Analytic CRM and Database Marketing

Database marketing is an important part of Analytic CRM. In a plain language, database marketing is a marketing technique that utilizes customer databases. Formally, database marketing is a form of one-to-one direct marketing in which databases of customers (or potential customers) are used to generate personalized communications in promoting products or services. Due to the richness of customer information, database marketing places emphasis on the use of analytic methods in selecting or targeting customers for marketing purposes. Database marketing may include;

Analytic Database Marketing Techniques

A prominent feature of database marketing is that there is readily available rich information about customers. Customer databases can become information gold mines. Analyzing them can lead to new marketing opportunities, thereby deepening customer relationships. Common analytic techniques used in database marketing are profiling, segmentation, and scoring. They are described in the subsequent sections.

Customer Profiling

Understating customers is very important in any business. Profiling customers can lead to better understanding about customers. For example, credit/insurance risk levels, marketing responsiveness, churn risk level, and so on. Although customer profiles may be invaluable for conducting businesses, obtaining them may not come easily. The main problem is with the size of customer information. There many dozens or hundreds of customer data fields. Without advanced software tools, analyzing them and extracting most relevant information is a non-trivial task. CMSR Hotspot Profiling is an advanced tool for analyzing complex customer information and extract most relevant information.

Customer Database Segmentation

People with similar attributes tend to exhibit similar tendencies in purchasing patterns. This leads to development of various segmentation techniques. Segmentation is performed in a way that keeps customers of a segment to have similar attributes (or profiles), while customers in different segments have dissimilar attributes. The intra-group homogeneity and the fact that people with similar attributes are likely to respond somewhat similarly to a given marketing strategy lead to segmentation marketing strategies customized for specific segments. The following figures show CMSR Neural Clustering Segmentation tools. The figures clearly show that similar customers are clustered within or nearby segments.

Database Marketing Methods

Database marketing techniques can be applied to different marketing approaches. The following sections describe various database marketing methods.

Direct Marketing - Mails, Emails, SMS, Telemarketing

Direct marketing is a form of marketing in which marketers promote products directly to customers. Common forms of direct marketing include postal mails, emails and telemarketing. The first two send catalogs to customers. The first one involves sending catalogs to customers through postal mails, a costly marketing method. The last involves more cost than other two. The cost includes not only phone calls but also costly human labor works.

The problem with direct marketing is with the fact that success rates of direct marketing is very low. For example, some survey suggests that national average of catalog sales success ratio is about 2%! With such success ratio, selling low profit margin products through direct marketing may not be feasible. Analytic methods that can select customers who are more likely to buy products are needed. The following techniques can be used in selecting customers;

• Customer profiling
• Customer segmentation
• Customer scoring

These techniques can reduce marketing cost by eliminating customer groups who are unlikely to place orders. The techniques are extensively described in Direct Mail Marketing and RFM Marketing. Details can be found from the linked pages.

Cross-selling and Basket analysis

Cross-selling is to sell other products to existing customers. To increase the success rate, other products tend to be co-products or related products. For example, a customer who bought a pocket MP3 player will be highly interested in buying rechargeable batteries along with a charger. The following chart shows customer purchasing behaviors;

Cells in red color indicate that there is positive relationship between two products. That is, when customers buy one product, they tend to buy the other product as well. Cells in blue color indicate the opposite. When customers buy one product, they tend to not buy the other product. Brightness of cells indicates the relative strength of relationships. The chart shows that the following product pairs have strong positive relationship;

• Product P ~ Product O
• Product P ~ Product F
• Product O ~ Product F

Also notice that most product pairs shows negative relationship (in blue). Knowing negative relationship can prevent from wasteful marketing efforts! The strongest pairs as follows;

• Product E ~ Product H
• Product G ~ Product J
• Product G ~ Product H

Up-selling

Up-selling refers to selling more expensive products or services to existing users of the same type items. The normally involves upgrades or replacements to more expensive products or services. It is noted that up-selling is a bit tricky business. Selling more expensive products to customers who cannot afford is no business. Rather it can irritate customers and lead to defection to rival providers. (See next section "Customer Retention"). So careful sensible customer targeting is needed! The following methods can be applied;

• Customer profiling
  Profiles of customers who are using higher-cost-version products are developed. The profiles are then used to identify customers who are in the profiles but still using lower-cost-version products. For obvious reasons, they are candidates for up-selling marketing efforts.
• Customer scoring
  Customers are given scores 0 and 1 depending on product versions they are using, i.e., for low-cost-version users are given 0 and the others 1. With the scores, a neural network predictive model is developed. The model is then applied to customer database records to compute predicted scores. Customers scoring high but still using low-cost-version products are targets for up-sell marketing. This method is similar to that of Direct Mail Marketing and Customer Scoring. For more information, read the linked page.
• Customer segmentation
  Segmenting customers may produce segments with many customers using higher-cost-version products. Customers using lower-cost-version in those segments can be good candidates for up-sell marketing.

Customer Retention

In many industries, customer churn is a big problem. For example, telecommunication industry has very high customer churn rate, especially in wireless services. It is noted that acquiring a new customer is several times more expensive than keeping an existing one. Therefore retaining customers is a very important management issue.

For obvious reasons, the most important strategy in customer retention is to identify groups that are likely to leave (potentially to rival providers) through performing retention (alternatively defection) rate analysis. Once they are identified, preventive measures can be developed and actions can be followed to prevent defections. The following techniques can be used to identify customer groups of defection risk;

• Defector profiling
  Develop profiles of risky groups based on demographic, geographic and psychographic attributes.
• Defection scoring
  Build neural network predictive models that can predict likelihood of defection.
• Customer segmentation
  Segment customers based on similarity in terms of demographic, geographic, and psychographic attributes.

  
  For more, please read Customer Retention.