What happens to a ticket from intake to closure?
- Intake: the user reaches the desk by phone, email, portal or chat, and a ticket is created every time
- Classification: category, impact, urgency and the affected service are recorded
- First-line diagnosis using a knowledge article or runbook
- Resolution, or escalation to the right resolver group with the evidence already gathered
- Confirmation with the user before closure, then a short satisfaction check
Why does ticket categorisation matter so much?
Categories drive routing, reporting and automation. If half your tickets end up in 'Other', you cannot see which systems cause the most work. Keep the category tree short (two or three levels), name categories after what users experience rather than internal system names, and review the 'Other' bucket monthly.
What is shift-left, and how does it reduce cost?
Shift-left means moving work to the earliest level that can safely do it: from engineers to the help desk, and from the help desk to self-service. Each time an engineer solves a repeat issue, the fix is written up so the help desk can handle it next time. Each time the help desk answers the same question repeatedly, it becomes a self-service article.
Continue reading
How should the knowledge base be maintained?
- Agents create or update an article whenever they solve something that is not yet documented
- Each article states the symptom in the user's words, the cause, and the fix step by step
- Articles are reviewed when a system changes, and retired when it is replaced
- The help desk tracks which articles resolve tickets and which ones lead to escalations
What does a good escalation look like?
Engineers lose time when escalations arrive without context. Agree a minimum standard so every escalated ticket includes:
- Who is affected and how many users
- Exact error messages, screenshots and timestamps
- What was already tried and the result
- Business impact and any deadline the user mentioned
Which help desk KPIs matter most?
| KPI | What it shows |
|---|---|
| First contact resolution | How much is fixed without escalation or follow-up |
| Mean time to resolve | Speed of resolution by priority |
| Backlog age | Whether tickets are quietly ageing in queues |
| Escalation rate | How much work reaches engineers, and whether it should |
| Reopen rate | Tickets closed before the problem was really solved |
| User satisfaction | How the experience feels to the person asking for help |