My problem with telemetry is that it's a strictly static, quantitative data source. It only shows data the system designers told it to collect. It may help analyse the key drop issue by contrasting the numbers for instances opened vs. instances successfully completed, for example, but it cannot help identify equally fundamental issues such as crafting bugs, or Ux issues.
In essence, telemetry is quite similar to automated testing (just with a user-provided data source), whereas customer feedback is the analogue to conventional testing by professional testers / end users.
Automated testing is brilliant to cut testing efforts for software components that evolve, i.e. where it's important that previously working functionality still works after a new version comes out. But do not doubt that the real meat in software QA always comes in the form of feedback from qualified, experienced testers, and is indispensable when innovation is disruptive.





Retour en haut








