Redefining test automation
summary
Argues that automation should serve testing not replace it, trading unfocused UI scripts for targeted checks and tooling across layers; moves from Selenium-heavy end-to-end flakiness to measuring testability, trimming checks from 2,500 to 400 UI checks and adding 500 API checks, and automating support work — e.g. created 89 instances for 14 devices in 4 minutes to save a week of regression — recommends TRIM model, observability, risk-first selection and automating repetitive setup not human judgment.
AI-written summary · about
chapters (24)
- 0:00Intro
- 1:26The journey
- 3:42Why did I fail
- 4:41Designing checks
- 7:50The flawed approach
- 9:48The game
- 12:40Testing and check
- 14:47Testability
- 19:17Checks
- 27:22Checks are good
- 27:42Taking a job in London
- 32:40Repetition
- 36:01My journey
- 37:07Exposure to tools
- 37:26What tool vendors say
- 40:15Automation in testing
- 41:03Human centric
- 42:51Observability
- 43:32Testability over replicate
- 44:16Automate ability
- 44:56Testing skills over coding expertise
- 45:50Risk over coverage
- 46:39Problems before tools
- 48:11Summary
related talks
40mThe Test Automation Firehose
13mBeyond the hype: What test automation really costs
42mOpening Keynote: Blunders in Test Automation
21mThe Essential Pillars of Automation Testing for Success
08mReviewing your test automation strategy"
47mWhat Successful Test Automation is... and Why it's NOT Testing!
55mShifting the Testing Role Pendulum
01h 08mTest Automation != Less Testers || Faster Testing || More Time For ET