Shipping a DeFi Funnel Pipeline in 90 Days
I was Growth Lead at Kyber Network from January to April 2026. The DeFi aggregator handles $200B+ in cumulative trading volume across 18 EVM-compatible chains. I joined to own growth for the core aggregator. The role scope shifted materially in Q1, so I shipped what I could and exited cleanly. This is what I built.
The Problem
Kyber had partial analytics coverage but no unified view of the user funnel. Acquisition, wallet connection, first trade, and retention all lived in separate tools. Nobody could answer: “Where exactly are we losing users between landing and first trade?”
What I Built
A funnel analytics pipeline in Bun/TypeScript tracking 14 data sources across 6 conversion stages. The pipeline joined on-chain signals (wallet connects, token approvals, swap executions) with off-chain signals (page views, button clicks, campaign UTMs) into one view.
The 6 stages:
- Landing (UTM-tagged traffic)
- Wallet connection
- Token selection
- Quote preview
- Swap approval
- First completed trade
Each stage measured drop-off rate, time-to-next-stage, and source attribution.
What I Found
40% drop-off between wallet connection and first trade. Users connected their wallet but never completed a swap. The bottleneck wasn’t awareness or acquisition — it was the gap between “I’m interested” and “I’ve done the thing.”
This became the team’s top onboarding priority. The pipeline gave the product team a clear, quantified target: close the wallet-to-trade gap.
The AI Automation Layer
I also automated the GTM reporting workflow using AI agents. Analytics query generation, campaign reporting, and competitive analysis that previously took hours per cycle ran in minutes. Not a novelty — a practical time multiplier for a small growth team.
The Exit
Short tenures are only failures if the output is nothing. I shipped a production analytics system in 3 months, identified a top-priority growth lever, and left clean.