Define a shared KPI dictionary with precise formulas

Create a shared KPI dictionary for Reduce support backlog while maintaining reliable resolution quality..

Current metric names and definitions:
Tickets closed per week; average resolution time; backlog. Teams disagree about whether reopened tickets count as new work and whether weekends count toward resolution time.
Available source fields and operational context:
Ticket records: ticket_id, created_at, first_resolved_at, latest_resolved_at, status, priority, team. Status history includes reopened events. Business calendar and metric owners are not yet documented.
Audience and review cadence:
Weekly support operations review; dashboard refreshed each morning.

Select at most six metrics that directly inform the goals, retaining existing metrics when useful and explaining any proposed replacement. Distinguish outcome measures from possible leading indicators; label an unvalidated leading indicator as a hypothesis.

Produce a registry table with:
- Metric name and plain-language purpose.
- Formula, numerator and denominator where applicable, unit, and whether higher or lower is desirable.
- Inclusion and exclusion rules, event or entity grain, aggregation window, timezone, and handling of missing, duplicate, or late records.
- Required source fields, refresh cadence, responsible owner or "owner needed", and known measurement limitations.
- Target or alert threshold only when supplied or justified by a stated decision rule; otherwise "not agreed".

Identify two definitions that could produce different numbers under the same metric name, and resolve the ambiguity or record the decision needed. Give a small validation checklist for reconciling the dictionary against source records before adding metrics to a dashboard.

Do not invent data availability, benchmarks, targets, business owners, or measured relationships. Mark proposed definitions as proposed. End with the decisions that the relevant stakeholders must settle.