Strategic Testing Decisions
summary
Argues that automated tests are just tests are one form of feedback and recommends choosing tests by product phase using Kent Beck’s 3x model: explore, expand, extract. Proposes spike-and-stabilize (ship cheap spikes, refactor survivors), favor market/manual feedback early, then scale automated suites later. Implements serverless walking skeletons, blue/green deployment or canary releases (keep 90–95% on old version), Sentry-driven rollbacks and kanban stages to limit wasted work when “90% of code could be waste.”
AI-written summary · about
chapters (14)
- 0:00Welcome and introduction
- 1:28About the speaker
- 2:31Testing as dogma
- 4:00Costs and benefits of testing
- 8:56Tests as feedback
- 14:363X model explained
- 17:10Product development triathlon
- 20:01Testing in explore phase
- 22:07Less tests in early stages
- 25:38Spike and stabilize
- 27:45Plugin deployment and canary releases
- 32:01Reducing waste with user research
- 33:47Book and training offer
- 34:28Q&A session
related talks
37mTransforming Testing from a Task to a Strategic Mindset
55mShifting the Testing Role Pendulum
58mTesting Strategy - New Model, Better Outcomes
08mTesting In Agile
01h 06mDecisions, Decisions — Dan North on Trade-offs in Software
15mTesting in Production
22mTesting in Production
22mTest Effective Development