Automating Ad Campaigns Across 4 Platforms

The situation

Universal Music Group's marketing team was creating ad campaigns manually across Meta, Google Ads, TikTok, and Airtable. Each campaign meant logging into 4 different platforms, copying creative assets, setting targeting parameters, and configuring budgets — separately, for each platform, every time. It took hours per campaign and was prone to human error.

What I built

I solo-architected and built an automation platform that unified the entire workflow. The team defines a campaign once — creative, targeting, budget — and the platform pushes it to all 4 advertising platforms simultaneously. What used to take hours became instant.

The backend is a Django REST API designed with an extensible provider pattern. Each advertising platform (Meta, Google, TikTok) is a pluggable provider module — adding a new platform means writing one adapter, not touching core logic. Airtable serves as the campaign data source that the marketing team was already comfortable with.

Campaign execution runs through event-driven Celery tasks — each platform push is an independent async job with retry logic and error handling, so a failure on one platform doesn't block the others.

The whole system shipped with 90% test coverage and was deployed on AWS with enterprise security and rate limiting.

Why this approach

The provider pattern was the key architectural decision. UMG's advertising needs would inevitably expand — new platforms, new campaign types, new targeting options. Building a monolithic integration would've meant rewriting core logic every time something changed. The provider pattern means the team (or another developer) can extend the system without understanding or modifying the internals.