Show Me the RIGHT Numbers! Are Our Users Happy?
summary
Argues that reliability work must be outside-in SRE and measure actual user happiness not sysadmin metrics, using personas and scenarios to define SLIs/SLOs, anomaly detection (percentiles, STL), passive feedback widgets and Net Promoter surveys, and Watson sentiment on forums. Reports that passive UI feedback can return only ≈0.5% response, warns against averages hiding problems, and shows instrumenting an app to measure scenario divergence and button-repeat or abandonment rates to detect real user pain.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 4:17Definitions
- 5:27Tools
- 7:06Scenario Divergence
- 9:33Asking Users
- 11:40Asking People
- 13:00Combining Methods
- 14:52Caveat
- 17:02Interactions
related talks
40mPanel Discussion: What Users Really Want
20mDelighted, Satisfied or Angry? Building empathy into business metrics to put users 1st
29mUsability testing based on user's mood
45mQuantifying Empathy with Service Level Objectives
29mMeasuring Reliability: What Got Us Here Won't Get Us There
18mProduct Reliability in
04mThe SRE I aspire to be
35mSRE 101