Business Context

An Autodesk engineering organization supporting Autodesk Construction Cloud changed its team process 8 weeks ago: it moved from ad hoc intake to a structured sprint model with WIP limits, a required design-review checkpoint, and a weekly bug triage. The Director of Engineering wants to know whether the new process is actually improving execution or just making dashboards look cleaner.

Metric Scenario

Before the change, the team averaged 62 story points completed per sprint, median cycle time of 9.5 days, 31% sprint spillover, 14 escaped defects per month, and team-reported process satisfaction of 6.1/10. In the 4 sprints after the change, throughput increased to 68 story points, but median cycle time rose to 11.2 days, spillover fell to 18%, escaped defects dropped to 9 per month, and satisfaction increased to 7.4/10. However, PMs report that urgent customer issues in BIM collaboration workflows are taking longer to resolve, and one enterprise account escalated two delayed fixes.

Requirements

1. Define the primary metric you would use to decide whether the new team process is working, and explain why it should not rely on throughput alone.
2. Identify leading and lagging indicators that together give a balanced view of process health.
3. Decompose the metric movement to determine whether the process improved predictability, quality, speed, or only shifted work between categories.
4. Explain what additional cuts or comparisons you would make before recommending the process be kept, modified, or rolled back.
5. Propose 2-3 guardrails to ensure the team does not optimize one metric at the expense of customer impact.

Data Available

• jira_issues: issue_id, issue_type, priority, created_at, started_at, resolved_at, story_points, sprint_id, team_id, component
• sprint_summary: sprint_id, committed_points, completed_points, spillover_points, unplanned_work_points
• incident_defects: defect_id, severity, source, created_at, linked_issue_id, customer_account_tier
• support_escalations: escalation_id, account_id, product_area, opened_at, closed_at, severity
• team_pulse_surveys: employee_id, sprint_id, process_clarity_score, focus_time_score, satisfaction_score