Introduction
Measuring L3 support team effectiveness is a non-trivial task. Standard metrics (closed tickets, response time) don't account for incident complexity and solution quality. In this article, I present a comprehensive KPI system based on ITIL methodology.
Block A (70%): Weighted SLA with Complexity Coefficients
The main evaluation block accounts for task priority and complexity. Each priority has its target SLA and weight:
- P1 (Critical) — SLA 4 hours, weight 4.0. Complete service failure.
- P2 (High) — SLA 8 hours, weight 2.5. Significant functionality reduction.
- P3 (Medium) — SLA 24 hours, weight 1.5. Partial degradation.
- P4 (Low) — SLA 72 hours, weight 1.0. Cosmetic issues.
Complexity Modifiers
Additional coefficients account for task specifics:
- R&D Interaction (×1.3) — tasks requiring coordination with the development team
- Astra Linux Security Modules (×1.5) — working with certified security systems
- Multi-service Incident (×1.4) — affecting multiple TermiDesk components
- Proactive Detection (×1.2) — engineer discovered the issue before customer report
Proactivity Scoring
Proactivity is a separate metric in Block A. It measures incidents prevented or discovered independently: log analysis, metric monitoring, automated checks. This encourages engineers to be proactive rather than purely reactive.
Block B (30%): Personal, Group, and Corporate Responsibility
Block B is divided into three levels: Personal (40%) covers schedule adherence, documentation quality, and training. Group (35%) covers peer assistance, knowledge sharing, and mentoring. Corporate (25%) covers process improvement, initiatives, and feedback.
Pilot Period
KPI system implementation requires a 2-3 month pilot. During this period, metrics are collected but don’t affect evaluation. The goal is coefficient and threshold calibration on the team’s actual data.
Data Collection Methodology
Block A data is collected automatically via Jira API: response time, resolution time, priority, category. Block B data comes through regular manager evaluations and peer review. Collection automation is critical for objectivity.
Conclusion
A two-block KPI system with weighted SLA and complexity modifiers provides fair evaluation that accounts for the real complexity of L3 engineering work. Pilot periods and iterative calibration ensure metric adequacy.