Conversation, Agent, Label, Inbox & Team Reports¶
These five reports are the backbone of ChannelX analytics. They all answer the same kinds of questions — how much work came in, how fast you responded, how long resolutions took — but each one groups the data along a different axis. Learn the metrics once and you can read any of them.
Open them from the Reports entry in the left sidebar, then pick the report you want from the list.
What each report tells you¶
| Report | Best for |
|---|---|
| Conversations | The overall health of your conversation volume and response performance. |
| Agents | Productivity of individual agents and where someone might need support. |
| Labels | Performance broken down by conversation category, exposing knowledge gaps. |
| Inbox | How a single channel performs and how conversations are distributed across it. |
| Team | Productivity and trends for a particular team. |
Labels deserve a special mention: once you tag conversations consistently, the Labels report shows how your team performs on each type of issue. That makes it easy to find topics where responses are slow, where the help center is thin, or where a bot is falling short.
The shared metrics¶
Every report presents a graph per metric. Here is what each one means.
Conversations¶
The number of conversations created in the selected period. Only newly created conversations are counted — if a conversation from an earlier period is reopened, it does not appear here again. For deeper insight into when conversations arrive, pair this with the traffic heatmap in the Overview report.
Messages received¶
The count of inbound messages across all channels for the period.
Messages sent¶
The count of outgoing messages from the account, including those sent by both agents and bots.
First Response Time (FRT)¶
The average time a human agent takes to send the first reply, measured across conversations created in the period. It is the gap between the first human message and the last non-human activity before it — typically the moment the conversation opened, or the point a bot handed it off to a person. FRT is your clearest read on responsiveness.
Customer waiting time¶
How long customers waited for a reply from your account. It resembles FRT but is calculated across every outgoing message rather than just the first. A higher number means individual messages are sitting unanswered longer.
Resolution Time¶
The average time to resolve a conversation, measured from when it first opened to when it was marked resolved, across conversations created in the period.
Reopens affect resolution time
Reopened conversations are not treated as new. If a conversation is reopened and resolved again later, the additional time is folded back in, which pushes the average up.
Resolution Count¶
The number of conversations resolved on a given day within the period.
Reading the graphs¶
Each graph plots a trend across the period. Hover over any point to see the exact value for that day, and — for time-based metrics like FRT and resolution time — the number of conversations the average was calculated from. That conversation count matters: an alarming average built on two conversations is far less meaningful than one built on fifty.
Customising the reports¶
Duration¶
Choose the date range you want to analyse. The default is the last seven days.
Grouping¶
After picking a range, choose how the data is grouped (by day, week, and so on). The available grouping options depend on the length of the range you selected — wider ranges offer coarser groupings.
Trend indicator¶
Beside each metric, a trend figure shows the percentage change versus the previous comparable period. Analysing a week compares it against the week before. The calculation is:
trend = ((current - previous) / previous) * 100
A positive or negative arrow makes it easy to see at a glance whether a metric is moving the right way.
Business hours¶
Toggle the business-hours switch at the top right to recalculate metrics against your configured working hours instead of the full clock. With it on, time spent outside business hours is excluded — so an overnight gap before a morning reply won't inflate your response times. This requires business hours to be configured for the relevant channels.
Common questions¶
Why does a reopened conversation not show up again in the Conversations graph? That graph counts creation events only. Reopens keep the original creation date, so they are not recounted, but they can still affect resolution time.
Should I judge an agent purely on First Response Time? No. Read it alongside resolution time, resolution count, and the conversation volume each figure is based on. A fast FRT on a tiny number of conversations tells you very little on its own.