ITSM Metrics and KPIs: MTTR, First-Contact Resolution, SLA Compliance and Dashboards
“What gets measured gets managed”, but only if you measure the right things. ITSM generates plenty of data; the skill lies in choosing a small set of measures that show whether services are healthy, whether users are satisfied and where to improve. This lesson gives the main metrics, the formulas and the traps.
Metric, KPI and Target
- A metric is anything you measure.
- A KPI (key performance indicator) is a metric chosen because it shows progress toward an important goal.
- A target is the level you want to achieve for a KPI.
Not every metric is a KPI. Tracking fifty numbers hides the five that matter.
Core ITSM Metrics
| Metric | Formula | What it tells you |
|---|---|---|
| MTTA (mean time to acknowledge) | Total time from creation to first response ÷ number of incidents | How quickly work starts |
| MTTR (mean time to restore or resolve) | Total resolution time ÷ number of incidents | How quickly service is restored |
| SLA compliance | Tickets resolved within target ÷ total tickets × 100 | Delivery against agreed times |
| First-contact resolution (FCR) | Contacts resolved at first contact ÷ total contacts × 100 | Desk effectiveness and knowledge quality |
| Reopen rate | Tickets reopened ÷ tickets closed × 100 | Quality of fixes |
| Backlog and age | Open tickets, and how long they have been open | Capacity and hidden delays |
| Change success rate | Successful changes ÷ total changes × 100 | Quality of change assessment (see Change Enablement) |
| Incidents caused by change | Count or share of incidents traced to a change | Risk of release practices |
| Repeat incidents | Incidents with a known cause that recur | Effectiveness of problem management |
| CSAT (customer satisfaction) | Positive responses ÷ total responses × 100 | How users feel about the support |
| Self-service resolution | Issues solved through portal or articles ÷ total issues | Value of knowledge and automation |
| Cost per ticket | Total desk cost ÷ tickets handled | Efficiency, to be read with quality measures |
A Worked Calculation
In a month, a desk records 240 incidents with a combined resolution time of 960 hours, of which 210 met their SLA targets and 12 were reopened. It receives 300 contacts, of which 132 are resolved at first contact. Of the 50 changes made, 46 succeeded.
| Measure | Calculation | Result |
|---|---|---|
| MTTR | 960 ÷ 240 | 4 hours |
| SLA compliance | 210 ÷ 240 × 100 | 87.5% |
| Reopen rate | 12 ÷ 240 × 100 | 5% |
| FCR | 132 ÷ 300 × 100 | 44% |
| Change success rate | 46 ÷ 50 × 100 | 92% |
Numbers like these only mean something against a target and a trend. An FCR of 44% may be low for a desk handling simple requests but healthy for one that supports complex engineering systems. Compare with your own history first.
Averages Can Mislead
If nine incidents take one hour and one takes ninety-one hours, the mean is ten hours, which describes none of them. Use the median and percentiles (for example, 90% of incidents resolved within six hours) alongside the average, and look at the longest-running tickets individually.
Balance the Scorecard
Any single measure can be gamed. Speed targets encourage closing tickets early; cost targets encourage cutting quality. Use a balanced set across four angles:
| Angle | Example measures |
|---|---|
| Speed | MTTA, MTTR, fulfilment time |
| Quality | FCR, reopen rate, change success rate, repeat incidents |
| Experience | CSAT, effort scores, complaints |
| Efficiency | Cost per ticket, self-service share, automation rate |
Dashboards That Work
- Know the audience: executives want service health and trend; managers want queues and breaches; agents want their own work.
- Show a few measures with trend lines and targets rather than dozens of tiles.
- Use consistent definitions: write down when the clock starts and stops.
- Pair every chart with a question: what will we do if this moves?
- Review regularly and retire measures nobody uses.
Common Mistakes
- Measuring activity (tickets closed) instead of outcomes (users able to work).
- Rewarding a single metric, which invites gaming; remember that when a measure becomes a target it often stops being a good measure.
- Comparing teams with different workloads using the same target.
- Changing definitions mid-year, which breaks trends.
- Reporting data and never acting on it.
- Choose a small number of KPIs linked to goals; not every metric is a KPI.
- Know the formulas for MTTR, FCR, SLA compliance, reopen rate, change success rate and CSAT.
- Use medians and percentiles as well as averages, and compare with your own trend.
- Balance speed, quality, experience and efficiency to avoid gaming.