By Raza Inam Khan, Systems Consultant
Teradata, a division of NCR

Jan 06 - 12, 2003


We know the built-in benefit of successful data warehousing: the ability to collect, store, analyze and act on massive amounts of critical customer data. That's the good news. Today's challenge? To take the business intelligence locked inside a data warehouse to the next level, to turn it into the kind of predictive insight that separates your business from your competitors. But uncovering that insight means removing a barrier of intensive computational steps of data preparation, pre-analysis and analysis of huge volumes of detail data.

Data mining enables you to efficiently apply analytic techniques to discover and interpret a variety of data patterns that represents customer behavior and business trends. It helps businesses answer complex questions by discovering previously unrecognized relationships from hidden information, including:

*Who are your most profitable customers? And why are they so profitable?
*How can you increase your levels of customer satisfaction, loyalty and lifetime value?
*Which customers will purchase across your product base?
*Which channel is the most effective way to reach each customer?
*How can you respond to each customer as an individual?
*How can you predict sales trends and business drivers that affect sales?

Data mining allows you to get to know your customers on an individual level, and use that knowledge to retain them and enhance their profitability. And there's more. Data mining also provides you with a better understanding of business trends - added knowledge that lets you react dynamically to the changing market. With insight into your Customer's behavior you can:

*Personalize customer interactions and react dynamically to their needs.
*Deliver more effective targeted marketing and cut costs by reaching out to those most likely to respond.
*Design winning strategies and make decisions based on real-world business conditions.


Teradata Data Mining, is a predictive analysis solution that analyzes detail data. It uses intelligent software to detect patterns through an exhaustive analysis of all the data in your warehouse. The result: the ability to integrate data mining insight into all your analytic applications, such as Customer Relationship Management (CRM), to predict future customer behavior. Data mining supercharges your analytic applications by providing:

*Scores - the likelihood a customer will purchase a product
*Predictions - how much a customer will spend in the coming year
*Descriptions - what characteristics define profitable customers
*Profiles - the common characteristics of each customer segment
*Business drivers - the key characteristics that identify fraudulent behavior


Teradata Data Mining brings you one more significant advantage - the ability to have a clear view of customer behavior across your total enterprise. Built on the foundation of Teradata Warehouse technology, this solution overcomes the massive processing bottlenecks caused by analyzing ever increasing volumes of data. In fact, the Teradata Warehouse provides a powerful infrastructure for handling complex analytic processing of thousands of variables and millions of transactions - every day. Teradata is the first to integrate data warehousing and data mining, a combination that allows you to perform cross-channel analysis against integrated customer data from the web, call centers, demographics, transactions and usage. And we've built the Teradata Warehouse to allow you to handle the rapid growth and change that come with increased data, user volumes and more frequent updates. In addition, Teradata tools and utilities streamline the loading of fresh, clean data that is critical to building accurate models and leveraging knowledge for dynamic customer interaction.


Teradata Warehouse Miner is a tool that provides a comprehensive set of data mining functions. It enables you to build and deploy analytic models directly in the Teradata Database. While, traditionally, businesses have been limited to mining small samples of enterprise data, the Teradata solution allows you to analyze the growing volume of detail data required in today's environment of increasingly complex questions. Teradata Warehouse Miner allows analytic applications to scale up to the complexity of business problems, then delivers results in the form of scores, predictions and descriptions. And, just as important, those results can be managed in a single location, providing your entire enterprise with access to the information they need for more informed decision making. Teradata pioneered "in place" data mining technology that allows you to improve data mining by:

•Eliminating data movement - there's no need for multiple iterations of data extraction for sampling, exploration, model development and scoring.
•Minimizing errors - our centralized data mining architecture eliminates errors that come with data movement across platforms, tools and databases.
•Simplifying data management - that same centralized architecture provides a single version of the truth and eliminates redundancy.
•Leveraging the parallelism of the Teradata Database - leverage the high performance and scalability of the Teradata Database.
•Building accurate models - enabling analysis of large volumes of detail data.





Customer Segmentation

What market segments do my customers fall into and what are their characteristics?

Personalize customer relationship for higher customer satisfaction and retention.

Propensity to Buy

Which customers are most likely to respond to my promotion?

Targeting customers based on their needs to increase their loyalty to your product line. Also increase campaign profitability by focusing on those most likely to buy.

Customer Profitability

What is the life time profitability of my customer?

Make individual business interaction decisions based on overall profitability of customers.

Fraud detection

How can I tell which transactions are likely to be fraudulent?

Quickly detect fraud and take immediate action to minimize cost.

Customer Attrition

Which customer is at risk of leaving?

Prevent loss of high-value customers and let go of lower value customers.

Channel Optimization

What is the best channel to reach my customer in each segment?

Interact with customers based on their preference.


• Descriptive Statistics provide a variety of descriptive statistical functions, giving the analyst detailed knowledge of the data to be analyzed. These functions are also used to uncover data quality issues that can jeopardize the accuracy of any analytic model.
• Transformation Components provide a variety of column transformation that are useful prior to designing the analytic model.
• Data Reduction Functions reduce the amount of data required for analytic algorithms. Data reduction matrices include correlation, co-variance and other often used analytic algorithms.
• Data Visualization provides graphics and charts for interpreting results.
• Organization & Partitioning Functions enable restructuring of tables and generation of samples and partitions.
• Multivariate Statistical Analysis provides Linear Regression, Factor Analysis and Logistic Regression techniques to build analytic models.
• Machine Learning Algorithms provide Decision Tree/Rule Induction and Clustering Algorithms to build analytic models.
• Model Deployment allows use of analytic models directly in the Teradata Database.