Mind Games
Segmenting customers by the advertising that appeals to them MOVE OVER, demographics, psychographics and RFM. There’s a new segmentation tactic in town, armed with customer reactions to pretty pictures and catch phrases. And it’s taking no prisoners.
OK, maybe that’s a little strong. Joel N. Greene, director of database marketing for Sterling Jewelers, Akron, OH, isn’t ready to throw out the firm’s traditional analytic practices yet. But he’s considering augmenting them with creative response segmentation, which places customers into unique categories based on their reaction to advertising.
Getting Creative Creative response segmentation acknowledges that advertising’s artistic content can either awaken or dampen a customer’s instinct to buy. The trick is classifying a prospect base by its anticipated reaction to creative efforts, and marketing to it accordingly.
Sterling uses direct mail to build traffic in its stores. By tracking the customers who come in after being solicited, and comparing purchases made after a mailing with their buying history, the chain is able to gauge the effectiveness of a given mail piece.
The jewelry manufacturer started by generating a series of visual elements, such as product shots, store images and photographs of jewelry customers. Complementing these were written phrases that highlighted qualities such as product brands, customer service and the shopping experience. Typically, marketers using creative response segmentation generate dozens of these, taken from their own material as well as that of competitors, focus groups and internal brainstorming sessions.
Sterling recruited a panel of its existing buyers who had previously received traffic-building direct mail pieces to come in and offer feedback on these elements. These individuals were exposed to 80 combinations of three or four of the visual and verbal elements.
Customers were then asked to rate each combination on a scale of 1 to 9 regarding how well it described something they would want in a jewelry store, and how likely they would be to visit a jewelry store that used each of these elements in its advertising.
Splitting Personalities Analyzing several elements in different combinations allowed market research firm Moskowitz Jacobs Inc., which pioneered creative response segmentation, to split the client sample into two unique groups based on their reactions to the various messages.
What distinguished these two segments was their attitude toward the jewelry-buying process. For instance, customers were classified as to whether their relationship with the store was important, whether the overall shopping experience was more or less of an issue than the item purchased, and what level of salesperson interaction they felt was appropriate.
Howard Moskowitz, president of the White Plains, NY-based research firm, is quick to point out that there are no pejorative attributes. It’s not whether the client is receptive to being solicited but rather how he or she should be approached.
Moskowitz Jacobs waited until this phase of its research was completed before appending data to each customer. By waiting, it was able to assign individuals to each of the two segments while reducing bias based on demographics.
When it did so, Sterling found that members of each group cut across most demographic classifications. The chain identified fewer than a dozen attributes that served as differentiators. Both Sterling and Moskowitz Jacobs refused to hint at the relative importance each element carries, or even their nature.
Sterling purchased overlay data and appended it to a small sample of its customer base, allowing the firm to assign these people to one of the two segments.
In spring 2000 Sterling sent out a test mailing to four sample groups. One of two new creative packages went to each of the two segments the company had identified. Two of the four groups received messages appropriate to their classification. Two received cross-solicitations – messages targeted to the opposite classification. And a much larger section served as a control, receiving Sterling’s standard direct mail traffic-building solicitation.
Moskowitz admits that he was nervous – “on shpilkes,” as he puts it – before the initial test. He needn’t have been. Of the four test cells, three pulled better than the control. The only one that didn’t was one of the “cross-tests” in which inappropriate content was sent to one of the segments.
Future Predictions Greene says that Sterling Jewelers is not yet ready to abandon traditional segmentation based on demographics and psychographics in order to implement creative response segmentation. He offers the caveat that the initial experiment was just that – a test.
“I think any time you get positive or negative responses you want to continue,” he explains. “One test alone is always inadequate. You need to have a pattern of responses.”
At press time Sterling had not scheduled further analyses. However, Greene indicates that future research might include trying to isolate further subgroups and refining the messages to them.
Moskowitz would like to see the company adapt the approach to the Internet. This would allow online testing and instant generation of images and wording most effective in stimulating a sale – or at least a visit – to an individual who’d been segmented. Ideally, by appending data to its entire client file, Sterling could generate images and text appropriate for any of its customers’ visits to its site.
