Time Anchors

Time Anchor (also known as "fact date") is the timestamp used to determine when the value of a particular metric aggregates.

Each out-of-the-box (OOTB) report in Gladly is attached to a specific Time Anchor, which guides each report's data output, whereas Insight Builder allows you to use different Time Anchors. When you begin to analyze various reports, it's essential to know which Time Anchor a report uses, as each report's output will differ.

How Time Anchor works

Let's say you run a report that queries data between 1 PM and 2 PM. A timeline showing the duration from 1 PM to 2 PM clearly marked.

For things like individual events that only happen at one point in time, finding the right set of data is pretty straightforward—did the event happen in that period of time? Graph showing data points between 1pm and 2pm with highlighted elements.Green icons represent events.

For things like Contacts and Conversations, which have a “state” (e.g., Queued, Fulfilled, and End state), it’s imperative to know which event/time it’s anchored to. This determines the set of data you’re considering for analysis. Timeline showing events Q, F, and E between 1pm and 2pm.

Let's use the Inbound Contacts report, which uses Queued as the Time Anchor. The report would then consider these Contacts: A timeline showing events labeled Q, F, and E at different times.

The Abandoned Contacts report uses Ended as the Time Anchor. The report would consider these Contacts: Timeline showing events labeled Q, F, and E between 1pm and 2pm.

You can see how it wouldn’t make sense to conclude by comparing (i.e., creating an average) information from the two reports that use two different Time Anchors because you would be analyzing different data sets.

On the other hand, if both reports were anchored to the same event, you could make comparisons and draw conclusions from the same data set.

A simplified example

Here’s a simplified example. Let’s assume we compare the 2 reports mentioned above, Inbound Contacts and Abandoned Contacts.

  • Inbound Contacts report is anchored to Contact Queued.

  • Abandoned Contacts report is anchored to Contact End.

Graph showing inbound and abandoned contacts over time with highlighted abandoned contacts.

If you tried to make a comparison between these two reports, you might reason that there are:

  • 2 Inbound Contacts (Queued)

  • 3 Abandoned Contacts (Ended)

3 Abandons / 2 Inbound = 150% of Contacts between 1-2 pm are abandoned!

The insights are mixed up. If you’re comparing apples to apples, you’d either say:

  • 100% of Contacts ended between 1-2 PM are abandoned

  • 0% of Contacts started between 1-2 PM are abandoned

So you can see why Time Anchor is important.

Time Anchor definitions

To truly understand how Time Anchors are used in reports, you should know how each Time Anchor is defined.

Conversations

Conversation last closed at

A Conversation that was closed-reopened-closed-reopened-closed would be anchored to the very last close event. In this example, it's considered a single close event.

Conversation Closed at

A Conversation that was closed-reopened-closed-reopened would be anchored to every close event. In this example, there are two close events.

Conversation reassigned at

When a Conversation was last reassigned.

Contacts

Contact queued at

Time a Contact was queued.

Contact fulfilled at

Time a Contact was fulfilled.

Contact initiated at

Time a Contact was initiated.

Contact transferred at

Time a Contact was transferred.

Contact ended at

Time a Contact ended.

Contact ended date (UI), Contact ended at (API)

Specific to the Work Session report.

  • UI - Contact ended date is the date a Contact is ended.

  • API - Contact ended at is the date a Contact ended.

Durations

Duration Started

Time an Agent started a given duration of being 'Active' or 'Away.'

Events

Event occurred at

Time when the relevant event occurred.

Time of event - when Answer was used

Time an Answer was last used.

Time of event - Conversation reopened

The time a Conversations is closed-reopened-closed-reopened would be anchored to every close event. In this example, there are two reopen events.

Time of event - Conversation created at or closed

Time a Conversation was either created or closed.

  • Created – When a Conversation was created during a specific time frame.

  • Closed – Any time when a Conversation was closed during a specific time frame. This can happen multiple times for one Conversation.

Time of event - Task created or closed

Time a Task was either created or closed.

  • Created - When a Task was created during a specific time frame.

  • Closed - Any time when a Task was closed during a specific time frame. This can happen multiple times for one Task.

Time of event - Declined or Missed Contact

Time a Contact was either declined or missed.

  • Declined - Any time when a Contact was declined during a specific time frame.

  • Missed - Any time when a Contact was missed during a specific time frame.

Time of event - when Topics were updated on a Conversation

Time a Topic or set of Topics was added to a Conversation.

Time of event - when Throttle % is changed

Time throttle % was changed as a reaction to incoming chat volume.

Time when a link is clicked

Time of event - when a search happens

Time a search was started/initiated.

Time of event

Time an event occurred.

Tasks

Task closed at

Time a Task was closed.

Task created at

Time a Task was created

IVR

IVR started at

Time when a call entered the IVR.

Payments

Date Payment Request

Date when the card payment request was requested.

Time Buckets

Bucketed

Segments of time across different "buckets" or time granularities specified using the report's rollup filter. For example, an Agent going "available" at 9:01 and then "Away" at 12:15 has a reportable segment of over 3 hours. If the rollup is set to "half-hourly", one might approach the scenario one of the following ways:

  • Anchor to the start of the timespan. So, the 9-9:30 granularity would say 3h 14m.

  • Anchor to the end of the timespan, so the 3h 14m would be in the 12-12:30 rollup.

  • Distribute the timespan to each of the "bucket" of time. This is how reports using "bucketed" as time anchors work:

    • 9-9:30 - 29m

    • 9:30-10 - 30m

    • 10-10:30 - 30m

    • 10:30-11 - 30m

    • 11-11:30 - 30m

    • 11:30-12 - 30m

    • 12-12:30 - 15m

Voice

Voice Outreach Time

Time when a Proactive Voice recipient was updated to a complete status.