Why does the model use daily snapshots?

Snapshots preserve point-in-time state to enable historical analysis, trend detection, and accurate reporting despite the limited data retention periods of platforms like Meta.

Which data sources are compatible with this scoring model?

The architecture is designed to integrate data from Google Analytics 4 (GA4), Google Ads, Meta (Facebook) Ads, and raw product feeds.

Can I customize the scoring rules?

Yes, the skill provides a framework for defining custom exclusion and inclusion rules based on specific thresholds for ROAS, revenue, view velocity, and availability rates.

How does this skill handle variant-level data?

It uses an aggregation pattern where metrics are collected at the variant level but scoring is calculated at the product (item_group_id) level, while maintaining stock overrides for specific variants.

What is the primary purpose of the Expanly Scoring Model?

The model categorizes e-commerce products into three priority tiers (High, Standard, Low) based on performance metrics to optimize bidding strategies and reduce wasted ad spend.

Expanly E-commerce Scoring Model

Name: Expanly E-commerce Scoring Model
Author: Expanly

byExpanly

Marketing Automation

Implements and manages multi-tier product scoring models to optimize e-commerce advertising spend and performance.

About

This skill provides a comprehensive framework for developers and data engineers to implement the Expanly scoring model architecture. It facilitates the creation of a three-tier priority system (Low, Standard, High) that segments products based on real-time performance signals from GA4, Google Ads, and Meta. By leveraging SQLX-based data modeling and daily snapshots, it helps automate product prioritization, reduces wasted ad spend on 'zombie' products, and ensures that dynamic remarketing campaigns focus on high-ROAS, in-stock inventory.

Key Features

Rule-based evaluation for product exclusions, inclusions, and default classifications
Three-tier priority scoring system (0, 1, 2) for automated campaign allocation
Multi-layered SQLX architecture covering raw sources, snapshots, and reporting
0 GitHub stars
Product-level aggregation logic for variant-heavy e-commerce catalogs
Automated custom label generation for Google PMAX and Meta Advantage+ campaigns

Use Cases

Identifying and scaling 'Trending' or 'Top Seller' items across multiple ad platforms
Reducing wasted ad spend by automatically down-weighting out-of-stock or low-ROAS products
Building historical performance dashboards using point-in-time product state snapshots

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add expanly/expanly-claude-code-agents expanly-scoring-model

For use in Claude.ai and ChatGPT

Download Skill

GitHub