Download - Lecture 8
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 1
Data Mining and Customer Relationship Management
“It is now conventional wisdom among marketers, and IT personnel who support them, that investments in customer data have
a demonstrable return” said Phillip Russom, director of data warehousing and Business Intelligence, Hurwits Group
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 2
Styles of data miningPredictive model
The virtuous cycle of data miningImplementing Data Mining for Better CRM
Outline
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 3
Two Style of Data Mining
Directed data mining: is a top down approach, used when we know what we are looking for.
Undirected data mining: is a bottom up approach that lets the data speak for itself.
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 4
Sometimes we do not care how the model works. It is a black box and we just want the best prediction possible.
Sometimes we want to use the model to gain insight into the data. We need to understand how the model works; it is more like a semitransparent box.
Data mining uses both black box models and semitransparent box.
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 5
Predictive Model
Prediction model answer questions such as: Who is likely to respond to to our next offer, based on the history of previous marketing campaigns?
What is right medical treatment, based on past experience?
Which machine is most likely to be the next one to fail?
Which customers is likely to leave in the next six months?
What transactions are likely to be fraudulent, based on known examples of fraud?
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 6
The Virtuous Cycle of Data Mining
The virtuous cycle consist of four major business process Identifying the business problemTransforming data into actionable resultsActing on the resultsMeasuring the results
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 7
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 8
Identify the Business Problem
Important that technical people understand what the real business needs are Do this by talking to domain experts: the people who understand the business Business people need to be kept informed of the development, so that they may make continuing contributions to the project and the focus remains on the business needs It is important to think outside the confines of domain experts' knowledge to understand the real problem
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 9
Identify the Business Problem
Business people's expertise allows you to answer the following questions:
Is the data mining effort necessary? Should we focus on a particular segment or subgroup? What are the relevant business rules? What do they (the business people) know about the data? What do their intuition and experience say is important?
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 10
Identify the Business Problem
Checking the opinion of domain experts:
The data can be used to check that the experience and intuition of domain experts is correct
Example: If a domain expert believes that the best customer is aged 24 to 31 with at least one university degree, we can check to see if this is supported by the data.
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 11
Transforming Data into Actionable Results
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 12
Transforming Data into Actionable Results
Identifying and obtain dataThe right data is often whatever is available, reasonably clean, and accessible Data must meet the requirements for solving the business problem Data must be as complete as possible When doing predictive modelling the data needs to be complete enough that we can determine the outcome of what we are modelling
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 13
Transforming Data into Actionable Results
Validate, explore and clean the dataIs there any missing data and will this be a big problem? Are the field values within legal bounds? Are the field values reasonable? Are the distributions of individual fields explainable? Data fields which are not used are often inaccurate compared to critical data fields
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 14
Transforming Data into Actionable Results
Transport the data to the right granularity
Granularity is the level of the data that is being modelled
Some data mining algorithms work on individual rows of data, so all data describing a customer must be in a single row
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 15
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 16
Transforming Data into Actionable Results
Add derived variables
Derived variables are calculated based on combinations of other values inside the data
Prepare the model set
Model set: data used to build the data mining models
The model set can be divided into training, test, and evaluation sets
Choose the modelling technique and train the model Check performance of the models
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 17
Acting on the Results
Insights: new facts learned during modelling may lead to insights about the customers and about the business One-time results: results may be focused on a particular activity and that activity should be carried out Remembered results: information in the results should be accessible through a data mart or a data warehouse Periodic predictions: periodically score customers to determine what ongoing marketing efforts should be Real-time scoring: model may be incorporated into another system Fixing data: may have to fix data problems that have been uncovered
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 18
Implementing Data Mining for Better CRM
Data mining provides companies with the ability to specifically target customer segments and their respective needs.
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 19
Implementing Data Mining for Better CRM
Identifying the customer: there are three levels of customersTransaction levelAccount levelComplete customer
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 20
Implementing Data Mining for Better CRM
The development and implementation of a successful data mining solution can be broken down into 5 steps:Setting goalsData collectionData preparationAnalysis and predictionMeasurement and feedback
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 21
Implementing Data Mining for Better CRM
Goals
Customer profiling and segmentation
Retention and attrition
Risk avoidance
Cross-sell
Shopping patterns
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 22
Data Collection
Types of data
Which data?
Implementing Data Mining for Better CRM
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 23
Implementing Data Mining for Better CRM
Data preparationData qualityData preparation
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 24
Analysis and prediction
Building the model
Testing the model
Score
Implementing Data Mining for Better CRM
04/10/23 01:27 PM Ibrahim Elbeltagi Info 2007 25
Measurement and feedback
How well your campaign perform?
Are groups with the highest predicted likelihood respond at a significantly higher rare than the average?
Make sure that this information is returned as a component of your promotional history table in your database and used as learning experience.
Implementing Data Mining for Better CRM