Hour 7: Business Intelligence & ERP

1
Hour 7Business Intelligence ERP

• ERP offers opportunity to store vast volumes of
  data
• This data can be data mined
• Customer Relationship Management

2
Data Storage Systems

• Data Warehousing
• Orderly accessible repository of known facts
  related data
• Subject-oriented, integrated, time-variant,
  non-volatile
• Massive data storage
• Efficient data retrieval
• CRM one data mining application
• Can use all of this data
• Common ERP add-on

3
Granularity

• Definition level of detail
• Most granular each transaction stored
• Averaging aggregation loses granularity
• Data warehouses usually store data at fine levels
  of granularity
• You cant undo averages aggregates

4
Data Marts

• Different definitions
• Small version of data warehouse
• Temporary storage of data
• possibly from multiple sources
• for a specific study

5
On-Line Analytic Processing

• OLAP
• Multidimensional databases
• Display data on selected dimensions
• Time
• Region
• Product
• Department
• Customer
• Etc.

6
Data Quality

• Problem causes
• Data corrupted or missing
• Failure of software transferring data into or out
  of data warehouse
• Failure of data cleansing process

7
Data Integrity

• No meaningless, corrupt, or redundant data
• Part of data warehousing function to clean data
• Data standardization
• Remove ambiguity (different ways to abbreviate)
• Matching
• Associating variables (unique mapping)

8
Database Product Comparison
9
Data Mining

• Analysis of large quantities of data by computer
• Micromarketing
• Versatile
• Apply to a wide variety of models
• Scalable
• Can analyze very large data sets

10
Types of data mining

• Hypothesis Testing
• Traditional statistics
• Knowledge Discovery
• No predetermined expectation of relationships

11
Business Data Mining Applications
12
Customer Relationship Management

• Determine value of customer
• Identify what they want
• Package products (services) to keep them
• Maximize expected net present value of customer

13
Data Warehouse Use

• Wal-Mart
• Fingerhut

14
Wal-Mart Data WarehouseFoote Krishnamurthi
2001

• Wal-Mart dominates retail market
• Heavy user of information technology
• Supply chain distribution to 2,900 outlets
• A critical success factor
• Data warehouse of 101 terabytes
• Possibly worlds largest
• Investment over 1 billion
• Can handle 35,000 queries per week
• Benefits over 12,000 per query

15
Wal-Mart

• Initial data warehouse
• point-of-sale shipment data
• Added data
• Inventory
• Forecast
• Demongraphic
• Markdown
• Return
• Market basket information

16
Wal-Mart Data Warehouse

• Process 65 million transactions per week
• 65 weeks of data per item
• By store
• By day
• Support decision making
• Many users have access
• Including 3,500 vendor partners

17
FINGERHUT

• Founded 1948
• today sends out 130 different catalogs
• to over 65 million customers
• 6 terabyte data warehouse
• 3000 variables of 12 million most active
  customers
• over 300 predictive models
• Focused marketing

18
Fingerhut

• Purchased by Federated Department Stores for 1.7
  billion in 1999 (for database)
• 2002 more recent developments
• Fingerhut had 1.6 to 2 billion business per
  year, targeted at lower-income households
• Can mail 400,000 packages per day
• Each product line has its own catalog

19
Fingerhut

• Used segmentation, decision tree, regression,
  neural network tools from SAS and SPSS
• Segmentation - combined order demographic data
  with product offerings
• could target mailings to greatest payoff
• customers who recently had moved tripled their
  purchasing 12 weeks after the move
• send furniture, telephone, decoration catalogs

20
Advanced Technology ERP

• Bolt-ons
• Middleware
• Security

21
Technology ERPManetti 2001

• Mobile commerce other IT makes ERP extensions
  possible, attractive
• Broader use of web-enabled systems
• Greater AI-driven applications
• Greater use of ERP in mid-sized manufacturing
• Flexible modular systems
• More bolt-ons (3rd party applications)
• Creates security issue

22
Conflict ERP Open Systems

• Original concept of ERP closed
• Easy to control access
• Openness creates security issues
• But there are too many good things to do with
  open systems
• ERP vendors also provide such products

23
Example Bolt-OnsMabert et al. 2000
24
Middleware

• ERP interfaces to external applications difficult
  to program
• Middleware is an enabling engine to allow such
  external applications eto ERP
• Data oriented products - shared data sources
• Messaging-oriented - direct data sharing

25
Web ERP

• J.D. Edwards OneWorld
• SAP mySAP.com
• Trends
• More web links
• More functionality

26
Middleware Data Acquisition

• Bar-code data collection
• Radio frequency data collection
• Web portals

27
Portals of Major ERP VendorsStein Davis
1999 Stein 1999
28
Other Vendor PortalsStein Davis 1999
29
ERP Security Threats
30
Summary

• ERP security originally was not problematic
• Only few internal users could access
• Open systems driven by external applications
• Creates security issues
• Web access especially problematic
• Special ERP Security aspects
• Data quality
• Control over data access

31
Bolt-On/Middleware Examples

• Kellogg Company Brown et al. 2001
• Dow Corning Teresko 1999

32
Kellogg Company Bolt-On

• Kellogg developed their own ERP
• Forecast demand
• Take customer orders
• Coordinate raw material purchasing
• Coordinate production of over 100 food products
• Coordinate distribution
• Added linear programming Kellogg Planning System
  (KPS)
• Production, inventory, distribution planning
• Budgeting capacity expansion

33
History

• Long user of MRP, DRP (distribution resource
  planning)
• 1987 realized product line growth, international
  expansion led to need for more computer support
• Developed KPS in 1989, modified over time
• By 1994 strong cost system in place
• Saved 4.5 million in 1995

34
Kellogg LP

• Minimized total cost
• Purchasing, manufacturing, inventory,
  distribution
• Variables product, package size, case size
• 30 week planning horizon
• Constraints
• Line, packaging capacities, flow constraints,
  inventories, safety stocks
• 700,000 variables, 100,000 constraints, 4 million
  non-zero coefficients

35
Kellogg LP

• Continuous model took several hours to run
• Generated starting solution for managers
• Probabilistic features dealt with through safety
  stock
• Example of bolt-on to ERP
• Linear programming generated better plans

36
Dow Corning System Integration

• 1995 adopted SAP R/3 to integrate global business
  practices
• Also adopted SAP data warehouse
• Consolidated information generated internally,
  externally
• Internal plant-floor data, patent information,
  benchmarking
• Allowed deeper data analysis

37
Dow Corning System

• Over 4,000 users had access
• Integration data compatibility problems dealt
  with by data warehouse
• Added automated data collection system
• Required middleware
• Middleware allowed expansion into supply chain
  management

38
Summary

• Customer Relationship Management very promising
• Has not reached all expectations as ERP add-on
• Quite expensive to get needed data storage
  capability
• Still an opportunity to use all the data
  generated by an ERP
• Many other useful bolt-ons