The Business Problem

A hospitality AdTech company needed to optimize pricing for targeted hotel ad placement across billions of auctions. The challenge involved processing massive auction data (5 TB) and determining optimal bid prices based on complex features.

Additionally, identifying the most relevant websites for ad placement required scraping and classifying hundreds of thousands of sites at scale.

The INM Consulting Approach

We developed and deployed a comprehensive ML-based pricing and placement optimization system, handling massive scale data processing and deployment.

Implementation Details

• Queried, grouped and aggregated auction data from billions of hotel ad auctions (5 TB in size) using BigQuery
• Analyzed and modeled in Python to calculate optimal bid pricing
• ML model optimized bids based on hotel ad features and user-level data tracked using cookies
• Scraped hundreds of thousands of websites and performed content classification
• Identified most relevant sites for ad placement
• Deployed using Airflow for orchestration
• Results visualized using Looker

Technologies Used