Business Context

FinDesk is a B2B SaaS company that handles customer support through email and chat. The Head of Operations wants a simple but reliable way to tell whether the support process is improving or degrading over time after a recent staffing and workflow change.

Metric Scenario

In January, the team handled 48,000 tickets with median first-response time of 2.1 hours, median resolution time of 19 hours, backlog older than 72 hours of 1,200 tickets, reopen rate of 6.5%, SLA attainment of 91%, and CSAT of 4.4/5. In March, volume rose to 55,000 tickets; median first-response time improved to 1.8 hours, but median resolution time worsened to 26 hours, backlog older than 72 hours increased to 2,050 tickets, reopen rate rose to 9.2%, SLA attainment fell to 86%, and CSAT dropped to 4.1/5. Leadership is asking whether the process is actually getting better, or whether one improved metric is masking broader degradation.

Requirements

1. Define the primary KPI you would use to assess overall operational health and explain why.
2. Identify which metrics are leading indicators versus lagging indicators.
3. Show how you would determine whether the process is improving or degrading over time rather than on a single month snapshot.
4. Describe how you would decompose the KPI to isolate likely drivers.
5. Recommend specific actions if the data suggests the process is degrading.

Data Available

• tickets: ticket_id, created_at, first_response_at, resolved_at, channel, priority, issue_type, customer_tier, agent_id
• ticket_events: ticket_id, event_time, event_type, status_from, status_to, reopen_flag
• sla_log: ticket_id, sla_target_minutes, met_sla_flag
• csat_responses: ticket_id, score, response_time
• agent_roster: agent_id, team, tenure_months, shift, region