Contact usRequest a demo

Analytics KPIs

This is a preview feature. It may be subject to change or removal with no further notice.

To enable preview features, set com.unblu.platform.enablePreview to true.

For more information on preview features, refer to the Unblu release policy.

Each widget on an analytics dashboard displays one key performance indicator (KPI). This article describes the KPIs you can add to the dashboard: what each one measures, how it’s calculated, and the groupings and filters it offers.

How KPIs are calculated

Every KPI is calculated over the widget’s timeframe. Five KPIs are counts of certain events: Calls, Collaboration layers, Conversations ended, Messages, and Queue interactions. Three are time averages: Average handling time, Average reservation time, and Average waiting time.

Common groupings

Every KPI offers four common groupings:

  • Conversation template

  • Locale

  • Recipient

    Combines agents, teams, and named areas. A record matches if the conversation was directed at any of the recipients you select.

  • Engagement type

Each KPI also offers groupings of its own, listed in the sections below.

Global filters

The dashboard offers six global filters: the four common groupings, which double as filters, plus Assigned agent and Assigned team. A global filter applies to a KPI only if the KPI supports it. For example, Average waiting time doesn’t support the Assigned agent or Assigned team filters. Whether a widget picks up the dashboard’s global filters also depends on its filtering type, described in Grouping and filtering.

In addition, each KPI offers its own filters in the widget’s Filter by section, listed in the sections below.

Average handling time

The average time an agent spent handling a conversation, across all conversations resolved within the timeframe. A conversation is resolved when the agent’s handling of it ends: the conversation ends, the agent’s participation ends, or the conversation is pushed to a queue or forwarded. These outcomes are the Resolution type values. A single conversation can produce several handling time entries: one for each agent who handled it, and more if an agent handles it more than once. To count only the agent who closed the conversation, set the Resolution type filter to Conversation ended.

Additional groupings: Agent, Team, and Resolution type.

Additional filters:

  • Resolution type: Conversation ended, Participation ended, Pushed to queue, or Forwarded.

  • Used feature: matches conversations that used any of the selected collaboration features (Screen sharing, Embedded co-browsing, Context migration co-browsing, Mobile co-apping, Document collaboration, Whiteboard, Universal co-browsing, Document camera sharing, or Calls).

Average reservation time

The average time it took for a queue reservation to be resolved within the timeframe. A reservation is a queue item reserved for an agent, who can accept it, decline it, or let it time out. It covers only reservations, not other queue interactions.

Additional groupings: Queue interaction type (the reservation’s outcome), Agent, and Team.

Additional filters:

  • Queue interaction type: the reservation’s outcome (Reservation redeemed, Reservation canceled, Reservation declined, or Reservation timed out).

  • Min reservation time duration and Max reservation time duration.

Average waiting time

The average time visitors waited for an agent. A conversation can go through several waits, since each time it’s queued or forwarded it starts a new one. The KPI covers every wait that ended within the timeframe, regardless of whether an agent accepted the invitation.

Additional groupings: Waiting time outcome, Waiting time type, Invitation target, Accepting agent, and Accepting agent team.

Additional filters:

  • Waiting time outcome: Accepted (an agent accepted the invitation) or Aborted (the invitation was unanswered, revoked, or superseded, for example by a forward).

  • Waiting time type (how the wait started): Initially queued, Reboarding queued, Push to queue, Agent forwarding, or Team forwarding.

  • Invitation: the agents, teams, or named areas the conversation was queued for. A record matches if it was queued for any of the values you select. A conversation queued for several targets, such as one an agent forwards to their own team, is counted once for each target.

  • Accepting agent and Accepting agent team.

Calls

The number of audio and video calls that ended within the timeframe. Each call counts once, regardless of how many participants it had.

Additional groupings: Call type, Call end reason, Participant count, Starting person, and Starting person team.

Additional filters:

  • Call type: Audio or Video.

  • Call end reason: Not answered, Aborted by initiator, Declined, Declined because unavailable, Last mandatory participant left, Conversation ended, Recording denied, System shutdown, or Unknown.

  • Min overall participants and Max overall participants.

  • Starting person and Starting person team.

  • Min active duration and Max active duration.

Collaboration layers

The number of collaboration layers that ended within the timeframe.

Additional groupings: Layer type, Layer stop reason, Starting person, and Starting person team.

Additional filters:

  • Layer type: Screen sharing, Embedded co-browsing, Context migration co-browsing, Mobile co-apping, Document collaboration, Whiteboard, or Universal co-browsing.

  • Layer stop reason: Stopped by user, Approval declined, Removed, Inactivity, Recording stopped, Conversation offboarding, System shutdown, Activation error, or Rendering Service stopped.

  • Starting person and Starting person team.

  • Min active duration and Max active duration.

Conversations ended

The number of conversations that ended within the timeframe.

Additional groupings: Agent, Team, Person rating, and End reason.

Additional filters:

  • Conversation end reason: Ended by participant, Mandatory participant left, Request declined, Request not answered, External channel closed, Solved by bot, Canceled, Call ended, or No agent available.

  • Min rating and Max rating: the conversation rating, normalized from 0 (lowest) to 1 (highest). This is independent of the display scale set by com.unblu.messenger.conversationRatingScale (1 to 10 by default).

  • Used feature: matches conversations that used any of the selected collaboration features (Screen sharing, Embedded co-browsing, Context migration co-browsing, Mobile co-apping, Document collaboration, Whiteboard, Universal co-browsing, Document camera sharing, or Calls).

Messages

The number of messages sent within the timeframe.

Additional groupings: Agent, Team, Sender type, Sender, Sender team, and Message type.

Additional filters:

  • Sender person type: Agent, Visitor, System, or Bot. By default, all sender types are counted; use this filter to narrow to specific ones.

  • Sender person and Sender team.

  • Message language.

  • Message type: Text, System, File, Card, List, Approval request, Multiple-choice question, Rating question, Text question, Reply, Question aborted, Recording available, or Template message.

  • Min character count and Max character count.

Queue interactions

The number of queue interactions that were resolved within the timeframe. A queue interaction is an invitation or reservation offered to an agent, or a visitor leaving the queue; it’s resolved when it reaches a final state, given by its Queue interaction type. Each interaction counts once. Unlike Average reservation time, this KPI counts every kind of queue interaction, not only reservations.

Additional groupings: Queue interaction type (the interaction’s final state), Agent, and Team.

Additional filters:

  • Queue interaction type: Visitor left, Invitation redeemed, Invitation revoked, Reservation redeemed, Reservation canceled, Reservation declined, or Reservation timed out.

  • Min reservation time duration and Max reservation time duration.

See also