Predictive dialling

Higher call volumes -with less hassle- for research and telemarketing operations

Bob Martin

Already well-established in the USA, predictive dialling systems are beginning to impact on European telephone research and telemarketing practice. A form of computerised 'call management', the systems dial up new prospects or respondents at the pace of the operator teams, leaving operators free to concentrate on live interviewing. This article explains why the new technology was needed, and some of the claimed benefits: improved productivity (leading to significant cost savings) and better quality results.

ALL COMPANIES who use the telephone to call large numbers of people are conscious of the importance of direct and well targeted contact. The essence of a telephone-based activity, whether to sell direct or to collect market research data, is interactivity. In a well-conducted call, the individual is treated as a person, rather than with the impersonality of mass marketing or a self-completion questionnaire. By contrast though, mass telephony procedures can be complex and operationally challenging too. Moreover, many features of mass telephoning are still imperfect from the user standpoint.

If we consider a team of operators or research interviewers, their problems often include:

  • The time it takes to dial, particularly if there is a high incidence of non-replies. Operator dialling inevitably produces some mis-dials leading to wrong numbers.
  • The loss of tempo caused by the need to raise or call up a prospect card. Paper-based research system (now increasingly being replaced by CATI - computer-assisted telephone interviewing), are of course more disruptive still.
An early remedy to some of these problems was afforded by the so called 'power diallers'. These could generate very large numbers of initial contacts on behalf of the operators, but presented other problems.


Given the very large numbers of businesses now engaged in telemarketing - these range from publishing to advertising, broadcasting, retail, banking and financial services - there is a widespread opportunity for real improvements in the available technologies. One of these, the subject of this article, is predictive dialling.

The purpose of predictive dialling is to rid the telemarketer, or other users of telephone interviewing, of the administrative burden and obstacles which hinder the efficiency of mass telemarketing campaigns. Such systems are programmed to dial long lists of telephone numbers, switching only the answered telephone calls to operators.

So the system filters out busy signals, ringing phones, and pre-recorded telephone answering machines: on getting a reply it instantly switches an answered call to an available operator. This results in a large number of effective connections in a shorter period of time. As a connection is made, detailed information on the called party appears on the operator's workstation screen. This information, e.g. name, age, address and purchasing history etc. is used by the telephone operator to help the interaction proceed smoothly.


Predictive dialling systems emerged in the US where, because of the low cost of using the telephone, $400 billion worth of products and services are now bought and sold over the phone every year.

As mentioned earlier, one preceding development was what are known as 'power diallers', a crude means by which thousands of telephone numbers on a computer database were integrated with an outbound telephone dialling switch, designed to alleviate the operator of the dialling burden. These implementations were fraught with problems, not least that the system would dial people but would then have no operator free to take the call on answer. The called party would either be left hanging on the phone or be played a recorded 'advertising' message while waiting to be connected to a free operator.

People became alienated to these unsolicited calls, and legislation was introduced to stop the intrusion, but the telephone marketing industry in the States was left with a bad image. In the UK, fortunately, even in these early stages of the growth of the industry, the playing of recorded messages is illegal.


As the name suggests, predictive diallers are so called because of their ability, through sophisticated software, to predict when the next operator will be available to take a call - virtually eliminating the problem of no operator being available to take the call, otherwise known as 'answer-no-operator'. In the unlikely event of 'answer-no-operator', predictive dialling systems drop the call immediatly and automatically tag it until it is re-dialled at a later stage. Upon re-dial the system will secure a free operator before completing the call, insuring 100 per cent that the call will be picked up.

The effect for users, is to increase the amount of time operators spend on the phone, while reducing the time spent on administration, dialling and follow-up. All these functions are, in effect, automated.

Standard features of a predictive dialling system will depend on the application, but usually include, scripting, system activity reporting, and verification functions. More recent systems have the capability to integrate with existing computer systems and database formats through gateway software.

Whereas in the US the technology is well-accepted and widespread, acceptance of predictive dialling is just beginning in the UK and Europe. Valley Forge Information Systems (VFIS), a medium-sized telephone interviewing agency based in the States, uses predictive dialling technology to complete over half a million research interviews per month and claims literally to have trebled capacity overnight. According to Bob King, VFIS director of sales, 'With manual calling we did not have the capability to bid aggressively enough on large projects with tight deadlines. Now we are more competitive and profitable, and our customers are paying less for better work with a faster turnaround. Manual telephone dialling in market research is about as effective as hand typing thousands of letters for a direct mail campaign.'


One problem with predictive dialling systems is the actual introduction of dialling technology to the operators and interviewers themselves. There is initial worry, apprehension and scepticism about the effect on their working habits, and the time it takes to switch answered calls to waiting operators.

In practice, installation of these automated call processing systems goes very smoothly. By taking out non-productive time-wasting elements like choosing who and who not to call, mis-dialling, engaged tones etc., frustration and fatigue are perceptibly reduced.

Such changes appear to lead to higher employee morale and motivation, lower employee turnover and absenteeism, and lower cost-per-contact indices. Rachel Thomas, telemarketing manager at RAC Motoring Services describes the implementation and acceptance of their predictive dialler as a major step forward.

'Our operators love the system as it drastically reduces all the administration and dialling burden associated with the job … we have been able to expand operations to take other work, and are creating new jobs as well.'

Finally, worries about the time it takes to pass answered calls to the operator seem unfounded. Calls are sent to the operator normally within 1/50th of a second. This eliminates awkward pauses on the line and prevents the called person knowing that the call is being handed by a machine to a live operator.

(As soon as the call processing system starts to work, it continually monitors the operators, to establish average call times and call patterns on the inside, and call response rates on the outside. This helps the dialler achieve the accuracy needed to keep the operators busy, but not overload them with too many calls.)

These systems are best suited to large volume outbound telephoning projects, whether in-house or agency. Multiple campaigns can be run over a single system, even over the same operator stations, without campaign interruption. The dialler can be programmed to accept more than one call list and to keep track of several agent groups working each list. Paperwork is, of course, virtually eliminated. Predictive diallers also measure the average length of calls and pace caller accordingly. This allows operator productivity, with little supervisor intervention, and provides the basis for more thorough and detailed campaign reporting.


According to industry experience with predictive diallers, operators talk-time can more than double. In a manual system operators typically spend only 20-25 minutes per hour talking on the phone. Operators using this technology can typically spend 40-50 minutes per hour talking.

Interestingly, operators seem to need less rest and fewer breaks between calls, because the boredom and stress that comes from waiting has been largely removed. Additionally, managers can discreetly monitor operator performance to lend assistance if needed. Overall productivity increases of between 100-300 per cent have been claimed.

However, a predictive dialling system represents a considerable financial investment. Typically a starter system would be between six to eight workstations, the minimum needed to realise the potential of predictive dialling. Another aspect which potential users need to consider is their existing computer infrastructure.

EIS has, however, developed an open architecture which logically and operationally integrates the dialling and switching functions with the controlling computer platform. This allows the system to be interfaced into a greater variety of computer environments, and facilitates enhancements in the system hardware and software. It is also possible to use the system with third party software.


Most of the above remarks focus on direct selling and telemarketing. However, one of the other major applications, market research, also requires examination.

The problems of conducting well-structured and administered research interviews by telephone are becoming increasingly difficult because of:

  • Rising referral rates;
  • Tight client budgets;
  • The dwindling number of hours when respondents are to be found at home.
In Germany, GfK are pioneer users of predictive dialling and computerised call management. Telephone interviewers are getting important increases in sample exploitation and list penetration, contacts made per hour, and surveys completed per hour.

The technology has also freed GfK supervisors to concentrate on coaching and motivating staff instead of worrying about keeping the project moving, supervisors continually monitor activity and real-time statistics to see how well the research project is going. Progress of a study can be analysed in steps. Necessary adjustments can be made within minutes instead of hours.

Although the GfK call processing system represented a heavy initial investment, its cost-benefits exceeded expectation and, at the time of writing, was expected to pay for itself within the eighteen months.


Bob Martin

Bob Martin

    Robert Martin has been managing director of EIS Limited since February 1994. Prior to joining EIS, he was managing director of Saztec International, a company involved in the development of large-scale information and image databases.From 1979 to 1989, he served as managing director and head of European operations for Computer Entry Systems, a manufacturer of automated processing systems. Prior to this, he held various management level positions in sales and marketing with InterScan Corporation and Scan Data Corporation.