An intelligent advertising case analyzer that diagnoses delivery issues like volume drops or high costs and generates structured Redoc analysis reports.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ads-case-analyzer
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install ads-case-analyzer using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is ads-case-analyzer?
The ads-case-analyzer is a sophisticated diagnostic tool designed for the content consumption advertising industry. By processing campaign, advertiser, and brand IDs, it automatically pulls post-delivery and bidding chain data to perform deep-dive funnel analysis and root cause inference. It serves as a specialized assistant for performance marketing teams who need to understand why specific campaigns are underperforming or why account consumption has suddenly decreased.
This skill leverages Openclaw Skills infrastructure to bridge the gap between raw data logs (like Hive and ClickHouse) and actionable business insights. It automates the extraction of core metrics across multiple dimensions—including bidding ratios, subsidy coefficients, and predicted CTR/CVR—to identify exactly where a campaign is losing its competitive edge in the auction process.
ads-case-analyzer Use Cases
- Investigating sudden drops in daily ad spend or impression volume.
- Diagnosing high Cost Per Action (CPA) and identifying model overestimation via PAOA analysis.
- Analyzing internal brand competition where multiple accounts use the same creative assets.
- Evaluating overall account health and identifying performance gaps between different placements like feed and search.
- Troubleshooting low retention rates or conversion rate (CVR) calibration issues.
How ads-case-analyzer Works
- Input Resolution: Parses the provided advertiser, campaign, or brand IDs and determines the required data filtering logic.
- Data Retrieval: Pulls daily granularity metrics from post-delivery logs and hourly bidding chain data from offline and online trace logs.
- Funnel & Auction Analysis: Executes a step-by-step breakdown of the advertising funnel (Recall -> Pre-ranking -> Ranking -> Auction) to find where impressions are being filtered.
- Root Cause Inference: Uses LLM reasoning to correlate data signals (e.g., bid_ratio drops, pcvr fluctuations) with specific delivery issues.
- Report Generation: Automatically calls the Redoc utility to produce a structured, professional analysis document containing core conclusions and recommended actions.
ads-case-analyzer Setup
To use this skill within the Openclaw Skills environment, ensure you have the necessary system dependencies and database access.
# Ensure dependencies are installed
sudo apt-get install curl jq python3
# The skill requires access to internal Hive/DWS tables
# Ensure your environment has the appropriate data permissions
ads-case-analyzer Data Schema & Taxonomy
The skill organizes data from several key sources to build its diagnostic model:
| Data Source | Description | Key Metrics |
|---|---|---|
ads_data_common.dws_ad_full_log |
Post-delivery performance | Imp_cnt, click_cnt, fee, advv, conversions |
bidding_cube_trace_log_hourly |
Offline bidding parameters | bid_ratio, given_cpa, subsidy_ratio |
bidding_pivot_trace_log_feed/search |
Online real-time bidding | pcvr, pctr, pgmv, cpaBid, request_cnt |
adtrace_log_checkpoint_rt |
Funnel stage analysis | is_pass, filter_reason, request_label |
Results are output as a Redoc Markdown document following a structured template (TL;DR, Trend Analysis, Bidding Chain, Root Cause, Recommendations).
ads-case-analyzer Advanced Features
- Internal Brand Competition Analysis: Detects if other advertisers under the same brand are 'stealing' volume using identical notes.
- Multi-Dimensional Drill-down: Automatically switches aggregation levels based on input (e.g., Advertiser -> Campaign -> Ad ID).
- Automated PAOA Calculation: Compares predicted vs. actual conversions to identify model bias in high-cost scenarios.
- Calibration Gap Analysis: Identifies discrepancies between original pcvr, calibrated pcvr, and actual conversion rates.
SKILL.md
Loading
Related Openclaw Skills
Parallel Coding with Git Worktree
Enable multi-tasking by deploying multiple AI coding agents across isolated git worktrees within a single repository.
Obsidian CLI
A powerful command-line interface to interact with a running Obsidian instance for note management and plugin development.
Xianyu Auto Ops
A professional bilingual workflow for generating high-conversion Xianyu listings, batch processing SKU data, and managing buyer communications.
AWS China What's New
A specialized tool for querying AWS China service announcements across Beijing and Ningxia regions.
Chery Information Assistant
A professional AI agent skill designed to retrieve real-time brand intelligence, product specifications, and financial performance data for Chery.
Task Quest
A seamless gamification layer that transforms routine task completion into an engaging experience with XP, levels, and achievements.
METADATA
Featured*
