysis (creative performance, audience segments, keyword/ad-group detail). It covers a multi-channel data pipeline that standardizes metric names across platforms like Google Ads and Facebook, multi-touch attribution models (time-decay, linear, u-shaped weighting), Plotly-based visualization (ROAS trend with spend overlay, channel comparison, conversion funnel, budget allocation), real-time alert thresholds for metrics like ROAS drops and CPC spikes, statistical significance testing (chi-square for conversion rates, t-test for continuous metrics like CPC/ROAS/CTR), data-quality validation queries, and rule-based optimization insight generation such as budget-increase or creative-refresh recommendations.

Use this skill when building a campaign performance dashboard that needs to serve multiple audiences - executives, campaign managers, and tactical analysts - from one underlying data pipeline, with attribution modeling and statistical rigor rather than raw metric dumps. It gives concrete performance targets: a 5-second dashboard load goal, data sampling above 1M rows, a 15-minute cache refresh cycle for frequently used metrics, and a 5-second rule for surfacing critical insights, plus a consistent color scheme (green for positive, red for negative, blue for neutral) and drill-down from summary to detail. It is not a guide for single-channel, single-metric reporting - its architecture assumes multiple ad platforms feeding into standardized metrics and multiple attribution models compared side by side, which is more infrastructure than a one-channel report needs.

-- Data validation queries for campaign metrics
SELECT 
    campaign_id,
    date,
    CASE 
        WHEN cost > 0 AND conversions = 0 AND clicks > 100 THEN 'No Conversions Alert'
        WHEN roas > 10 THEN 'ROAS Anomaly'
        WHEN ctr > 20 THEN 'CTR Anomaly'
        ELSE 'Normal'
    END as data_quality_flag
FROM campaign_performance
WHERE date >= DATE_SUB(CURRENT_DATE, INTERVAL 7 DAY);

Given raw multi-channel campaign data, the skill produces a CampaignDataProcessor class that standardizes metrics across platforms and calculates CTR, CPC, ROAS, and conversion rate, a multi-touch attribution function returning per-touchpoint weights, a Plotly subplot dashboard covering ROAS trend, channel performance, conversion funnel, and budget allocation, a JavaScript alert-threshold checker flagging ROAS drops, CPC spikes, conversion drops, and budget pacing issues by severity, a chi-square/t-test significance-testing function for campaign variants, the data-quality query shown above, and a rule-based insight generator that flags budget-reallocation opportunities (high ROAS plus high budget utilization) and creative fatigue (CTR declining 30%+ week over week).

Marketing analysts and growth teams building a campaign dashboard that needs to serve executives, campaign managers, and tactical analysts from the same data pipeline, with attribution modeling and statistical significance testing rather than vanity metrics. It suits teams running multiple channels (Google Ads, Facebook, LinkedIn, email, display) who need standardized cross-platform metrics, automated executive reporting with month-over-month trends and goal-variance analysis, and scheduled, role-personalized report delivery.