Making the case for Brevo's customer data platform
Brevo had not chartered a CDP. I found the opportunity, made the case layer by layer, and built the tribe that shipped it.
+$10M
ARR (15%+ revenue lift)
3 squads
25 engineers, 3 PM reports
~90K
unified profiles resolved daily
1M events to 90K peopleIdentity resolution across 30 sources, resolved daily.
30 sources supplied the signals.
Identity resolution processed roughly 1M events per day.
The result was about 90K unified profiles per day.
Sources, step 1 of 3
The opportunity
Brevo was a $200M+ ARR SMB tool with 500K customers across Europe, the US and LATAM. The data needed to serve enterprise buyers existed, but it was scattered across 30 sources and nobody could act on it.
The product
I built identity resolution that turned roughly a million raw daily events into about 90,000 unified customer profiles a day, with a segmentation engine on top. Segmentation 2.0 lifted upsell 20% ($4M, $25M annualised) within three months of launch.
Building the case
I built the case with each layer in turn instead of pitching once at the top and waiting. It was slower, but it made the charter inevitable and helped it survive budget season.
The investment
It took nine months of unfunded discovery before a tribe existed. That was the work required to originate something nobody had asked for.
How we structure context across short-term reasoning and long-term storage in agentic loops. Defining the boundaries between RAG, vector stores, and persistent world models. — Aadhar
Reinforcement Learning from Human Feedback isn't just a technical training phase; it's the new "User Research". How to structure human feedback loops to align agentic outputs with product outcomes. — Aadhar
