Ad Performance Decision Engine

The problem

Advertisers often struggle to interpret raw performance data, leading to emotional decisions, premature budget cuts, or wasted spend on failing creatives. Without a structured diagnostic framework, it is difficult to isolate whether a performance drop is caused by the ad creative, the landing page, or the bidding strategy.

What it does

• Analyzes multi-platform ad data (Meta, Google, TikTok, LinkedIn) to classify creatives into actionable categories: Scale, Optimize, Test, Hold, or Pause.
• Diagnoses funnel leaks by mapping metrics across impression, click, landing page, and conversion stages.
• Evaluates creative performance against business economics like target CPA, break-even ROAS, and contribution margins.
• Identifies specific bottlenecks such as weak hooks, high friction offers, or audience saturation.
• Generates structured test hypotheses for underperforming assets to isolate variables before full resets.

Why this beats prompting it yourself

This skill enforces a rigorous DATA to DECISION workflow that prevents common AI hallucinations regarding statistical significance. It uses specific diagnostic modes for different business objectives like Awareness, Leads, or Ecommerce, ensuring recommendations align with actual financial goals rather than just surface-level engagement metrics.

Use cases

• Auditing a weekly Meta Ads report to identify which high-spend creatives are actually driving profitable sales.
• Troubleshooting a high-CTR campaign that is failing to convert on the landing page.
• Determining if a new TikTok creative has sufficient data to be declared a winner or if it needs more budget.
• Prioritizing optimization tasks across a multi-ad account based on confidence levels and trend stability.

Known limitations

Does not execute API calls to ad platforms or make live budget changes. Performance outcomes are never guaranteed.

Ad Performance Decision Engine helps advertisers make better decisions about ads that are already running. It does not focus on generating ad copy. Instead, it analyzes campaign evidence and determines the most appropriate next action: 📈 SCALE — when performance and economics support continued investment 🔧 OPTIMIZE — when a specific bottleneck can be improved 🧪 TEST — when more controlled experimentation is needed 👀 HOLD — when performance is still developing ⏸️ PAUSE — when evidence shows persistent unacceptable performance 📊 GATHER MORE DATA — when the available evidence is insufficient It evaluates the complete funnel rather than blindly optimizing one metric. Core Workflow DATA → CONTEXT → DIAGNOSE → CLASSIFY → DECIDE → EXPLAIN → NEXT ACTION