Question

Which dataset should I use to report on omnichannel work?

Answer

Currently there are three datasets available in Explore to report on omnichannel tickets. The dataset that you need to use depends on your goal:

Omnichannel - Agent productivity
Omnichannel - Agent state
Omnichannel - Agent state daily

Agent productivity dataset

This dataset focuses on the work offered and assigned to agents. It details how agents use their capacity, such as number of tickets handled or response times.

Use this dataset when you want to analyze how well agents manage their workloads and productivity levels, or to assess agents' performance based on the number of tasks completed and capacity.

Example use case: I want to evaluate how effectively agents handle their workload

Analyze how many tickets each agent resolves in a given period, compare that to their available time, and identify top performers or those who may need additional training or support.

Agent state dataset

This dataset tracks how long agents spend in different states, such as Online, Away, Transfer Only, Invisible, and Offline. It records state changes with timestamps.

Use this dataset when you need detailed insights into individual agent behavior and specific timing of state changes, to understand agent availability and patterns during peak times.

Example use case: I want to manage staffing more effectively

Analyze how often agents switch between the status Online and Away during a busy period. This allows you to identify time frames when agents aren't available and correlate this data with call volumes or ticket inflow.

Agent state daily dataset

This dataset provides aggregated information summarized on a daily basis about how groups and agents spend their time across channels.

Use this dataset to get an aggregated overview of agent performance across days and track general trends in agent behavior.

Example use case: I want to quickly see trends in agent availability or inactivity

Build a high-level overview of agent performance over a week, without the need to analyze detailed timestamps.

For more information, see these articles: