- Always start with a question and ask what you want to learn.
- A good hypothesis is a clear articulation about your design, the effect on consumer behavior, and effect on the key metric.
- The goal of data and design is to understand people. All numerical data reflects human needs and behavior underneath.
- Be open minded to all forms of data, and aware of data quality.
- Testing in small samples can reveal big problems.
- Collaborate in the definition phase to develop holistic metrics that reflect the impact well.
- Consider limitations to resources and prioritization.
- Roll out experiences slowly and most of all, serve the user.
Goal → Problem Opportunity → Hypothesis → Test → Result