• Customer acquisition (often via targeted marketing)
• Customer retention (minimizing churn)
• Cross-sell and up-sell (extending the relationship)
• Customer win-back (recovering customer lost to competitors)
• Customer support (increasing customer satisfaction)
• Maximizing lifetime value (repeat purchase, subscription services).

Succesful CRM has two foundations:

• relevant, timely data
• powerful analytical tools

To date, CRM vendors have focused almost exclusively on gathering data. The goal has been to obtain a 360-degree view of a customer; that is, instant access to information across all customer touch points (stores, call centers, web sites). In other words, vendors seek to have a comprehensive record of every interaction between the customer and the firm. While this initial focus on data acquisition and organization is important, it is also important to know what to do with the information you have gathered. The full value of such rich data stores cannot be realized without using modern data mining tools.

Why Data Mining?

Data Mining represents a quantum leap forward in the sophistication and power of analytical tools. The best data mining tools are capable of discovering patterns in data in a few hours that expert human quantitative analysts might not find in years of work. These tools can automate many phases of data analysis, radically increasing both the productivity and the accuracy of human researchers. In one telling example, a major direct mail marketing firm compared the performance of our CART decision tree to their in-house models developed by six Ph.D.-trained quantitative analysts over a period of two years. The firm believed it had perfected its targeted marketing models. It took one analyst working one day with CART to match their best performance. Had they managed to perfect their models as they claimed? Perhaps. Could they have achieved the same or better results with far less expense and effort? Absolutely. Similar experiences have been reported by many other users.

How do our Data Mining Tools integrate with other Software and Systems?

Salford data mining tools are extremely friendly to other software systems. CART, MARS, TreeNet,and RandomForests can read from almost any data base format and can write results back to your data base. Most users simply run our software side-by-side with their other systems. We can also make an API available to allow you to easily integrate our software into larger corporate structures. Serviceable integration can be achieved by one programmer in a few days and full integration, including customized interfaces and customized reports, in a few weeks.

Why Salford Systems?

We have been developing data mining software since 1990 (before it was called data mining), working with the world's best machine learning researchers at UC Berkeley and Stanford the entire time. We have superbly trained analysts who know software as well as mathematics, and we have extensive consulting experience. These knowledgeable analysts provide a superior expertise, as shown by our winning the KDDCup 2000 international data mining competition, beating all other major vendors for first place finishes in two key categories (Best CRM Analysis, Most Accurate Models).

How do you get started with Data Mining?

If you are already gathering and organizing your customer data, you are ready to start immediately. Using our products, you can begin developing models to address any of the topics listed above. If you have not yet pulled your customer data together, useful analyses still can be conducted, although they will not be as predictive as models based on more comprehensive data stores.

For further information, download our white papers or try our software on extracts from your own data. Walkthroughs showing how you might conduct an analysis can be followed for both CART decision trees and MARS predictive models. Fully functional evaluation versions of the software can be downloaded for immediate use.