Predictive Dialing For Market Research

This kind of performance in market research has been unheard of. Before we explain why it is now possible, let’s start by dismissing the idea that market research needs a special kind of dialer design. It simply isn’t true. Let’s add some emphasis to this – it totally isn’t true.

Just as in other outbound markets, there is a widely held view in market research that the right predictive solution is an algorithmic one based on the specific behaviour of interviewers. An algorithmic approach (based around mathematical formula) is not just wrong for market research, it is wrong for all outbound activity. The only way to achieve reasonable predictive performance and avoid unacceptable nuisance calls is to adopt the simulation approach taken by Sytel.

Some readers brought up with other dialers in market research may not necessarily agree with this. A debate on the merits of specific dialers is beyond the scope of this page, but we are very keen that the reader should set aside any bad dialing design ideas they may have picked up. So if you think that market research needs a different approach and the material we provide, here and elsewhere, doesn’t convince you otherwise, then contact us so we can set matters right for you.

Dialers and performance in Market Research

There’s just one reason to buy a predictive dialer and that is to generate predictive gain. This is the extra talk time per agent hour that can be obtained on a predictive campaign, compared with a progressive campaign when only one outside line is dialled each time an agent becomes free.

The way it used to be

Historically the predictive gain on market research campaigns was small. There were several reasons for this. Often the lists were ‘friendly’, as opposed to being based on Random Digit Dialing (RDD) techniques – so consumers were more willing to talk for lengthy periods of time, and in some cases were standing by, waiting for their regular call.

Even though predictive dialers have been widely used in market research for many years the actual predictive gain under conditions such as this have usually been very small – rarely more than a few extra minutes of agent talk time per hour. And that of course shouldn’t be surprising when you consider that some interviews could last 20 minutes or more.

Predictive gain in today’s markets

Today many market research campaigns are little different from telemarketing campaigns in that overall average talk and wrap times are low (often around the 60 second mark) and live call rates are also low, as calls divert to voicemail or go unanswered.

Under these conditions, even with low numbers of interviewers, say under 10, a well-engineered dialer operating under compliance can yield significant benefits, without the dialer causing lots of nuisance calls, for example by abandoning lots of live calls or placing them in hold queues. And as interviewer numbers grow on a campaign to say 20-30, then the predictive benefit with Sytel can be as much as 50%, or in other words an increase in talk time per interviewer hour from say 30 minutes to 45 minutes.

And this kind of performance is achievable under the compliance limits set by bodies such as the Federal Trade Commission (FTC) in the US and Ofcom in the UK.