campaign-retrospective-analyst

What This Skill Does

The campaign-retrospective-analyst skill is designed to transform raw advertising data from major platforms into actionable business intelligence. It automates the complex task of multi-channel performance analysis, providing a structured framework for root-cause identification and future campaign planning. Rather than providing generic advice, this skill dissects performance metrics—such as ROAS, CPA, and funnel conversion rates—across Meta, Google, TikTok, YouTube, Amazon, and DSPs, ensuring that your growth decisions are backed by rigorous data review.

Installation

To integrate this skill into your environment, run the following command via your terminal: clawhub install openclaw/skills/skills/danyangliu-sandwichlab/campaign-retrospective-analyst

Use Cases

This skill is ideal for performance marketers, growth leads, and agency managers looking to scale ad spend efficiently. It excels at:

• Performing deep-dive root-cause analysis when ROAS dips unexpectedly.
• Creating data-driven post-mortem reports for quarterly business reviews (QBRs).
• Evaluating whether to increase or decrease budget allocations across different funnel stages.
• Identifying winning creative patterns versus underperforming audience segments.
• Validating cross-platform performance parity to optimize media mix modeling.

Example Prompts

• "Analyze our Meta and TikTok performance for the last 30 days. Why has our CPA increased by 15% even though spend remained flat? Provide a breakdown by creative asset."
• "Compare our Amazon Ads ROAS against Google Shopping intent-based campaigns for the holiday season. Which platform is driving higher quality traffic at a better marginal cost?"
• "Our DSP campaigns saw high frequency but low conversion this month. Perform a retrospective analysis to determine if audience saturation or creative fatigue is the root cause."

Tips & Limitations

To get the best results, always define your metric scope and time range clearly. The skill functions best when provided with specific attribution windows; if you leave these blank, it will default to the platform-specific standards. Remember that this tool is directional when sample sizes are low. Always check the provided confidence score before scaling spend based on recommendations. The tool will explicitly flag if it suspects tracking breakages or data discrepancies, so pay close attention to the 'Interpretation and Caveats' section of its output. Use the rollback conditions provided in every recommendation to mitigate financial risk.