L’interface est localisée ; les présentations et documents des Skills sont traduits automatiquement par IA dans votre langue. Le code, les commandes, chemins et termes techniques restent en version originale, et le document de l’auteur fait foi.
Explorer les Skills
Productivité · SKILL.md

ttl-marketing-report-agent

ttl-marketing-report-agent

Generate comprehensive marketing reports, calculating ROI, CPA, and other key financial and performance metrics for Texas Total Loss.

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentation de la Skill

GitHub

name: ttl-marketing-report-agent description: Generate comprehensive marketing reports, calculating ROI, CPA, and other key financial and performance metrics for Texas Total Loss. trigger: > Use this when asked for marketing reports, ROI analysis, cost per lead summaries, or financial performance of specific campaigns or the site as a whole.

Goal

Provide accurate, data-driven financial and performance reports to guide marketing investment decisions.

Instructions

  1. Gather Data
    • Aggregate cost data from docs/campaigns/ (look for spend or budget fields).
    • Aggregate conversion/lead data from docs/marketing-insights.md or data/analytics/.
    • If available, pull revenue/case-value data from Supabase or manual logs.
  2. Calculate Key Metrics
    • CPA (Cost Per Acquisition): Total Spend / Number of Leads.
    • ROI (Return on Investment): (Total Revenue - Total Spend) / Total Spend.
    • Conversion Rate: Total Leads / Total Traffic.
  3. Analyze Trends
    • Compare performance across different cities (Dallas vs. Houston vs. San Antonio).
    • Identify high-performing vs. low-performing channels (Organic, Paid, Email).
  4. Generate Report
    • Create a dated report in docs/reports/marketing-roi-<date>.md.
    • Use tables and concise summaries for easy reading.
  5. Suggest Allocations
    • Based on ROI, suggest where to increase or decrease marketing spend for the next period.

Constraints

  • Clearly label "estimates" vs. "verified data".
  • Do not invent data; if a metric is missing, report it as "N/A" and suggest how to start tracking it.
Skills enregistrées