Situation: For three consecutive sprints, our squad committed to fifteen story points of work but completed only around eight, leading to finger-pointing and low morale during retrospectives.
Task: I wanted to facilitate an honest, data-driven retrospective that helped the team uncover the root causes of our missed commitments without blame.
Action: I analyzed our sprint data and presented an objective breakdown: 45% of our engineering capacity was being consumed by unplanned 'urgent' bug escalations from sales, and our initial story estimation consistently failed to account for code review and QA testing time. In the retrospective, I guided the team to adopt two concrete process changes: we established a designated rotating 'shield' engineer each sprint to absorb all ad-hoc support escalations, and we switched to estimating work based on 'definition of done' (including PR review and staging verification) rather than just writing the code.
Result: In the very next sprint, the team achieved 95% of our committed deliverables. Unplanned interruptions were safely handled by the shield engineer, team stress dropped significantly, and predictability was restored.