How to read duration and reaction metrics on the Appeals tab
When you need to assess how fast appeals are handled over a week or month — and understand why average reaction time sometimes looks unrealistically high — open the Appeals tab in native statistics and scroll to Duration indicators and Reaction time.
When to use this
- You need an SLA view: how fast operators reply after joining a dialogue.
- Average operator reaction time for the period looks unrealistic (hours or days).
- You want to see how long queue waits or operator transfers affect reporting.
- You compare workload: dialogue duration, queue time, closures by inactivity.
What to know
The Reaction time block has three different KPIs — do not mix them:
| UI label | Meaning | When to use |
|---|---|---|
| Average operator reaction time | Average time until the operator joins the dialogue | Historical parity with Metabase reports; queue impact |
| Average first reaction time | Average time from join to the operator’s first reply | SLA for “first response after taking the chat” |
| Dialogs with reaction time anomaly (>1 day) | Count of dialogues where time until join exceeded 1 day | Spot outliers that pull the average up |
Duration indicators block:
| KPI | Summary |
|---|---|
| Average dialogue duration | Average operator time in one dialogue |
| Total dialogue duration | Sum of all dialogue time in the period |
| Time in queue | How long clients waited in queue before an operator joined |
| Closed by inactivity | Dialogues closed automatically by timeout |
Before you start
Steps
- Open Menu → Statistics.
- Go to Main statistics.
- Select the Appeals dashboard tab.
- Set Date and other filters, then click Apply filters.
- Scroll to Duration indicators — review the four duration KPIs in one row.
- Below — Reaction time: compare average operator reaction time and average first reaction time.
- If connection time is high, check dialogs with anomaly (>1 day) — a large count means outliers in queue or long-lived tickets.
- For volume, see Quantity indicators (incoming / outgoing dialogues).
Typical scenario
Symptom: monthly average operator reaction time is several hours, but average first reaction time is a few minutes.
Conclusion: operators reply quickly after joining; the high average reflects queue wait or time until assignment, not reply speed.
Next steps: review time in queue, staffing, routing, and skill groups; use the anomaly counter to gauge outlier volume.