Portfolio project
Paid Media Performance & Optimization Tool
A small pipeline that blends spend data across Google Ads, Meta Ads, and LinkedIn Ads, flags underperforming campaigns, tracks period-over-period trends, and writes an exec-ready summary — built with Claude Code.
The story it tells
Marketing teams normally spend hours each week pulling exports from every ad platform, reconciling column names, and writing up a summary for leadership. This tool compresses that to about two minutes and makes it repeatable every reporting period — a hands-on example of using AI to build practical marketing tools, not just talk about them.
View the reports
Two versions of the same report, generated by the same pipeline:
How it works
- Ingests and normalizes three differently-shaped platform exports into one common schema
- Flags campaigns missing a target CPA or CTR, with per-channel thresholds
- Computes period-over-period trends per campaign and per channel
- Generates a data-driven written narrative — no hard-coded copy, it changes if the data changes
- Renders a single self-contained HTML report with hand-rolled charts (no external dependencies)