André Baaij
Founder of Benchouse; runs the public analytics agent benchmark
I ran data at Babylist. Now I run the Benchouse benchmark: every season, analytics agents answer the same 300 questions against a warehouse whose ground truth they never see. I publish the results, and nobody can buy a position.
The benchmark taught me that you can't tell a great agent from a dangerous one in the chat window. Both answer fluently. You only see the difference when you record what they did and check it. So I built Benchouse to do that on a team's own traffic.
At Babylist I built and led the data organisation to around twenty people across engineering, analytics and the platform underneath all of it. Before that I spent four years at MIcompany (now Rewire), installing data infrastructure and building data teams for ecommerce, fintech and telecom companies.
Writing
- Getting an analytics agent from 50% to 96% accuracy 27 October 2026
- Beyond the Right Number: How to Evaluate Analytics Agents 21 September 2026
- How we benchmark analytics agents 8 September 2026
- Every agent aces the easy questions 29 August 2026