SuppCo case study
Making supplement quality easy to check
I lead product design for SuppCo, now part of Function Health. My focus is TrustScore, the quality score on every product card, search result, and comparison.
At a glance
About the company Seed-stage, backed by Union Square Ventures, when I joined. Acquired by Function Health in May 2026.
SuppCo by the numbers
- App Store rating, from more than 31,000 ratings
- 4.8★
- supplement routines analyzed by the time of the acquisition
- 500,000+
- supplements in the database the app helps people search and compare
- 160,000+
In this project
Nobody checks the label
Supplements are a $60 billion market with almost no oversight. A brand can print any ingredient and dose on a label, and nobody verifies it. People are left asking what to buy, what they need, and who to trust.
SuppCo rates products on dozens of quality signals. In March 2026 it launched TESTED, which buys supplements off retail shelves, tests them in a lab, and publishes the results.
Make the score explain itself
The core design problem is trust. SuppCo has to educate people and hold opinions about brands without sounding like it sells for them. TrustScore is how we do that.
Every product gets one number. It has to read in a second and still hold up when someone questions it. So the model is open: one tap shows exactly why a product or brand earned its score.
Lead with what people use most
We had an unusual problem for a seed-stage company. The product did too much, and new users could not tell what it was for. By solving every problem a little, we had not solved any of them well enough to learn what people wanted.
Removing features is hard. Instead, we studied what people used and needed most, invested in those features, and made them the obvious paths. Less-used features stayed, but moved out of the way.
One unit at a time
Peptides have become popular with biohackers, and reliable guidance is scarce. SuppCo does not recommend them. But many people already run protocols, so we designed a tool to track schedules and calculate doses.
Units and timing are where mistakes happen, so we designed them to be hard to misread. We started with the 10 most common peptides and tested the flows with colleagues who have run their own protocols for years. The list grows based on what people actually use.
What I’d do differently
In my first months, I held back on structural issues I saw early, like a design system that would not scale with the product. I treated them as someone else’s call.
Now I raise structural problems in my first weeks, before more features depend on them.
