A number can be precise and still be difficult to interpret. Page views, response times, and error counts become useful only when we know what was measured and why.
Look at units, time windows, and the people or events included. An average can hide very different experiences. A change may reflect a new measurement method rather than a change in the system.
Start with a question you want to answer. Then choose the measure that speaks to it, and record the limits of what that measure can tell you.
- Check the unit and time window.
- Ask which events are included.
- Compare the number with the question it should answer.
An example to consider.
A list of unfinished tasks might suggest workload, confusion, or changing priorities. One conversation about the task can help explain what the count alone leaves uncertain.
Put it in perspective.
Notice what happens when a tool meets an imperfect condition: a small display, an interrupted session, or an unfamiliar file. Those edges reveal its assumptions.
Follow a related question
Distinguish local preferences from account data.
Where a cloud application livesName the activity the tool should improve.
New tools, familiar human questionsKeep learning
Related background to continue exploring this subject.
W3C: web standards W3C: accessibility, usability, and inclusion
