Data Mining

Techniques for CRM

Seyyed Jamaleddin Pishvayi

Customer Relationship Management

Instructor : Dr. Taghiyare
Tehran University
Spring 1383
Outlines
 What is Data Mining?
 Data Mining Motivation

 Data Mining Applications

 Applications of Data Mining in CRM

 Data Mining Taxonomy

 Data Mining Techniques

2
Data Mining
 The non-trivial extraction of novel, implicit, and actionable
knowledge from large datasets.
 Extremely large datasets
 Discovery of the non-obvious
 Useful knowledge that can improve processes
 Can not be done manually
 Technology to enable data exploration, data analysis, and data
visualization of very large databases at a high level of
abstraction, without a specific hypothesis in mind.
 Sophisticated data search capability that uses statistical
algorithms to discover patterns and correlations in data.

3
Data Mining (cont.)

4
Data Mining (cont.)
 Data Mining is a step of Knowledge Discovery in
Databases (KDD) Process
 Data Warehousing
 Data Selection
 Data Preprocessing
 Data Transformation
 Data Mining
 Interpretation/Evaluation
 Data Mining is sometimes referred to as KDD and
DM and KDD tend to be used as synonyms
5
Data Mining Evaluation

6
Data Mining is Not …
 Data warehousing
 SQL / Ad Hoc Queries / Reporting

 Software Agents

 Online Analytical Processing (OLAP)

 Data Visualization

7
Data Mining Motivation
 Changes in the Business Environment
 Customers becoming more demanding
 Markets are saturated
 Databases today are huge:
 More than 1,000,000 entities/records/rows
 From 10 to 10,000 fields/attributes/variables
 Gigabytes and terabytes
 Databases a growing at an unprecedented rate
 Decisions must be made rapidly
 Decisions must be made with maximum knowledge
8
Data Mining Motivation
“The key in business is to know something that
nobody else knows.”
— Aristotle Onassis

PHOTO: LUCINDA DOUGLAS-MENZIES

PHOTO: HULTON-DEUTSCH COLL

“To understand is to perceive patterns.”

— Sir Isaiah Berlin

9
Data Mining Applications

10
Data Mining Applications:
Retail
 Performing basket analysis
 Which items customers tend to purchase together. This
knowledge can improve stocking, store layout strategies, and
promotions.
 Sales forecasting
 Examining time-based patterns helps retailers make stocking
decisions. If a customer purchases an item today, when are they
likely to purchase a complementary item?
 Database marketing
 Retailers can develop profiles of customers with certain
behaviors, for example, those who purchase designer labels
clothing or those who attend sales. This information can be used
to focus cost–effective promotions.
 Merchandise planning and allocation
 When retailers add new stores, they can improve merchandise
planning and allocation by examining patterns in stores with
similar demographic characteristics. Retailers can also use data
mining to determine the ideal layout for a specific store. 11
Data Mining Applications:
Banking
 Card marketing
 By identifying customer segments, card issuers and acquirers
can improve profitability with more effective acquisition and
retention programs, targeted product development, and
customized pricing.
 Cardholder pricing and profitability
 Card issuers can take advantage of data mining technology to
price their products so as to maximize profit and minimize loss of
customers. Includes risk-based pricing.
 Fraud detection
 Fraud is enormously costly. By analyzing past transactions that
were later determined to be fraudulent, banks can identify
patterns.
 Predictive life-cycle management
 DM helps banks predict each customer’s lifetime value and to
service each segment appropriately (for example, offering
special deals and discounts).
12
Data Mining Applications:
Telecommunication
 Call detail record analysis
 Telecommunication companies accumulate detailed call
records. By identifying customer segments with similar use
patterns, the companies can develop attractive pricing and
feature promotions.
 Customer loyalty
 Some customers repeatedly switch providers, or “churn”, to
take advantage of attractive incentives by competing
companies. The companies can use DM to identify the
characteristics of customers who are likely to remain loyal
once they switch, thus enabling the companies to target
their spending on customers who will produce the most
profit. 13
Data Mining Applications:
Other Applications
 Customer segmentation
 All industries can take advantage of DM to discover discrete
segments in their customer bases by considering additional
variables beyond traditional analysis.
 Manufacturing
 Through choice boards, manufacturers are beginning to
customize products for customers; therefore they must be able to
predict which features should be bundled to meet customer
demand.
 Warranties
 Manufacturers need to predict the number of customers who will
submit warranty claims and the average cost of those claims.
 Frequent flier incentives
 Airlines can identify groups of customers that can be given
incentives to fly more.
14
Data Mining in CRM:
Customer Life Cycle
 Customer Life Cycle
 The stages in the relationship between a customer and a
business
 Key stages in the customer lifecycle
 Prospects: people who are not yet customers but are in
the target market
 Responders: prospects who show an interest in a product
or service
 Active Customers: people who are currently using the
product or service
 Former Customers: may be “bad” customers who did not
pay their bills or who incurred high costs
 It’s important to know life cycle events (e.g.
retirement)
15
Data Mining in CRM:
Customer Life Cycle
 What marketers want: Increasing customer
revenue and customer profitability
 Up-sell
 Cross-sell
 Keeping the customers for a longer period of time
 Solution: Applying data mining

16
Data Mining in CRM
 DM helps to
 Determine the behavior surrounding a particular
lifecycle event
 Find other people in similar life stages and
determine which customers are following similar
behavior patterns

17
Data Mining in CRM (cont.)

Data Warehouse Customer Profile Data Mining

Customer Life Cycle Info.

Campaign Management

18
Data Mining in CRM:
More
 Building Data Mining Applications for CRM
by Alex Berson, Stephen Smith, Kurt
Thearling (McGraw Hill, 2000).

19
Data Mining Techniques

Data Mining Techniques

Descriptive Predictive

Clustering Classification

Association Decision Tree

Sequential Analysis Rule Induction

Neural Networks

Nearest Neighbor Classification

Regression
20
Two Good Algorithm Books
 Intelligent Data
Analysis: An
Introduction
 by Berthold and Hand
 The Elements of
Statistical Learning:
Data Mining, Inference,
and Prediction
 by Hastie, Tibshirani, and
Friedman

21
Predictive Data Mining

Honest
Tridas Vickie Mike

Crooked

Wally Waldo Barney

22
Prediction

Tridas Vickie Mike

Honest = has round eyes and a smile

23
Decision Trees
 Data
height hair eyes class
short blond blue A
tall blond brown B
tall red blue A
short dark blue B
tall dark blue B
tall blond blue A
tall dark brown B
short blond brown B

24
Decision Trees (cont.)
hair
dark blond
red

short, blue = B short, blue = A

{tall, blue = A}
tall, blue = B tall, brown = B
tall, brown= B tall, blue = A
short, brown = B

Completely classifies dark-haired Does not completely classify

and red-haired people blonde-haired people.
More work is required
25
Decision Trees (cont.)
hair
dark blond
red

short, blue = B short, blue = A

{tall, blue = A}
tall, blue = B tall, brown = B
tall, brown= B tall, blue = A
short, brown = B
Decision tree is complete because
1. All 8 cases appear at nodes
eye
2. At each node, all cases are in blue brown
the same class (A or B)
short = A tall = B
tall = A short = B
26
Decision Trees:
Learned Predictive Rules

hair

dark blond
red

B A eyes
blue brown

A B

27
Decision Trees:
Another Example

Total list
50% member

0-1 child 2-3 child 4+ children

20% member

$50-75k income $75k+ income $50-75k income $20-50k income

15% member 70% member 85% member

Age: 20-40 Age: 40-60

45% member 80% member

28
Rule Induction
 Try to find rules of the form
IF <left-hand-side> THEN <right-hand-side>
 This is the reverse of a rule-based agent, where the rules are
given and the agent must act. Here the actions are given
and we have to discover the rules!
 Prevalence = probability that LHS and RHS
occur together (sometimes called “support factor,”
“leverage” or “lift”)
 Predictability = probability of RHS given LHS
(sometimes called “confidence” or “strength”)

29
Association Rules from
Market Basket Analysis
 <Dairy-Milk-Refrigerated> → <Soft Drinks Carbonated>
 prevalence = 4.99%, predictability = 22.89%
 <Dry Dinners - Pasta> → <Soup-Canned>
 prevalence = 0.94%, predictability = 28.14%
 <Dry Dinners - Pasta> → <Cereal - Ready to Eat>
 prevalence = 1.36%, predictability = 41.02%
 <Cheese Slices > → <Cereal - Ready to Eat>
 prevalence = 1.16%, predictability = 38.01%

30
Use of Rule Associations
 Coupons, discounts
 Don’t give discounts on 2 items that are frequently bought
together. Use the discount on 1 to “pull” the other
 Product placement
 Offer correlated products to the customer at the same time.
Increases sales
 Timing of cross-marketing
 Send camcorder offer to VCR purchasers 2-3 months after
VCR purchase
 Discovery of patterns
 People who bought X, Y and Z (but not any pair) bought W
over half the time

31
Finding Rule Associations
Algorithm
 Example: grocery shopping
 For each item, count # of occurrences (say out of 100,000)
apples 1891, caviar 3, ice cream 1088, …
 Drop the ones that are below a minimum support level
apples 1891, ice cream 1088, pet food 2451, …
 Make a table of each item against each other item:

apples ice cream pet food

apples 1891 685 24
ice cream ----- 1088 322
 Discard cells below support threshold. Now make a cube for triples, etc.
pet food
Add 1 dimension -----on LHS.-----
for each product 2451

32
Clustering
 The art of finding groups in data
 Objective: gather items from a database into
sets according to (unknown) common
characteristics
 Much more difficult than classification since
the classes are not known in advance (no
training)
 Technique: unsupervised learning

33
 The K-Means Clustering
Method
10
10
10 9
9
9 8
8
8 7

6 7
7
5 6
6
4 5
5 3
4
4
Assign
2
Update 3
3 1

2 each 0 the 2

cluster
0 1 2 3 4 5 6 7 8 9 10
1
1
of the 0
0
0 1 2 3 4 5 6 7 8 9 10 objects means 0 1 2 3 4 5 6 7 8 9 10

to
most reassign reassign
similar 10 10

K=2
9 9

center 8 8

7 7

Arbitrarily choose 6 6

K objects as initial
5 5

4 4

cluster center 3

2
Update 3

1 the 1

cluster
0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

means

34
Thanks

Seyyed Jamaleddin Pishvayi

Customer Relationship Management

Instructor : Dr. Taghiyare
Tehran University
Spring 1383