Article : What Is Predictive Dialing?
Predictive dialing technology is used in call centers to automatically dial phone numbers and screen out non-productive call attempts such as busies, no answers, fax/modem, answering machines, or Standard Information Tones (SIT).
Only successful voice connections are transferred to agents in the call center while simultaneously delivering customer information to the agent's desktop.
With this efficient model, call centers can benefit from increased talk time with customers or prospects and noticeable time savings since agents do not waste time analyzing non-voice call types. Predictive dialing systems have evolved from primarily hardware-based systems to software-based systems.
Sponsor message - content continues below this message
Employee Engagement Awards NOW OPEN!
Enter your center,and compete for this prestigious award to attract new staff and show investors and customers you are #1 in your nation!!
Content continues ….Software-based predictive dialing solutions do not require additional hardware such as adjunct switches or proprietary telephones and are typically less expensive than comparable hardware-based solutions. Lower priced predictive solutions have made there their way from large, enterprise enterprise-only solutions to more affordable solutions for small to mid sized businesses. Predictive dialing technology, traditionally used for telemarketing and collections, is now also being utilized for new functions in the call center such as proactive outbound customer service.
There are some basic features to look for when evaluating Predictive Dialing dialing solutions.
The pacing algorithm is a key feature of predictive dialing. This is the method used to maximize the efficiency of your call center, by lowering your abandonment rate while making sure all your agent are as busy as possible. Typical predictive dialing algorithms take into account variables such as Anticipated Agent Availability, Number of Agents Logged On, Average Talk Time, Average Wrap Time, Answer Rates, Connect Rates, Abandonment Rates, Agent Pacing, and Queue Pacing. The algorithm automatically adjusts the calling rate in order to meet the service levels that are set.
The biggest issue with pacing is controlling the abandonment rate, or the number of "answered" calls dropped due to the lack of an available agent. A key feature is the ability to adjust the rate of abandonment in order to "lose" the fewest number of people answering the phone. With the recent amendments to the Telemarketing Sales Rule by The Federal Trade Commission (FTC), organizations must ensure that no more than three percent of calls answered by a person are abandoned.
2. Answer Detect
Answer Detect is the ability to detect who or "what" answers a call -- whether it is a human voice, an answering machine, a fax machine, a modem, a busy signal, or an SIT. This is a key feature especially when trying to detect the difference between a real human and a recorded message. Many predictive dialing solutions use Voice Detection Units (VDUs) to properly differentiate the call types, and most provide over a 90% accuracy rate in detecting a real human voice. The ability to leave a pre-recorded message if an answering machine or voice mail is detected is also a key feature when comparing dialing solutions.
Precision dialing is an ensemble feature that uses "alternate" dialing methods when required. Precision dialing ensures that an agent is available prior to dialing the last digit of a phone number. This feature is used to reduce the abandonment rate but it also, however, has the effect of slowing down the dialing. Other dialing methods typically available in predictive solutions include Manual Dialing, Preview Dialing, Progressive Dialing, and Power Dialing.
Manual – - Agent manually dials the phone number using their phone set
Preview - Agent reviews the customer profile information before the call is initiated
Progressive - Simultaneously Displays displays the customer profile simultaneous with the phone number being dialed
Power - (Often confused with Predictive dialing) When an agent becomes available, the dialer initiates multiple calls or "over dials" to get one contact. This method often results in a very high abandonment rate and requires more trunk lines to be available.
Campaign Management is another key feature of predictive dialing, and is the ability to create, execute, and monitor outbound calling campaigns. Features that may be included in the campaign management capabilities include:
Record Release Management - Controls the order, mix, and time constraints for the release of records within individual lists. This may include the ability to define a release priority for customer records in the list, define the conditions that must exist in order for a record to be released from a list, and define the number of customer records that should be released from a list before moving on to the next list.
Stop List Management - Prevents calls to be placed to individuals who have requested "no more calls" on current or future campaigns or on future campaigns.
Predicted List Exhaustion - The ability to predict when a list may become exhausted.
Recycling - Ensures optimal list penetration and contact rates through intelligent scheduling of callbacks. Customers are called back at later dates and times based on a variety of variables such as call outcome, date, day, and time of call attempt, or number of call attempts. The type of telephone, such as Home, Office, and Mobile, to call the customer back on can also be varied based on these variables.
Real-time Graphical Reports - The ability for the supervisor to see trends and statistics for each list and campaign to determine the level of success.
Real-time List Management - The ability for Administrators to dynamically create and delete lists, modify list parameters, and load new records into lists without stopping a campaign.
Desktop Automation is the application the call center agent sees and works with on a daily basis. Typical predictive dialing solutions includes such functions as customer profile, call status, product information, order entry, and schedule callbacks in the agent application. It also includes the scripting capability of the predictive dialing solution, where agents can be guided through a sales process with conditional branching capabilities and embedded text for product information and customer data.
Today, many predictive dialing solutions offer browser-based agent applications, enabling easier installation and maintenance. Browser-based applications also better support distributed call center environments where agents can literally be located anywhere in the world.
6. Call Blending
Call Blending is an optional feature of predictive dialing solutions that enables agents to be automatically moved between inbound and outbound campaigns based upon call volumes. The mixing of inbound and outbound calls allows call centers to maximize agent utilization by having agents available for the peak inbound calling periods and also available to handle outbound calls during slow inbound call volumes.
In order to determine the success of predictive campaigns, reporting is a required feature. This feature should provide both real-time and historical reports on lists, campaigns, agents, and call center statistics in general. Standard reports should be available in graphical and tabular form with the ability to customize or create new reports.
How To Implement
A Predictive Dialing Solution
The successful implementation of a Predictive Dialing Solution depends on the following phases:
Define the requirements
The number of agents in the call center and the number of target customers to reach.
Often, a call center will have an existing PBX, and whether the chosen predictive solution can support the PBX is a factor in controlling costs. The number of trunk lines in existence or new lines to be ordered will play a big part in this too.
Type of communication network, Analogue and Digital T1 or E1, will have an influence on the choice of VDUs (Voice Detection Units) that the solution can use to detect the call outcome.
Designing the solution based on the requirements
Contact scripts to guide the agents through a consistent approach.
Dialing strategy – - consider Consider when and how to get maximum list penetration? . For example, switching to progressive dialing on a campaign if the Predictive predictive solution cannot get a live voice, as agents should not be waiting indefinitely for a voice call.
Outcome handling – - designing Designing the recycling strategies for non-voice contacts. Also designing how answering machines will be handled (i.e., transfer the call so that an agent agent-friendly or automatic generic message can be left on the machine, or just hang up the call).
Deliver the solution - Check all prerequisites specified by the solution vendor such as: Functioning PBX
Sufficient trunk lines
Functioning network for all participating servers, agent workstations and telecommunication equipments (PBX, IVR, ACD, etc.)
Unit tests need to be conducted on individual components to ensure that they are functional.
System test a sample of customer records prior to going live. These are "controlled" customers whereby the outcomes are predefined and the testers know in advance how the system should handle these calls.
Going live – - things Things may go wrong the first time, so make sure specialists are on hand to deal with issues!
Refine the solution
Even though the system appears to be working, it pays to continue to monitor the system in order to gain maximum benefits. This is where good reporting and the ability to fine-tune the system come into play.
Today's Tip of the Day - Be Prepared!
More Editorial From Cincom
About Theresa VanLaeken:
Theresa VanLaeken is Product Manager for Cincom Systems’ Customer Experience Management Technology Group.
Cincom Systems delivers software and services to simplify complex business processes. For more than 43 years, we have empowered thousands of clients worldwide to outperform their competition by providing ways to increase revenue, control cost, minimize risk, and achieve rapid ROI.
Published: Tuesday, August 12, 2003