Profile on Plow & Hearth--Reaping What You Sow (2,623 words)
On the other hand, Plow & Hearth's item-specific upsell program was rather rudimentary, says Hay. "For example, an agent would try to upsell cushions to someone who was buying our presidential rocker," he says. "There was not real scripting. A box would pop on the agent's screen. It wasn't very friendly, it was difficult to read, and it would always pop up, regardless of whether the customer had already ordered the cushions or not."
Only the best, most experienced agents were able to work with the primitive system. That meant that during the busy holiday season, the 300 to 400 seasonal workers the company hired couldn't take advantage of either the phone specials or item-specific programs.
During these busy times, the company suspended its successful phone specials program. Explains Hay, "The rationale was that we would forgo these extra sales to cut down on the extra talk time to allow agents to take more orders, be more efficient and keep our service levels high."
With 50 percent of the company's sales coming in the two months before Christmas, cutbacks in upselling efforts were costly. For the 1998 holiday season, a technological upgrade solved the scripting problem, as well as the offering of already-ordered products. Using Decision POS, a PC-based script management software from ASA that integrates with ModelMax, Plow & Hearth could apply real-time predictive model scoring at the point of sale and establish simple rules-based logic for implementing cross-selling and upselling scripts. In other words, when a customer is most likely to accept an upgrade offer, Decision POS alters the script for the agent to make an item-appropriate suggestion.
While the roll-out only tested scripted suggestions of a dozen items, Hay's team saw immediate improvements. "With predictive modeling we saw the penetration and profitability of the phone selling increase, by basically eliminating most of the nonresponsive customers," he says. "Penetration rates tripled for the first two weeks after we implemented it."