Amazon Ads Automation System

Overview

I built a Python-based automation system to improve how Amazon Ads campaigns were analyzed, optimized, and managed across a large ecommerce catalog. The goal was to move beyond surface-level advertising metrics like ACOS and instead evaluate campaigns based on estimated net sales, profitability, placement performance, and incremental value.

The system helped analyze performance across campaign placements such as Top of Search, Rest of Search, and Product Pages, then generated recommendations for bid and placement modifier changes.

Problem

Amazon Ads performance was being managed across a large number of campaigns with significant annual ad spend. Manual analysis was time-consuming, inconsistent, and difficult to scale.

Traditional metrics did not always tell the full story. A campaign could look inefficient from an ACOS perspective while still driving valuable incremental sales, or it could appear strong while hiding weak profitability after costs and fees.

Solution

I created a set of Python workflows that pulled campaign performance data, calculated profitability-focused metrics, and generated structured outputs for optimization.

The system included:

• Campaign and placement-level performance analysis
• Estimated net sales calculations
• Same-SKU and other-SKU sales breakdowns
• Cost and unit-based profitability estimates
• Placement modifier recommendations
• Current modifier comparisons
• Excel reports with highlighted benchmark rows
• Database inserts for tracking campaign modifier recommendations over time

Impact

This project reduced the amount of manual campaign analysis needed and made advertising decisions more consistent, scalable, and profit-focused.

It also helped shift optimization away from simple revenue or ACOS-based decisions toward a more complete view of advertising efficiency, margin, and incremental value.

What I Learned

This project strengthened my ability to combine API integrations, business logic, data analysis, and automation into a practical decision-support system. It also reinforced the importance of designing metrics that match the actual business goal rather than relying only on default platform reporting.