01 · RESULTSA service became a product
Along the way, the pre-sales demos I ran helped close Zynga, CD Projekt Red and Google.
02 · MY PARTI led the product side, end to end
- Discovery: worked with client studios on what a team needs to launch a study and read the results without us.
- Vision and roadmap: one direction across 3 products, so they ship as one experience.
- Team: 3 PMs, one per product, with clear ownership of each product and of the interfaces between them.
- Trade-offs: made the calls that cross product lines and both sides of the platform.
- Alignment: kept the CEO, CRO, VP of Research and Head of Marketing on the same priorities, launches and go-to-market.
- Sales: ran product demos for enterprise deals.
03 · PROBLEMStudios liked the studies. They had to wait for us.
Each study needed our researchers to set it up, find players, run the sessions and write the report. Studios valued the results, but every study went through our team.
The goal: a studio team signs up, launches a study and reads the results alone, with the same quality a researcher would give them.
04 · DECISIONSThree products, one experience
The platform is two-sided. Studios launch studies on one side, players take part on the other, and an AI pipeline turns each session into insights. I split it into 3 products with one PM each, and kept the hard calls between them:
- what studios ask for, against a setup simple enough to run alone,
- enough players for every study, without lowering session quality,
- pipeline accuracy, against processing cost and speed,
- new features, against the technical work that keeps the platform fast.
Each feature had 1 main metric and 1 or 2 guardrails. A faster self-serve setup could not lower session quality.
05 · WHAT SHIPPEDLaunch, play, insight
- Studio platform: self-serve setup and launch, so a team runs a study without a researcher.
- Panellist app: players find studies, play, record their face and cash out.
- AI pipeline: combines gameplay data with facial-emotion signals. Dashboards build themselves once data collection ends, so insight takes 1 day, not 1 week.
06 · WHY IT WORKEDSame value, no waiting
The studies were already worth paying for. What changed was who could run them. Clear ownership let 3 teams move at once, and the quality guardrail stopped self-serve from trading quality for speed.