Practice : Safe-to-Fail Experiments
Purpose and Strategic Importance
Safe-to-Fail Experiments are deliberately designed actions that limit downside risk while maximising opportunities to learn. This approach encourages teams to explore uncertainty confidently, mitigating the fear of failure and enabling faster adaptation.
By embracing small, controlled experiments, teams can uncover insights, validate assumptions, and drive innovation without jeopardising system stability or delivery commitments.
Description of the Practice
- Experiments are scoped to be small enough to contain potential negative impacts.
- Clear criteria define when an experiment is successful, inconclusive, or failed.
- Teams plan for quick recovery or rollback if an experiment does not go as intended.
- Learning from experiments informs prioritisation, design, and process improvements.
- Experiments support both technical and product discovery efforts.
How to Practise It (Playbook)
1. Getting Started
- Identify areas of high uncertainty or risk suitable for experimentation.
- Design experiments that limit exposure—time-boxed, targeted, and reversible.
- Define success criteria and monitoring mechanisms upfront.
- Communicate intent and contingency plans to stakeholders.
2. Scaling and Maturing
- Incorporate safe-to-fail experiments into backlog and planning processes.
- Use automated monitoring and alerting to detect issues early.
- Capture and share learning from experiments broadly across teams.
- Refine experiment design based on prior outcomes and feedback.
3. Team Behaviours to Encourage
- Foster a culture that values learning over blame.
- Encourage risk-taking within safe boundaries.
- Collaborate on experiment design and contingency planning.
- Reflect openly on both successes and failures.
4. Watch Out For…
- Experiments that are too large or poorly scoped to be safe-to-fail.
- Ignoring early warning signs or failing to stop failing experiments.
- Using experiments as excuses to avoid accountability.
- Lack of documentation or sharing of experimental outcomes.
5. Signals of Success
- Teams run frequent experiments that inform key decisions.
- Failures are treated as learning opportunities and lead to improvements.
- Risks are identified and mitigated earlier in the delivery cycle.
- Experiment outcomes influence roadmap and backlog prioritisation.
- Psychological safety increases as teams embrace controlled risk-taking.