Practice : Improvement Experiments as Work
Purpose and Strategic Importance
Improvement Experiments as Work is the practice of treating improvement efforts with the same visibility, priority, and rigour as feature delivery. Rather than abstract ambitions, changes to ways of working are planned, delivered, and tracked as part of the team’s actual workload.
This creates a culture of continuous learning, ensures retrospectives lead to tangible outcomes, and embeds improvement into daily flow—not as an afterthought, but as a strategic input to performance.
Description of the Practice
- Improvements identified in retrospectives or flow reviews are broken into small, testable experiments.
- These experiments are added to the team backlog or planning cycle alongside delivery work.
- Each experiment has a clear goal, ownership, and timeframe for review.
- Learning from each experiment is captured and used to guide future decisions.
- Teams use the results to decide whether to adopt, adapt, or discard the change.
How to Practise It (Playbook)
1. Getting Started
- When an issue or opportunity is surfaced, phrase the proposed change as a hypothesis (e.g. “We believe that trying X will help reduce blocker age”).
- Define what success looks like and how you’ll know if it worked.
- Add the improvement as a backlog item or sprint goal with a clear owner.
- Time-box the experiment (e.g. 1–2 sprints) and agree when you’ll review the outcome.
2. Scaling and Maturing
- Maintain a simple register of improvement experiments and their outcomes.
- Prioritise experiments using the same criteria as product work—value, risk, effort.
- Track impact using metrics like lead time, velocity, WIP, or team sentiment.
- Involve stakeholders in experiments that touch shared workflows or dependencies.
3. Team Behaviours to Encourage
- Propose small, safe-to-fail experiments—avoid wholesale process changes.
- Reflect regularly on “what we tried” and “what we learned”.
- Celebrate the act of experimentation, not just successful results.
- Share outcomes openly—even when experiments “fail”.
4. Watch Out For…
- Abstract improvement goals with no owners, timelines, or follow-through.
- Teams deprioritising improvement work when delivery pressure increases.
- Treating improvement as extra work rather than integrated delivery.
- Learning being lost because no one captured outcomes.
5. Signals of Success
- Teams can point to recent experiments and what they learned.
- Retrospectives routinely produce change, not just conversation.
- Improvements are delivered incrementally and sustainably.
- Metrics and sentiment show tangible signs of progress.
- Continuous improvement becomes part of the team’s delivery rhythm.