Measuring Wellbeing - Thrive Score

Thrive was flying blind. Recommendations keyed off engagement signals - did you open the notification, did you read the article - not whether your health was actually improving. Someone sleeping four hours a night could get the same generic content as someone thriving. The experience felt impersonal because it was.

That hurt on two sides at once. Users got advice that didn't fit their lives, and the business couldn't prove wellbeing improvement to the enterprise buyers paying for it - engagement is a weak proxy that doesn't hold up in a commercial conversation.

We assumed the concept was already validated - that the job was execution, not discovery. It wasn't. The first direction was a radar plot across wellbeing dimensions: visually interesting, hard to read. The VP of Product's feedback was the classic death sentence - “This looks nice, but I don't understand it.”We'd also weighed letter grades (too much judgment) and a wellbeing-vs-actual age (compelling but complex).

The veto was the right call. It forced the real question: had any of this been tested with users? It hadn't. So I took the concepts out - and heard the VP's reaction back, verbatim. Studying how others earn trust made the pattern obvious: Whoop and Oura work because sub-scores show their work; a credit score is distrusted because it's a number with no because.

That reframed the whole problem. The job wasn't “how do we measure users.” It was “how do we make recommendations relevant to what's actually holding this person back.” Users don't want a score - they want to feel better and don't know where to start.

A single composite score, built from five core behaviors - Food, Movement, Sleep, Stress management, and Connection. It's never opaque: you can always see what's driving it and what to do next. The sub-scores are what make the “because” visible.

The 23-question onboarding survey isn't friction - it's the product doing the work to understand you before it recommends anything. Without it, the score is just a number; with it, every recommendation is earned.

Adding 23 questions to onboarding during an engagement crisis was a non-starter. The Head of Design owned that flow and guarded its KPIs - and she was right to.

I got to yes with two moves:

• Redesigned the flow so users could skip cleanly - taking “you're forcing this on users” off the table.
• Ran user testing and came back with the number that ended the debate.