Jump to: navigation, search

Pacing Models

Pacing defines the way in which the contact attempts are made; you select the parameters that best satisfy your business requirements and the messaging system adjusts accordingly. Depending on the type of pass you're running, you can choose from three pacing modes:

Auto Pacing

Auto Pacing is designed to automatically set and adjust the pace at which voice, text, or email messages are sent in order to spread delivery out as evenly as possible through the entire delivery window.

You identify the desired completion time and the system sets a rate of attempts that ensures the entire list is attempted by the defined completion time. As new time zone windows begin and end, and as new contact passes begin and end, the pace is re-adjusted to account for the remaining messages to be delivered. When used with a voice pass, Auto Pacing takes into account first attempts, busies, and other retries as calculations and adjustments are made.

Important
  • The system adjusts to accommodate any pause in the sub-campaign by recalculating the number of attempts yet to be made and accelerating the rate within the time available once the sub-campaign is resumed.
  • Calculations are made during each pass, and if retries are a part of the call pass, they are also used as a part of the calculations.
  • Adjusting the rate up or down, using the Rate Adjustment % (Rate Adj %) functionality may cause an undesirable change to the system-defined constant rate per minute of outbound messages. For example, if you have 6,000 calls in one hour, the system will calculate a rate of 100 calls per minute and send the calls at that rate for one hour. If you manually adjust the rate at which calls are made, the system will recalculate and either speed up or slow down the calling rate per the percent indicated. This will likely result in the calls being completed early and possibly flooding the call center, or the list will not be completed.

Auto Pacing Scenarios

Use the Auto Pacing model if the following conditions exist:

Scenario 1
  • You are using a blended contact center (i.e. your agents take calls from multiple sources).
  • It is a priority to complete the list.
  • It is important to receive a steady flow of messages throughout the day.
Scenario 2
  • You are using a blended contact center.
  • You are using an IVR.
  • It is important to receive a steady flow of messages throughout the day.
  • You want to safeguard against flooding your contact center from consumers who opt out of the IVR in order to speak to an agent.
Scenario 3
  • You are using a blended contact center.
  • You have five or fewer agents in your contact center.
  • Your list size is small.
  • The average duration of each call is greater than 5 minutes.
Scenario 4
  • You are using a blended contact center.
  • You have five or fewer agents in your contact center.
  • The average duration of each call is greater than five minutes.
  • Direct Connect rates are low (1 to 2%).
Scenario 5
  • You are using Agent Text Portal.
Scenario 6
  • You are running alert campaigns, with no Direct Connects back to a contact center.

Auto Pacing with Voice

The method for determining the rate for Auto Pacing with voice is the same as for text and email, except that busies and other retries are taken into account.

Auto Pacing with Interactive Voice

When Auto Pacing is selected for interactive voice campaigns, Direct Connect duration and DC penetration can be used to estimate the expected number of agents needed to support calling at the rate determined to complete calling by the specified time.

There is a safety net of three times the user-specified number of agents on the agent group. To prevent severe issues at the contact center, the application will not allow a rate that projects agent usage greater than three times what is currently set on the Agent Group.

Auto Pacing with Text

The method for determining the rate for Auto Pacing with text is as follows:

  • Expected end time = Requested end time – 4 minutes

Because there are no retries, they don’t have to be taken into account. However, time zones are honored in the same way as with voice calling, and will affect the distribution of the messages across the available time period.

Auto Pacing with Email

The method for determining the rate for Auto Pacing with email is as follows:

  • Expected end time = Requested end time – 4 minutes

Because there are no retries or time zone issues with email, all messages will be sent out evenly over the specified time period for the pass.

Example: Auto Pacing and Time Zones

This example demonstrates the effects of time zones on Auto Pacing. Assume all the following statements are true:

  • You are using a blended contact center.
  • The contact window is 10 hours long, from 8am to 6pm, Local.
  • The total number of contacts in the contact list is 11,000 - 1,000 in EST, 1,000 in CST and 9,000 in PST.

In the Eastern and Central time zones, the contact attempt rate would be 100/hr (1000 contacts / 10 hours = 100/hr).

For those contacts in the Pacific time zone, the attempt rate would be 900/hr (9000 contacts / 10 hours = 900/hr).

Pacing autotime.png

Using Auto Pacing, the attempts would be made as follows:

  • 8am EST (7am CST and 5am PST) to 9am EST (8am CST and 6am PST) = 100 attempts per hour.
  • 9am EST to 10am EST (100 attempts/hr.) plus 8am CST to 9am CST (100 attempts/hr.) = 200 attempts per hour.
  • At 11am EST (8am PST) attempts are made in all three time zones = 1100 attempts per hour.
  • At 6pm EST, calling ceases in the Eastern time zone = 1000 attempts per hour.
  • At 6pm CST, calling ceases in the Central time zone = 900 attempts per hour, until 6pm PST when all calling is complete.

Pacing autozones.png

Execution and Escalation

Only parallel execution should be used with Auto Pacing, which is a default setting when creating the campaign strategy. Because sequential execution forces the system to favor earlier passes until all never-tried attempts are dialed, it will result in attempts being left on the table as the earlier passes are spread out over the time period.

Avoid multiple, overlapping passes with escalation when using Auto Pacing. For example, if Pass 2 overlaps Pass 1 and is scheduled to end before Pass 1 ends, then the attempts made at the end of Pass 1 that escalate to Pass 2 will never be tried a second time since the Pass 2 calling window will have finished before receiving the escalated contacts. Pass 1 needs to finish before Pass 2.

Fixed Pacing

The Fixed Pacing model lets you define, or fix, a precise number of attempts to be made per minute. The Fixed Pacing model is used with voice, text, and email passes and is selected as an option during setup.

Fixed Pacing Scenarios

Use the Fixed Pacing model if the following conditions exist:

Scenario 1
  • You are using dedicated contact agents (i.e. your agents take calls from a single source).
  • You have five or fewer agents in your contact center.
  • It is important to send a steady flow of attempts throughout the day.
Scenario 2
  • You are using dedicated contact agents (i.e. your agents take calls from a single source).
  • You have five or fewer agents in your contact center.
  • Your list size is large.
  • The average duration of each call is greater than five minutes.

Edit Calling Rate

To edit the calling rate, use the Requested Rate/Min dialog. Access the Requested Rate/Min dialog in one of the following ways:

  1. From either the sub-campaign Actions menu or from the Agent Groups page, select Pacing > Set Requested Rate/Min.
  2. Select the value in the Contact Attempts - Req/Min column, or right click the pass row and select Set Requested Rate/Min.

Based on historical data, the messaging system will report the projected number of agents required to manage the anticipated number of Direct Connects resulting from that calling rate in the Agents - Projected column on the Agent Groups page.

Important
This number has no direct relationship to available agents, so if the rate per minute is set too high and there are not enough agents in the agent group to manage the DCs, it could flood your contact center with more calls than the agents can handle. Calling is only slowed when an AutoManage rule is triggered.

When using Fixed Pacing, retries are not included as a part of the calculations, nor are any adjustments made in order to complete the list by the end of the contact window. If either of these criteria are important, it is recommended that you select a pacing model other than Fixed Pacing.

Predictive Pacing

Important
You should only consider Predictive Pacing if you have 10 or more agents. Anything less than that can significantly reduce performance.

The Predictive Pacing model is designed to maintain a consistent level of work for a defined number of agents in an agent group by regulating the speed of outbound attempts. Predictive Pacing is used with voice passes and is selected as an option during set up.

Once calling begins, the system continuously monitors and collects information regarding the number of busy agents, the number of consumers in the hold queue, their average hold time, and the average duration in seconds of each Direct Connect call. To do this, the system recalculates and adjusts the rate of outbound attempt every 60 seconds to ensure that the number of direct connect calls into the contact center does not exceed the number of agents available to take the calls.

To set or change the number of agents:

  1. From either the sub-campaign Actions menu or from the Agent Groups page, select Pacing > Set Number of Agents.
  2. Select the value in the Agents - Available column, or on the Agent Groups page, right click the standard agent group row and select Set Number of Agents.

For sub-campaigns assigned to a portal agent group, increase or decrease the number of available agents on the List Agents page by selecting New or Delete, or on the Agent Groups tab by right clicking the portal agent group row and selecting New Agent.

Predictive Pacing Scenarios

Use the Predictive Pacing Model if the following conditions exist:

Scenario 1
  • Your sub-campaign includes a Dialer pass.
Scenario 2
  • You are using dedicated contact agents (i.e. your agents take calls from a single source).
  • You have 10 or more agents in the agent group.
  • It is a higher priority is to maintain overall service levels and keep your agents busy than it is to complete the list.
Scenario 3
  • You are using a blended contact center.
  • It is a higher priority is to maintain overall service levels and keep your agents busy than it is to complete the list.
  • You want to cap the number of simultaneous calls to a specific number.

Predictive Pacing with Dialer

To support our clients who must adhere to regulatory requirements when using the Hosted Dialer to run marketing or sales campaigns, Predictive Pacing supports an abandon rate pacing model.

In the United States, clients are governed by the TCPA and in the United Kingdom, Ofcom. US regulations state that contact centers may not abandon any more than 3% of calls during a campaign, where an "abandoned call" is defined as any call connected to a live party which is not bridged to an agent in less than 2 seconds.

To adjust pacing settings for a Dialer pass, on a Campaign's Outbound tab, select the Dialer pass and adjust the following, as needed:

  • Enable Abandon Rate Pacing: —Select this option to enable abandon rate pacing.
  • Target Abandon Rate: —(Default = 5%) Use this option to define a target level by which the system adjusts the pacing to ensure calls connected to live parties are bridged to agents in the time frame specified in the Abandon Call Definition option.
  • Abandon Call Duration: —(Default = 2 seconds) Use this option to define the maximum time allowed for a call, which is connected to a live person, to be bridged to an agent before being identified as an abandoned call.

Example: Auto Pacing vs. Predictive Pacing

This example demonstrates the benefits of using of Auto vs. Predictive pacing in a blended contact center.

Assume all the following statements are true:

  • You are using a blended contact center.
  • The contact center is staffed with 50 agents.
  • The contact window is 10 hours long.
  • The number of contacts in the contact list is sufficient to keep 10 agents busy for 1 hour.

In this example, using the Auto Pacing model would keep 1 agent busy for 10 hours, consuming 2% of the contact center resources across the 10 hour contact window. Using the Predictive Pacing model would keep 10 agents busy for the first hour, consuming 20% of the contact center resources for that hour, and leaving all 10 agents idle for the remaining 9 hours.

Auto Pacing (blue line) distributes the contact attempts more evenly over time than Predictive Pacing (red line).

Pacing autovspred.png

Related Topics

Feedback

Comment on this article:

blog comments powered by Disqus
This page was last modified on November 17, 2017, at 04:22.