LAUNCHING A NEW PRODUCT FOR A $70M+ FUNDED AI STARTUP
The Challenge
The objective was to streamline a rapidly growing advertising ecosystem across multiple active departments. With operations scaling quickly, the challenge lay in structuring high-volume data streams, unifying tracking protocols, and preventing internal audience overlaps.
We were tasked with designing a highly organized, centralized tracking architecture—updating creative frameworks and deploying a rigorous testing environment capable of validating new localized product hypotheses at scale.
The Approach
We started by heavily restructuring the data pipeline. By flawlessly synchronizing Exa.ai's CRM with advertising channels, we gained end-to-end transparency: finally identifying exactly which keywords and variants drove qualified leads, and which wasted budget.
After the first two months of optimization, we identified that Performance Max (PMAX) operations could systematically generate a high-volume, highly stable pipeline. This generated a stream of leads at ~$20 per Sign In, and a stream of direct sales at ~$120–$170 per sale—ultimately locking in a ~1.6× to ~2.3× ROAS.
Key Learnings
- Competitor Queries:
PMAX deployed on competitor-related queries produced the most positive, scalable results. - Event Optimization:
The most effective optimization trigger occurs on 'checkout initiation'. Since checkout events trigger more frequently than final purchases, the algorithm learns far more robustly when explicitly trained for the checkout milestone.
Explore More Creatives
Average CTR
Achieved robust click-through velocity after overhauling query targeting and updating creatives across the network.
Average CPC
Sustained highly economical traffic pricing by removing audience cannibalization and consolidating bidding algorithms.
Cost per Sign In
A highly optimized, stable stream of qualified sign-ins generated primarily through our dedicated PMAX infrastructure.