 |
 |
|
|
Home
The four most important applications of data mining methods are response optimisation, cross selling, customer retention and credit rating.
The objective of response optimisation is to achieve an optimum customer response to a certain advertising effort, in other words, an optimum selection of customers on whom a certain advertising medium is targeted. With data mining tools the former buyer behaviour of customers (such as their response when they were sent a mail order catalogue a year before) can be analysed. The buyer behaviour for the next issuing of catalogues will be predicted on the basis of this analysis, so that catalogues can be sent more purposefully to those customers who are expected to make the greatest purchases.
This procedure requires that the customers have already carried out purchases in the past. If you have but little data about a customer, data mining procedures are usually of little use. It is recommended to deal with this group of customers separately. Data mining procedures are in fact useless for the acquisition of new customers, even if the procedures are complemented with data such as e.g. micro geographical data. An adequate data stock is without exception the principle prerequisite for successful data mining. Which means, too, that the history of a customer can be traced back sufficiently far into the past.
If you want to introduce a new product to your customers - one which they have never used before - you need to be able to conclude from their past consumer behaviour what their attitude towards the new product is likely to be. Let us illustrate this with the example of the use of credit cards in the context of payment transaction. Imagine you want to arrive at a determined promotion for the use of credit cards by the customers of a bank.
The basis is formed by the entirety of the bank data, including all the account movements in the past, passbooks, stock portfolios etc., with the exception of direct information about whether or not a bank client uses credit cards. With the help of a data mining analysis you will then determine the correlations between these informations and the actual use of credit cards. The customer profile emerging can be used directly in order to choose bank clients with whom a determined promotion for the use of credit cards will be likely to show some effect.
One of the most important objectives of marketing is the patronage of customers. With data mining techniques you can find out which of your customers are in danger of being lost to you and which are not. Knowing this, you can employ adequate marketing efforts more determinedly to retrieve your customer relationships than before (without the use of data mining). We would like to illustrate this by giving an example for the loyalty of customers towards their banks.
The basis is again the entirety of the customers' bank data in their chronological order. The objective is to discover quickly if and when a customer is about to cancel his or her relationship with the bank. In reality this means that one would compare former cases of customer losses, with an interval, however, of about 3 months in between the application of the bank data and the actual getting-out. As soon as an adequate model case is found, the achieved knowledge can be applied to the current banking transactions and customers in danger of getting out be identified in time.
A major goal of order processing in B2C or B2B is credit rating. Data mining closes information gaps or completes existing credit information on your customers. Dr. Imhoff Business Consulting has comprehensive experience in building credit rating systems of largest extent. Our real time scoring machine smoothly integrates into arbitrary system environments like e.g. SAP R/3. A typical SAP R/3 implementation only takes a few days and works completely smooth and frictionless in the SAP framework.
click to proceed to Preparation...