Successful customer management involves changing customer behavior. Companies incite that change by taking actions -both mass actions that affect all their customers and differential actions that affect only some. In order to choose the appropriate action to achieve a particular objective, marketers need to know how each customer will react to each possible action, and that is what predictive behavior modeling is all about.

Predictive marketing software makes it possible for businesses to react to what their customers are going to do rather than what they have already done. It begins by identifying a problem or opportunity -losing too many valuable customers or a need to grow the customer base, for example - and possible actions to address it. It models each customer's likely response to various possible actions and helps the marketer choose the most appropriate action for each of those customers.

Customer behavior modeling can play an important role in an integrated approach to marketing that encompasses risk, retention, revenue and response modeling.

* Risk. Companies have a need to manage risk, especially in the area of credit. If they extend credit to a particular customer, how likely is that customer to default? What is his or her potential profit contribution? How does the risk profile vary according to the nature of the credit extended? How does this customer's risk profile correlate with that of other customers in the company's portfolio? Predictive behavior modeling makes it possible to answer those and related questions..

Good high-tech marketing involves a marriage of art and science.

* Retention. This starts with a determination of how likely each customer is to defect over a given period of time, but perhaps more important is the likely effect of various actions the company might take to retain that customer. It makes no sense to spend money on an action that will have not have a positive effect on retention. The value of the customer and the net revenue at risk if that customer defects must also be considered. Combining all those factors enables a company to target customers with effective retention actions that make sense from an ROI perspective, if that is the company's goal.

* Revenue. Customers must be managed in a way that maximizes revenue contribution to the organization. Achieving that requires an understanding of historical revenue patterns, likely future revenue, net revenue and margin. More importantly, it requires a reliable means of projecting the effect different actions are likely to have on the value of different customers. A mistake many companies make is concentrating all their efforts on extracting even more revenue from their best customers. Not only can this be an unproductive use of resources, it risks alienating those customers. The greatest potential for increased value is often to be found in second-and third-tier customers.

* Response modeling. Every time a company considers a direct communication or other management decision affecting a customer, it should first create a model of the likely outcome of that action. A simple example is, "Should we be sending this piece of mail?" It will certainly have a cost, but it may or may not have an effect. If it does have an effect, will it be positive or negative? Is one direct mail piece more likely than another to have the desired effect? Which channel is most appropriate for this communication?

Predictive behavior modeling integrates with other aspects and components of broader marketing solutions. It helps marketers understand customers, positioning the organization to maximize its likely return from each customer. While campaign management tools allow companies to manage and track marketing campaigns, predictive modeling helps companies understand whom to target, with what message or offer, at what time and through which channel.

Data warehouses allow about customer characteristics and behaviors, but if that data just sits in the warehouse, it's a wasted investment. Behavior modeling lets companies leverage their investments in data warehousing by putting that data to work improving the value contribution of existing and prospective customers.

Today's marketing technology opens up new opportunities for businesses.

The type of information provided by predictive marketing software can be very useful to senior management for strategic planning. For example, customer modeling can alert a company to changes in the behavior of its customer base long before they might otherwise become apparent. Applied to product development, behavior modeling lets companies aggregate previous experience with components being used from existing products to predict customer reaction to new products. It can also be used to make scientific trials and product testing more effective.

Customer relationship management (CRM) itself is one of the hottest business trends to emerge in years, and with good reason. The very notion of businesses managing their relationships with customers makes so much sense, it borders on the simplistic. To some extent, companies have always tried to do this, but new technologies developed in recent years now offer options that never existed before.

There are two ironclad arguments for companies to adopt CRM programs. First, study after study. has demonstrated that it is much more expensive - five to 10 times more - to recruit a new customer than to do business with an existing one. Second, customer loyalty has become an increasingly scarce commodity, thanks at least in part to the global competition the Internet makes possible.

Today's CRM programs tend to be elaborate technology solutions. If not managed properly and aligned with a company's basic strategic objectives, CRM systems can become an end unto themselves - and a massive black hole for capital resources. At its core, CRM boils down to two basic propositions: knowing your customers and targeting them with consistent and appropriate communications.

When a CRM program is well designed and implemented, it can provide substantial benefits in three ares: cost savings, revenue enhancement and strategic impact. Annual sales increases of up to 10% per employee due to improved productivity are not uncommon in the first three years of a CRM program. Win rates improve, since better information supports earlier disengagement from unlikely or unfavorable deals. The improved customer knowledge CRM enables can result in less discounting.

CRM programs come in all shapes and sizes and can be tailored to meet the needs of almost any type of business. Initiating a CRM program should be a structured process and should include the following considerations:

Customer relationship is one of the hottest business trends to emerge in years.

* Identify which functions should be automated. Use a CRM audit that lists the business functions that need to be automated and the technology required.

* Choose partners and technology systems carefully. There are hundreds of vendors in the CRM marketplace. Make sure you choose the one that has the right technology to meet your company's needs. Make equally sure that the vendor's personnel have the experience and expertise not only to complete a successful implementation, but also to provide whatever level of ongoing support you will require.

* Make CRM an enterprise-wide initiative. Involve users early in the process to make sure the system addresses their needs.