Benchmarking your predictive dialer

Hands up who’s bought a predictive dialer (a.k.a. auto dialer, phone dialer) on the basis of a manufacturer’s claims, without getting independent verification of performance. OK, some of you are reluctant to own up, but you would be surprised what good company you are in. And we’d be the first to admit that we have customers who buy without always asking. But verification of predictive dialer performance is something that customers are entitled to demand. If they don’t, then this can mean big trouble, if compliant predictive dialing is required.

So how can evidence of effective performance be achieved?

1. One way is to set up a trial between different predictive dialers and compare the outcomes. A lot of care needs to be taken to ensure a level playing field, but it can be done. The result won’t tell you much about the effectiveness of either predictive dialer, just that one gives more talk time per hour than the other.
2. The next thing you can do is really simple: insist on running a campaign in progressive dialing and then predictive dialing mode. All predictive dialers can allow this (think twice about buying, if you are told that one can’t). What you will then get is an absolute measure of the “predictive gain”. If you are working under compliance you may get a shock at how small this difference is in the case of many predictive dialers, which is why you rarely see this test discussed!

Some pitfalls to be aware of:

1. Don’t be beguiled at any predictive dialer benchmarks in the public domain, or provided by any ‘testing authorities’. There are none that tell you how good predictive dialer performance is, especially under compliance. (N.B. We’d love to be proved wrong on this, but it won’t happen any time soon.)
2. Watch out too for a whole lot of mumbo jumbo about predictive dialer design. Even the most intelligent user (or consultant) is hard-pressed to pick his way through the competing design claims made by different vendors. It is of course flattering to have a vendor share the black magic of his predictive dialer design with you; a bit like having Stephen Hawking discuss the creation of the universe with you! You don’t really understand it but are flattered that he shared his views with you and are naturally predisposed to believe him because of his apparent knowledge and passion for the subject! But it’s no substitute for doing the kind of benchmarking we mentioned above.
3. And beware the ‘killer’ reference visit. No customer that you ever meet will confess to having spent his money unwisely, and you can be sure that any performance you see will be difficult (often impossible) to relate to your own circumstances.