Quantitative User Research

I spent seven months as the quantitative user researcher for IBM's Analytics Platform portfolio. Images and artifacts from that work aren't public. But, at a high level, my role was to determine IBM's strategy for utilizing usage metrics across a portfolio of SaaS products, and then to implement that strategy. My goals were to help User Researchers understand user behavior, to help Offering Managers understand key drivers of churn and growth, and to provide users with personalized experiences.

I worked with Meghan Corbett, a traditional user researcher, to develop a comprehensive framework for IBM design teams to integrate quantitative methods alongside their existing qualitative process. We built off of existing frameworks including Google HEART and IBM's universal experiences.

I implemented a software instrumentation pipeline, utilizing I developed principles for how users should be modeled on the backend, and guidelines for making the pipeline extensible, so that new feature requests didn't derail the project. I analyzed data in R and Python, and built a custom "command station" web application. This app showed a live stream of the most important metrics on a T.V., displayed prominently in the office.

Some of my more recent work on Immersive Insights has also touched on the idea of gaining data driven insight on user behavior. You can read about that work here.

During my time as a quantitative user researcher, I was further developing ideas that I had begun thinking about two years prior. Those initial thoughts, on using machine learning to improve UX, can be read here.