Engage Digital | Understanding measures and attributes

Last updated on November 24, 2021

Within Engage Digital, Measures refer to computational expressions of numerical or quantitative data, and Attributes refer to non-measurable descriptors used to break down metrics and measures. In other words:

Measures are qualitative data; attributes represent qualitative data.
Measures are data that can be ‘sliced’ by attributes.

For example, Queue Duration — the duration of time the contact spends waiting for an agent — is a measure. Agent Team is an attribute that can be used to create insights about that measure. If you were to create an insight, Queue Duration can be ‘sliced’ by Agent Team to display, for instance, a bar graph of the sums of queue durations by agent teams.

There’s more to creating insights than just understanding insight types. You also need to know the sources of the data from the data catalog, and what their relationships are to each other. Not all measures can be ‘sliced’ by all attributes; certain attributes don’t work on certain measures. Logical Data Models (LDMs) can explain the relationships between different types of data.

About Logical Data Models (LDMs)

The data used in creating measures and attributes for your contact center’s database is organized into datasets. The dataset — a set of related measures, attributes, or both — is the basic organizational unit for these data.

Datasets are connected to one another to create exclusive relationships. For example, Dataset A might be connected to Dataset B, but not to Dataset C. Dataset B, on the other hand, is connected to Dataset C. This means that A can interact with B but not with C, and vice versa. However, B can interact with both A and C.

Try to imagine datasets as boxes with data in them. Datasets that are connected to one another constitute a logical data model (LDM). LDMs are used to determine which data can interact with other data.

The two most important datasets are segment and agent state.

Segment

A segment is the smallest independently reportable part of an interaction or conversation that is handled by a single agent. A segment has a definite beginning and end time, and is, optionally, associated with specified categories.

An interaction or conversation can have one or many segments associated with it. By definition, one segment must end before the next one can begin. An interaction or conversation is finished when all associated segments are finished. When analyzing all segments together, you will have a complete picture of a given interaction or conversation.

For example, in a typical transfer use case, one interaction may be handled by three agents. Each leg of the interaction (that is, each segment) has a wait time, a handle time, and an agent associated with it. The total average wait time and total average handle time for the interaction will be calculated by adding up the times for each segment, then dividing that figure by the number of segments. The categories that apply to the interaction will include all the categories for each segment.

Mode types

All segments have an attribute called a Mode type, which denotes whether the segment pertains to one of two “modes”: Routing or Inbox. Mode types are assigned according to these criteria:

If the channel belongs to a channel group, all segments are assigned the mode type Routing.
Otherwise, all segments in this channel will have Mode type = Inbox.

Note that if a channel belongs to the Routing mode, there won't be any segment that belongs to the Inbox mode even if the conversation was handled in the Inbox mode.

Excluding a channel from statistics

When configuring a channel in Admin, you can hide analytics metrics for a channel by selecting the Hide the channel from statistics checkbox for that channel. No segments will be reported for a channel with this setting.

For more information on how channels are configured, see Configuring channels.

Agent State

The Agent State dataset includes data about the amount of time that agents are available and handling messages, as well as the amount of time that agents are unavailable. Agent states refer to statuses that describe agents’ actions, such as Available, Busy, On Break, etc. The total duration of all states that occur when agents are logged in is referred to, within the Agent State LDM, as Login.

Now that you have a basic understanding of measures, attributes, and where they come from, let’s take a look at all the specific data under each category. These categories can be found in the data catalog in the left-hand panel.

Measures

Measures are computational expressions of numerical or quantitative data. Measures are data on which mathematical functions can be performed.

Agent State: Measuring by agent state allows you to see Duration data, which pertains to the total time an agent spends in a given state. You can filter specific states once you have configured Duration.
Segment: Measuring by segment allows you to see data for the smallest independently reportable parts of an interaction that are handled by a single agent. Categories include:

CSAT Score: The customer satisfaction (CSAT) score from a survey response.
First Response Duration: The interval between the agent accepting a transaction to the transaction’s first human response. First Response Duration also accounts for automated responses.
Handle Duration: The total amount of time that an agent spends on a transaction.
- For Routing mode segments, Handle duration is equal to the handling time as calculated by the system.
- For Inbox mode segments, Handle duration equals Talk Duration.
Hold Duration: The total amount of hold time during an interaction. Hold duration equals Talk Duration minus Handle Duration. This segment is only calculated for Routing mode segments.
Message Exchanges: The number of message exchange rounds between the customer and the agent.
Message Respond Duration: The interval between the customer’s first message and the agent’s first response.
Queue Duration: The amount of time a customer waits for an agent.
For Routing mode segments, Queue Duration is equal to calculated as the time the interaction was accepted by an agent minus the time the interaction last had a state change.
For Inbox mode, this duration is calculated as follows:
- For the first interaction, the duration is calculated as the time the conversation was opened or assigned minus the time of the first content import for that conversation.
- For the second interaction, the duration is calculated as the time of the first content reply within the segment minus the time that the conversation was opened, reopened, reactivated, or updated.
Queue Duration Business Hours: The amount of time, during regular business hours, that a customer waits for an agent during a given segment.
For Routing mode, Queue Duration Business Hours is calculated as the time the interaction was accepted by an agent minus the time the interaction last had a state change.
For Inbox mode, Queue Duration Business Hours is calculated as follows:
- For the first interaction, the duration is calculated as the time the conversation was opened or assigned minus the time of the first content import for that conversation.
- For the second interaction, the duration is calculated as the time of the first content reply within the segment minus the time that the conversation was opened, reopened, reactivated, or updated.
Talk Duration: The amount of time an agent spends on an interaction. This time includes only the actual focus time.

Attributes

Attributes are non-measurable descriptors used to break down metrics and measures and represent qualitative data. Attributes can also be used to filter your insights, KPI widgets, and dashboards.

Note: Attributes and filters that contain a start or end time use the timezone for the domain where the segment starts or ends. Attributes cannot be filtered by the agents’ timezones.

Agent: Filtering by agent allows you to sort metrics by the agents configured in your account.
Agent Team: Filtering by agent team allows you to sort metrics by the teams configured in your account. If an agent is not assigned to any team, their Agent Team value will display as N/A.
Channel: Filtering by channel allows you to sort metrics by the channels configured in your account.
- Channel: Displays the name of the channels.
- Channel Type: Displays the type of channel.
Disposition: Filtering by disposition allows you to sort metrics by the disposition categories configured in your account. A disposition is a type of label that agents apply to an interaction when solving a conversation.
End Time: Filtering by end time allows you to sort metrics by segment end time.
- End Half Hour: Displays segment end times in half-hour increments.
- End HH:MM: Displays segment end times in hour-and-minute format.
- End Hour: Displays segment end times in hour increments.
- End Quarter Hour: Displays segment end times in quarter-hour increments.
Final Category: Filtering by final category allows you to sort metrics by interactions’ final categories. The list of categories is displayed in the same hierarchy as configured in the Categories configuration page in Admin. Child categories of a parent are shown indented. If no category is assigned, the Final Category value will display as N/A.
Initial Category: Filtering by initial category allows you to sort metrics by interactions’ initial categories. The list of categories is displayed in the same hierarchy as configured in the Categories configuration page in Admin. Child categories of a parent are shown indented. If no category is assigned, the Initial Category value will display as N/A.
- Initial category: Displays the names of initial categories assigned to an interaction.
- Type: Displays the type of category, such as analytics or skill.
Segment: Filtering by segment allows you to sort metrics by the smallest independently reportable part of an interaction handled by a single agent. If no Segment value is assigned, this attribute will display as N/A.

Agent Connected: Possible values: True or False. Indicates whether or not an agent was connected to a call and handled the interaction.
Bot: Indicates whether a segment was handled by an agent or a bot. Yes indicates bot segments; No indicates agent segments.
Conversation ID: Displays a unique identifier number for the conversation.
Creation Reason: Indicates why an interaction was initiated.
- Possible values for Routing mode: Initial, Transferred, Deferred.
- Possible values for Inbox mode Initial, Transferred, Deferred, Reopened.
Final: Indicates whether a segment has been completed.
- Possible values for Routing mode: Completed, Destroyed, Deferred, Undelivered.
- Possible values for Inbox mode: Completed, Canceled.
First Contact: Possible values: Yes or No. Indicates whether a given segment is the first time the customer has contacted an agent for this specific issue.
Identity ID: Displays a unique identifier for the customer.
Index: Displays the number of the segment in the interaction.
Interaction Start: Displays the time when the interaction began.
Language: Displays the language of the interaction’s first message.
Mode Type: Displays the operational mode. Possible values: Routing or Inbox.
Reply Assistant Used: Indicates whether the Reply Assistant composed a reply to a customer.
Resolution Success: Indicates whether the customer’s issue was successfully resolved.
- Possible values for Routing mode: True (default) or False.
  
  Note: The value will be true by default and will turn false if there is a one-contact resolution failure for this segment. That is, the value will be false if the interaction with the customer did not require another contact from the customer after the time period defined in admin settings for a one-contact resolution failure.
- Possible values for Inbox mode: True (default) or False (if a conversation is reopened).
Resolved: Indicates whether the interaction was closed as a result of a given segment.
Service level metrics: The following metrics leverage the service level thresholds configured in Admin > Settings > Service Level thresholds. For example, your admin may have configured the thresholds 10s, 30s, 60s, 10m, 30m, 1h, 1d. For each KPI that supports these service-level metrics, this dataset will be used as a dictionary, and each row will contain information about each bucket individually. For example, in an insight with the measure Sum of Queue Duration, a value of 25 seconds corresponds to an SLI Queue Duration of 30s, as 21 belongs to the range greater or equal to 10s and less than the next threshold of 30s. The value is rounded up.
- SLI First Response Duration: Agent's first response time.
- SLI Queue Duration: The amount of time the customer waited for an agent.
- SLI Queue Duration Business Hours: The amount of customer wait time in a segment during normal business hours.
- SLI Talk Duration: The amount of time that an agent spent on the interaction.
Survey Responded: Possible values: Yes or No. Indicates whether the customer responded to a survey.
Termination Reason: Displays the reason that a segment ends.
- Possible values for Routing mode: Completed, Transferred, Deferred, Undelivered, Canceled.
- Possible values for Inbox mode: Completed, Transferred, Deferred, Canceled.
Thread ID: Displays a unique identifier for the message thread.
UII: Displays a unique identifier for the interaction. For more information about these identifiers, see the API Reference.
Virtual Queue ID: Displays an identifier that combines the channel, language, and routing categories.

Segment Dispositions Hash: Filtering by Segment Dispositions Hash allows you to sort metrics by a compact string representation that uniquely identifies a segment’s disposition categories. If no disposition is assigned, this value will display as N/A. For example, if you have two segments that are assigned three categories, the way to choose those segments is by filtering by hash.
Segment Final Categories Hash: Filtering by Segment Final Categories Hash allows you to sort metrics by a compact string representation that uniquely identifies a segment’s final categories.
Segment Initial Categories Hash: Filtering by segment Initial Categories Hash allows you to sort metrics by a compact string representation that uniquely identifies a segment’s initial categories.
Start Time: Filtering by start time allows you to sort metrics by segment start time.
- Start Half Hour: Displays segment start times in half-hour increments.
- Start HH:MM: Displays segment start times in hour-and-minute format.
- Start Hour: Displays segment start times in hour increments.
- Start Quarter Hour: Displays segment start times in quarter-hour increments.