Data Mining On the Fly
FutureCall scores the remaining records so that Partners First can prioritize leads. Within a few hours, FutureCall can apply the model and begin testing it. Results are reviewed daily and the model is refined according to response results. Or, if objectives change, the model can be rebuilt from the ground up.
The benefit of TactiCall is that it filters the targeted prospect file through a layer of current data. Branda says that FutureCall's modeling on the fly has given the company double-digit lift in sales and balance transfers.
Achieving significant improvement involves management of data, but also of the call itself through savvy script design. Branda notes that in the case of credit-card marketing, it's essential that phone reps don't push too aggressively to force prospects to take cards they don't want.
"Maybe you rebut once or twice, but you don't rebut five times," Branda says. "If they don't want the card, let it go."
While most companies prefer reps who close as many sales as possible, Partners First doesn't want to maintain credit accounts for people who don't use the card. Instead of boosting card acceptance rates through aggressive sales techniques, Partners First wants to boost its activation rate among the customers who hold the card. One way to do so is cross-selling a balance transfer or fee service at the time of card acceptance or as a follow-up.
"If you have a balance to move, once you put that on the account, you're going to be active by default," Branda says. By targeting those with a likelihood to complete balance transfers, Partners First has boosted its activation rate.
"It's a combination of good data mining techniques with good offers, good scripting and good performance tracking at vendor," he adds.
Partners First develops its scripts with a team consisting of Branda and Bellanger; Sandra Alexa and Reegan Albright from FutureCall; and Al Kleeman and Tracy Murgash of A.S. Kleeman and Associates, Partners First's telemarketing broker. Each script goes through the same testing period as a response model.