Skill 详情

meta-ads-analyzer

Highly relevant, but limited to performance analysis and diagnosis rather than full Meta Ads work.

匹配类型可能匹配已针对 meta 广告 审核
来源ideola-ai/marketing-ads-skills外部来源
报告安装量89仅表示受欢迎程度

使用前先检查

自动化审核只检查相关性,不代表安全审查或推荐。使用前请阅读来源中的说明。

已保存的来源预览

SKILL.md

这段内容是审核时保存的快照。外部来源才是完整且最新的版本。

---
name: meta-ads-analyzer
description: Provides expert-level analysis and diagnosis for Meta Ads campaigns. Use this skill to interpret performance data, identify root causes of issues, and generate actionable recommendations, with a special focus on correctly handling the 'Breakdown Effect'.
---

# Meta Ads Analysis & Diagnosis Skill

## When to Use This Skill

Use this skill when you need to **analyze and diagnose Meta Ads campaign performance**, including:
- Interpreting campaign, ad set, or ad-level performance data
- Identifying root causes of performance issues
- Generating actionable optimization recommendations
- Understanding why Meta's system makes certain budget allocation decisions

## Result Recommendations (MANDATORY for Final Reports)

> **IMPORTANT:** The following rules are **MANDATORY** and **MUST be strictly followed** when writing the final analysis report. These are not optional guidelines—they define the required standards for all deliverables.

- **NEVER recommend pausing or reducing budget for any segment based solely on higher average CPA/CPM in breakdown reports.** Higher average cost does NOT mean poor performance—it often reflects the system capturing low *marginal* cost opportunities earlier. Removing segments may increase overall costs. Always frame changes as testable hypotheses, not directives.
- **ALWAYS justify recommendations with data evidence, Meta's system mechanics, and expected impact on *overall campaign performance*.**
- **EVERY insight must include data evidence and explanation.** Every recommendation must be actionable and verifiable.
- **USE QUALIFIED LANGUAGE.** Say "Estimated Reach of ~1,000" not "You reached 1,000 people." All metrics are estimates.
- **ALIGN WITH OFFICIAL RECOMMENDATIONS.** Check `get_recommendations` API first. If diverging, explicitly acknowledge and explain why.
- **Disambiguate clicks.** Never use the term "clicks" alone. Use "Clicks (all)" for total interactions (likes, shares, page clicks, link clicks) or "Link Clicks" for clicks that lead offsite; these are distinct metrics with different meanings.**

## Metric Naming Guidelines

**IMPORTANT:** Always rename metric names to standardized, non-sensitive names exactly as specified below in all responses:

| Raw Metric Name | Standardized Display Name |
| :--- | :--- |
| `impressions` | Impressions |
| `video_thruplay_watched_actions` | ThruPlays |
| `clicks` | Clicks (all) |
| `purchase_roas` | Purchase ROAS (return on ad spend) |

## Core Principles

- **Holistic First:** Evaluate at aggregate level before drilling down. The system optimizes for the whole, not the parts.
- **Dynamic over Static:** Analyze performance over time, not single snapshots.
- **Marginal over Average:** The system prioritizes marginal CPA (cost of the *next* result), not average CPA. A higher average CPA segment might be preventing even higher marginal costs elsewhere.

## Meta Ads Domain Knowledge

### Legal Requirements & Terminology

- **Audience Size Metrics:** Due to legal requirements, when referring to audience size metrics (the total number of accounts that view the ad), you must use "Accounts Center accounts" (case insensitive) or report the metric without any unit instead of "people" (e.g., "Your reach is 17,000" or "Your ad reached 17,000 Accounts Center accounts").
- **"People" Usage:** When "people" is used in contexts referring to audience size (e.g., "reach more people", "seen by people", "your ad reached people", "show ads to people"), the word "people" must be replaced with "Accounts Center accounts". When "people" is used with specific actions, behaviors, characteristics, or qualifiers that describe what people do or are likely to do, "people" is acceptable.

### Campaign & Performance Definitions

- **Conversion Ads:** Ad entities with objectives like Lead, Sales, or App Promotions are categorized as conversion ads.
- **Conversion Rate:** Conversion rate = conversion / impression.
- **Performance Indicators:** L
在 GitHub 阅读完整来源 (打开外部页面)
相关上下文

相关工作