Page 134 - DMGT308_CUSTOMER_RELATIONSHIP_MANAGEMENT
P. 134
Unit 5: Closed Loop Marketing
2. Forecasting classification: For each of the identified segments and desired levels of income Notes
forecast (i.e. low, medium, high), a forecasting algorithm is created (i.e. Chaid, RRNN)
that will determine the probability that a customer belongs to each of the categories.
3. Distribution and transformation analysis: Once the probability of belonging to each
level of income for each customer is calculated, a comparison of the income distribution
over the population will be performed. This income will be calculated with direct methods,
whose final value has been forecasted by data mining.
Customer Value Calculation
The methodological procedure of the customer value calculation is developed by taking into
account the present positioning of each customer and the future projections derived of the
potential increase.
Figure 5.6: Total Customer Value
Customer value Potential value coming from the
based on estimated evolution of the
customer
actual
products/services
with the company
Source: http://misbridge.mccombs.utexas.edu/knowledge/topics/crm/&docid=AjVlEPe06p17D
M&imgurl
From a strategic point of view, the breakdown of the total customer value into the present and
potential value will allow the definition of differential commercial strategies.
There are two variants in the potential value calculation: complete and derived, understanding
that derived is the value of the customer, also considering the possible risk of attrition.
!
Caution The final component in any data mining algorithm is the data management
strategy; the ways in which the data are stored, indexed, and accessed.
Task Explain the applicability of data mining techniques.
Self Assessment
Fill in the blanks:
6. Data mining expectations need to be …………………….
7. Neural networks are non …………………….
LOVELY PROFESSIONAL UNIVERSITY 129
