Think About an Attitude

In our last issue we discussed how marketers are using order information as well as promotion history and demographic data to create segments and regulate customer contacts (“Learn About the Past,” DIRECT, November).

This column will examine still another type of data-attitudinal data-and how it can be used to identify market segments that should be treated differently.

Interestingly, this data is probably used the least for contact strategy support. This is partly because of the long lead times sometimes involved in fielding the research survey required to collect it, and also because many marketers-due to their inexperience or problems with applying the data to support a contact strategy-have a hard time justifying the financial investment.

Here’s why. Traditionally-regardless of whether one is interested in using attitudinal data to better understand the motivations and barriers for purchase, enhance a brand’s positioning against a particular competitor or leverage brand equity among a loyal market segment or specific market category-an attitudinal segmentation studyis undertaken with a research survey consisting of often hundreds of questions directed to a random sample of a customer file.

Using the data captured, practically any customer file can be segmented into four to eight relatively homogeneous groups. These segments, of course, should differ in terms of their members’ needs and wants, competitive sets, purchase rates, customer satisfaction and so on.

In addition to quantifying the attitudinal attributes for each segment, another activity is to profile each segment using demographic and behavioral variables as descriptors.

Unfortunately, many times these descriptors overlap from one segment to another, and consequently are usually viewed as directional rather than exclusive.

For example, an insurance company conducted a segmentation study to better understand the needs and wants of loyal and extremely loyal life insurance customers. An important finding of the study was that extremely loyal customers averaged much higher face amounts of insurance than less loyal ones, preferred to deal with agents and thought insurance was a great value and they could always use more.

Less loyal customers did not rank the agents as being so important, and didn’t think they needed more insurance.

While the demographic profiles of these segments were different-extremely loyal customers were usually older, better educated and held professional and managerial jobs-both groups had their share of what could be termed “average” customers (that is, people between the ages of 35 and 50, married parents earning under $75,000 a year).

As a result, although the segments did differ demographically, the fact both had a significant number of so-called average customers made such differences fuzzy.

This fuzziness has two consequences. First, it makes it hard to use behavioral and demographic data to predict segment membership. Second, it often means creative strategies targeted to each segment do not do dramatically better than “generic” control creative efforts which are not targeted to a particular segment.

One way to deal with this problem is to use each data type-behavioral, demographic and attitudinal-sequentially, to identify a customer along three separate dimensions.

Here’s how such a process works: Start with behavior and create three to five segments. In the case of one retailer, customers were assigned to categories based on sales in the last year. (Those on the file less than six months were put into their own behavior segment.)

Next, the company used a simple demographic scheme, similar to the one defined here but combining segments into just six major groups.

Then the company contacted the most heavily populated segments by phone to probe these customers’ attitudes about shopping in the store-and specifically to determine why certain segments seemed to satisfy only a few of their requirements there, even though the store offered a complete line of clothing and home furnishings.

No attempt was made to create attitudinal segments within the behavioral/demographic segments, although such segments might exist and might even be predictable. At this stage, the retailer had more than enough information to develop strategies and programs that might increase sales from customers within each of the major segments.

This process is one solution to overcoming the deficiencies of implementing contact strategies and marketing programs built around traditional segmentation practices. Among its benefits is that it’s relatively easy to do and also not costly to perform. Much of the expense of conducting the research is kept to a minimum since surveys are targeted to a small group rather than a large random sample, and are done over the phone.

Additionally, sequentially creating the segments one dimension at a time enables a company to better understand its customer file, thus allowing it to develop more compelling creative thrusts and targeted offers for each segment.