1. Customer Relationship Management (CRM) Chapter 4 Customer Portfolio Analysis Learning Objectives Why customer portfolio analysis is necessary for CRM implementation That there are a number of disciplines that contribute to customer portfolio analysis (CPA): market segmentation, sales forecasting, activity-based costing and lifetime value estimation How the CPA process differs between business-to-consumer and business-to-business contexts How to use a number of business-to-business portfolio analysis tools The role of data mining in CPA

2. Tools used to predicting future behavriour Decision Trees The graphical model output has the appearance of a branch structure. Decision trees work by analysing a data set to find the independent variable that, when used to split the population, results in nodes that are most different from each other with respect to the variable you are tying to predict.

3. Decision Tree: A decision tree can be constructed to develop credit risk profiles of customers. This will be useful in determining whether they represent good or bad value. Target Variable: RISK Risk is dependent and other columns are independent Independent variables: debt, income and marital status In the example, each of these is a simple categorical item, each of which only has two possible values. As you examine the data, as re-presented in the cross-tabulation of the dependent variable and all the independent variables in Fig

4. Tabular form:

5. Graphical representation: Each box is a node. The top node is the root node. The data from the root node are split into two groups based on income. The right-hand, low income box, does not split any further because both low-income customers are classified as poor credit risks. The left-hand, high-income box does split further, into married and non-married customers. Neither of these splits further because the one unmarried customer is a poor credit risk and the two remaining married customers are good credit risks

6. Decision Tree: As a result of this process the company knows that customers who have the lowest credit risk will be high income and married. They will also note that debt, one of the variables inserted into the training model, did not perform well. It is not a predictor of creditworthiness Several software packages, such as CART (Classification and Regression Trees) and CHAID (Chi-squared Automatic Interaction Detection), perform decision-tree analysis

7. Neural Networks: Neural networks are another way of fitting a model to existing data for prediction purposes. The expression ‘neural network’ has its origins in the work of machine learning and artificial intelligence researchers who were trying to understand and learn from the natural neural networks of living creatures. Neural networks can produce excellent predictions, but they are neither easy to understand nor straightforward to use. Neural networks represent complex mathematical equations, with many summations, exponential functions and parameters. Neural networks need to be trained to recognize patterns in sample data sets. Once trained, they can be used to predict customer behaviour from new data. They work well when there are many potential predictor variables, some of which are redundant.

8. CPA Tools: Cunningham and Homse were among the first to develop the concept of a customer portfolio, which they suggested was useful for improving the allocation of resources between customers and for ensuring that relationships with key customers were managed more effectively. Their recommendations were based on over 800 interviews in 160 companies. They discovered that sales volume was not the only criterion that companies regarded as an important attribute of a customer. Many companies valued what the researchers called technical development customers. Fiocca advanced customer portfolio theory when he developed a two-step customer portfolio model. At the first step customers are classified according to: The strategic importance of the customer The difficulty of managing the relationship with the customer.

9. CPA Tools: The strategic importance of a customer is determined by: The value/volume of the customer’s purchases The potential and prestige of the customer Customer market leadership General desirability in terms of diversification of the supplier’s markets, providing access to new markets, improving technological expertise and the impact on other relationships. The difficulty of managing the customer relationship is a function of: Product characteristics such as novelty and complexity Account characteristics such as the customer’s needs and requirements, customer’s buying behaviour, customer’s power, customer’s technical and commercial competence and the customer’s preference to have many suppliers Competition for the account, which is assessed by considering the number of competitors, the strength and weaknesses of competitors and competitors’ position vis-`a-vis the customer.

10. Fiocca’s CPA Model:

11. Strength of customer relationship: The strength of the customer relationship is determined by: The length of relationship The volume or dollar value of purchases The importance of the customer (percentage of supplier’s sales accounted for by this customer) Personal friendships Co-operation in product development Management distance (language and culture) Geographical distance.

12. Factors influencing the customer’s attractiveness:

13. Fiocca’s 2 nd step of CPA model:

14. Factors influencing the customer’s attractiveness: Fiocca picks out three main strategies than can be used across the portfolio. Improve the strength of the relationship (cells 1, 2, 4 or 5). Hold the position (cells 3, 6 or 9). Withdraw (cells 7 and 8).