Data-Driven Experimentation

Introduction

I partnered with data science, privacy, and engineering teams to build foundational experimentation infrastructure used across Microsoft 365 Consumer. My focus was on enabling cross-surface experiment tracking and creating low-cost tools for PMs to test hypotheses without engineering dependencies. While not immediately customer-facing, this work significantly increased experimentation velocity and insight quality across multiple teams.

Cross-Surface Experiment Scorecards

• Led the creation of new experiment scorecards capable of tracking user journeys that span multiple app surfaces.
• Collaborated closely with data science, privacy/compliance, and security teams to ensure scorecards respected user data boundaries and met Microsoft’s global data governance standards.
• Scorecards required solving technical and legal complexities to attribute downstream behaviors accurately while adhering to strict privacy rules.
• These scorecards are now used widely to interpret experiment impact beyond a single entry point, supporting more strategic decision-making.

Low-Eng Experimentation Platform

Initiated and led the development of a modular experimentation surface—a new platform that allows Product Managers to run messaging and UX experiments without engineering involvement.

Worked across design, engineering, and data science to:

• Define UI boundaries and injection points.
• Build governance guardrails to avoid overexposure to experiments.
• Establish variant targeting rules and success metrics 

The result was a self-service experimentation tool that has now powered dozens of tests—many of which would have been deprioritized due to engineering bandwidth constraints.