Segmentation with Attitude

As we've previously discussed, the goal of segmentation is to produce differentiated groups of customers that look, think, and behave similarly. Having distinct segments in place allows for the business - marketing in particular - to focus on desirable relationships, customize the content of communications, and manage and track the overall health and growth of the customer portfolio. In this article we will focus in on one particular part of segmentation that is often foregone in database marketing applications - customer attitudes.

True CRM requires a firm understanding of all aspects of the customer's relationship with your organization - and segmentation is a great foundation for that understanding. Typical database marketing segments address issues which are directly available or may be inferred from information contained in customer databases. For example, database segmentation solutions often address:

• What customers buy from you.
• When they make their purchases,
• How often they make their purchases,
• The value they create for you.

However, in order to fully understand the customer - their behaviors and motivations - we must also understand aspects of the relationship that are not readily available within the database - customer attitudes. These include:

• How customers think about you and the competition,
• Why customers purchase from you, and
• Why customers purchase from the competition.

As an example, we may wish to understand how customers make decisions about where to buy a product. They may consider a host of issues when making their purchase decision. Their purchase criteria may include the following items which we want to include in our segmentation:

• Selection: Is it important to have a wide range of products within the category from which to choose or are the top sellers sufficient?
• Service: Is the level of service (defined by knowledge, courtesy, or attentiveness of the customer service associate) important to the respondent or are your customers more driven by the ability to "self-serve"?
• Convenience: Is the convenience of the store location important? Will the customer go out of their way to select a specific retailer?
• Price: What is the relative importance of different levels of discounts?
• Brand Affinity: To what extent is there a natural loyalty to a specific retailer?

When completed, the analysis yields a grouping of respondents based on the similarities across these factors. Examples include Price Sensitive; Selection and Service Driven; and Brand Loyal.

While on the surface it is intuitive that attitudinal components be included in segmentation, one must understand why direct marketers have largely bypassed attitudinal dimension of customers when developing segmentation schemes to understand why it is often foregone in practice.

Traditionally, marketers have used primary research to understand consumers' motivations, attitudes, and decision-making processes whereas database marketers have analyzed customer demographics and behaviors to identify marketing opportunities. Historically, combining the two lines of research has been a difficult proposition at best. This is primarily due to a lack of tools accurate enough to project attitudinal information back to the customer database at a satisfactory level; after all, if there were readily available processes that could accurately project primary research response information back to a customer database, then everyone would be doing it!

The creation and modeling of the attitudinal dimension of a fully developed customer segmentation model requires combining primary research tools with behavioral data obtained from the database. Accurately projecting the attitudinal components back to the database to include the results within the overall segmentation scheme is now possible due to advances in the primary research tools available to marketers.

Our Process

At CAC Group, we have pioneered methodologies that allows for the inclusion of an attitudinal dimension as part of building customized segmentation solutions. Our process is highly consultative as we work with you at each step, and includes:

Planning
When developing an attitudinal component for database marketing segments, we must first develop a research plan. This plan must address the following issues:

• What do we want to learn?
• This would include identifying the relevant insights that will assist in understanding marketing's goals and challenges. However, it may also include gathering industry and product-level insights.
• What attitudinal information goes into the decision making process?
• This would include a measurement of a customer's price sensitivity, brand affinities, impact of convenience, the types of and level of service desired, and perhaps other meaningful characteristics (fashion? Selection?).
• Why is this important information for Marketing?
• The plan identifies how the information will be used for Marketing's purposes. This guides the survey design and analysis plan.

Evaluation
We must also understand and create criteria for evaluating successful development of attitudinal components of segmentation. The evaluation criteria include:

• Does the dimension identify important factors in the buying process which drive customer decisions.
• Marketing communications may be tailored to fit customers purchase drivers
• Is there meaningful differentiation among attitudinal groups?
• The solution must be able to speak directly to the salient attitudinal characteristics of each segment
• Attitudes must be differentiated to capitalize on communication opportunity
• Do the attitudes complement the other segmentation components?
• Do they "jibe" with other things we know about customers?
• Is the solution able to be successfully applied to database?
• All records need to be "scored" with the attitudinal models to be incorporated with the full segmentation model.

Development of Attitudinal Dimension
Once the foundation is laid, the attitudinal dimension is developed using a methodology called Discrete Choice Modeling. Discrete Choice methodologies stem from researchers desire to understand the relative importance of factors influencing decisions or customer attitudes. In essence, it helps us understand the important motivators that customers have when making purchase decisions, by asking respondents to "pick favorites" among alternative shopping or purchasing experiences.

Formally, the discrete choice model which makes this possible is called a Hierarchical Bayesian Multinomial Probit model. The ability to accurately map the results back to the data base segments comes as an effect of how the model is specified. The model is developed using a combination of the primary research data and behavioral information from the database. The two types of information are directly incorporated in the Discrete Choice Modeling process through Bayesian process.

When a fully developed segmentation is completed, our clients have the ability to differentiate messaging strategies across segments. Message differentiation may then include the ability to speak to customers regarding their drivers underlying purchase decisions. Some examples include:

• Focusing on a BEGINNING OF SEASON message for fashion-oriented shoppers (or "new technology" for early adopters)
• Focus on DISCOUNTING / CLEARANCE for your most price-sensitive customers
• Reduce communications to customers that have a low value to your company AND don't put much importance on the traits that you have (selection, etc.).

Overall, the attitudinal component plays a large role in versioning which allows our clients to:

• Think differently about our customer relationships
• Innovate communications
• Leverage attitudinal insights across the entire business
• Move from thinking about marketing programs to executing customer driven marketing

Segmentology, the CAC Group approach to segmentation, views the different pieces of information - behavioral, attitudinal and demographic - as meaningful and important differentiators between customer segments. In Segmentology, each aspect of the customer relationship is considered a unique dimension - each providing different insight into the overall relationship with the customer.