Backend Automation · Case study
Taboola Campaign Bid Optimizer
A Python automation engine that takes over Taboola campaign management — analyzing site-level ROAS, adjusting bids automatically, and blocking underperforming sites, so a paid-media team stops spending five hours a day on manual optimization.
ROAS +30-50%
Automated, data-driven bidding lifted return on ad spend across the portfolio.
5+ hrs/day back
Manual bid management eliminated — the team reclaims over a full workday each week.
Underperformers auto-cut
Bad placements get blocked instantly, so ad budget stops leaking to them.
the situation
The team was managing $10,000+ a month of Taboola spend by hand. Someone sat for five-plus hours a day nudging bids, working off incomplete data, and making calls that varied by mood and fatigue. Underperforming sites stayed live too long because no one got to them in time. Good money chased bad placements.
what i built
- ▸Connected the engine to the Taboola API to pull site-level ROAS data continuously instead of by hand.
- ▸Wrote automated bid adjustment logic that reacts to performance data — no human judgement in the loop.
- ▸Built automatic site blocking so underperforming placements get cut the moment they underperform.
- ▸Added weekly reporting so the team sees the optimization story, not just the spend.
built with
PythonTaboola APISupabasePandas