Using Data Science to Quantify User Journeys, by Sam Zaiss
Sam leads a team of data scientists (now called Experience Analytics) at Microsoft
He’s determined the goal of data science and user research is the same: use data to accomplish goals
Communication breaks down between data scientists and UX researchers because there is a lexicon barrier. When data scientists say there are “lies, damn lies, and observational studies” they mean “don’t assume correlation implies causality.” But UX researchers hear it as “you’re undermining us!”
We all know quantitative (or data science) tells us what and UX research answers why. Both are taking data and transforming it into insights. Insights communication is often where teams start, but in reality if data science and UX research work together at the data level, they can create the insights together.
Triangulation is when you see something happening, but before you share it you look for other areas to gather data to prove or strengthen it. UX researchers can triangulate via data science to strengthen the information.
- Example: users in interviews say they are confused about a choice. Data shows that 20% of users leave when they reach that choice, and never return. Using a data visualization of the user journey they could see more specifically where people were leaving and why, and determine a solution to the choice issue.
This is a framework for data scientists and UXers to discuss user journeys
- User journey objective (is it a destination or repetition?)
- Feature selection (is it filtered, hierarchical, or all features?)
- Desired output (is it a description, prediction, or classification?)
- We can create a Sankey diagram visualization of how the data supports the flow of the journey and how many users drop off along the way
- Sunburst diagrams can show what people are doing in a journey where we want them to be staying/repeating actions
In conclusion, data science is a great approach to help triangulate and strengthen UX research. The work should be a partnership.