Baselines and What Normal Looks Like
Without knowing the ordinary range, every reading looks like a finding. Establishing it is the first analytical task.
Reading
Reading the output
No experience measure should be published as a single average. Median, worst tenth, and the size of the affected group.
6 notes
Averages Hide the People Having a Bad Time
The single most important reading habit in this field: the average is fine and the tail is why you have a programme.
Segmenting Without Singling Anybody Out
Useful analysis requires breaking the data down. Where that stops being analysis and becomes a report about individuals.
Correlating Telemetry With Sentiment
The two sources disagree more often than they agree, and the disagreements are where the useful findings are.
When the Data and the Complaints Disagree
People say it is slow and the dashboard says it is fine. Both are usually accurate, and the resolution is the finding.
What a Score Cannot Tell You
The limits of quantified experience, stated plainly, so that the programme is not asked to answer questions it cannot.