Data Mining in Customer Relationship Management

1
Data Mining in Customer Relationship Management

• Presented by Tim Fletcher
• April 2005

2

• One of the fastest growing areas of the
  applications of decision sciences technologies
  and data mining is Customer Relationship
  Management (CRM). Berson, 44

3
Differentiating CRM from BI

• Data warehousing is not CRM and neither is
  business intelligence. Dyche, 15
• The ability fo access data is , by itself,
  immensely powerful, but many business
  intelligence environments simply use data to
  confirm already held hypotheses. The mandate of
  CRM is the ability to act on that data and to
  change fundamental business processes to become
  more customer-centric. Dyche, 16

4
Problems CRM tools address

• Identifying most profitable customers
• Leaky bucket syndrome
• Which promotions work for which people
• Closing the information loop
• Increasing the customers perceived value

5
CRM Components

• Marketing Force Automation
• Sales Force Automation
• Customer Service
• Operational CRM
• Analytical CRM

6
DM and DW in CRM

• Dos
• Integrate customer data
• Build DW with a requirement-based approach
• Act upon data and analysis
• Define CRM Metrics

• Donts
• Make decisions based on DM without business
  knowledge
• Run DM on operational databases
• Let your models become dated

7
DM for effective CRM

• DM for knowledge discovery
• Define business problem
• Build data mining DB
• Explore data
• Prepare data for modeling
• Build model
• Evaluate model
• Deploy model and results

1. Define business problem
2. Build Marketing DB
3. Explore data
4. Prepare data for modeling
5. Build model
6. Evaluate model
7. Deploy model and results

Source Two Crows Corporation
8
Types of DM most often applied to CRM

• Prediction models that predict future actions
  or behaviors based on historical results
• Sequence identify the sequence of events that
  trigger a customer action
• Association ex. Market-basket analysis

9
Click-stream Analysis

• Log a persons actions on a website
• May be stored in a separate data webhouse then
  the corporate
• Examine traits of abandoned shopping carts
• Combine click-stream analysis with customer data

understanding a customers navigation around a
site can help a company decide how to lure him
back Dyche, 137
10
Personalization and Collaborative Filtering

• Provide custom content
• Rules-base
• If hard coded can be hard to maintain for new
  products
• Collaborative filtering learns trends in similar
  purchasing behavior to target customers
• Amazon does this extremely well
• Dynamic pricing

11
Recognizing a need for DW to support CRM

• Union Bank of Norway
• 1990s success in driving customers to more cost
  efficient channels decreased in-branch bank
  staffs knowledge of customer
• Data dispersed across multiple systems
• Needed single 360-degree view of the customer
• Had Teradata build a data warehouse

12
CRM Trends

• Integrating CRM tools ERP tools
• Largest growth in On-demand CRM tools

13
References

• Berson, A., Smith, S., et al. Building Data
  Mining Applications for CRM. McGraw-Hill, 2000.
• Dyche, J. The CRM Handbook A business Guide to
  Customer Relationship Management. Addison Wesley,
  2001.
• Edelstein, H. Building Profitable Customer
  Relationships With Data Mining, Two Crows Corp.
  2000.
• Herbert, L. The Forrester Wave Hosted
  SalesForce Automation, Q1 2005. 2005.
• Introduction to Data Mining and Knowledge
  Discovery, Two Crows Corp. 1999.
• Tiwana, A. The Essential Guide to Knowledge
  Management E-Business and CRM Applications.
  Prentice Hall PTR, 2001.