Transcript Disco SA: Problems

Disco SA: Problems
• Didn’t know customer
– Needed to understand behavior of loyal customers
• Retail Analytics
• Reliant on OLTP systems and IT ppl to
solve/prepare information
– More employees needed more access to data to
try to figure out anomalies/patterns
– ‘full-service’ gas station
Disco SA: Solution
• Used data warehouse to centrally store data  data
mining used to find unknown patterns to bridge
analysis gap
– Descriptive Data Mining = finds patterns in data to
explain/describe behavior
• Segmenting = putting customers into distinct groups
• Clustering = describes customer in segment
– Few specific dimensions determine cluster
– Predictive Data Mining = finds patterns that are used to ID
trends
• Finding characteristics of customers who are likely to buy
particular product
• Decision Tree = visual & interactive model used to break data into
groups
Cascade Designs: Problem
• Diverse collection of loosely integrated
standalone applications  ‘legacy systems’
– Developed and supported by a few internal IT
people
– ‘full-service’ gas station
• System complexity limited its efficiency and
growth for company
Cascade Designs: Solution & Benefits
• ERP (Enterprise Resources Planning ) System
– Real-time INFO
– Individual Responsibility
– Better control/management of
inventories/procedures
– Better product decisions (effective choices)
• Ex: carabineers
– Discovery of 20/80 customers
Identifying BI Opportunities
• 1. Do Homework
– WHERE : BI will be used/needed
• Functional Areas = dept. of buiness (FIN, OPS, HR)
• Business Units = line of business that crosses funtions
– Cross-functional and business-unit applications have bigger
payoff potential  protecting competitive advantage
– WHO : will use BI and benefit from information
• BI at Higher Level = need for summarized data that supports
analysis of trends/patterns w/in and across functional areas
• BI at Lower Level = need detailed data that is operational in
nature and specific to functional area
BI Opportunities
• 1. Doing Homework (cont.)
– WHAT : information (measures/dimensions)
• Measures = KSF for functional areas /business units
– Base Measures = measures captured at transaction level
– Calculated Measures = computation of base measures
• Dimensions = ‘by,by,by’ = data you need for analysis
• Level of Detail = summarization can be derived from
detail
– Ex: target  POS down to hour instead of minute was
sufficient for good results
BI Opportunities
• 2. Sharing & Collecting Ideas
– Brainstorming Teams = specify measures &
dimensions
• Why ?’s  What ?’s  answers define SO
– Answering what you NEED out of system to perform
successful analysis
– BI blueprint = measure and dimension analysis
(p.127)
BI Opportunities
• Evaluating Alternatives= synthesis of BI blueprint to list
of BI opportunity areas
– Group requirements by Opportunity Areas
• Opportunity Areas = logical grouping of measure requirements w/
consistent data of dimensions
– Consistent set of requirements/data that can be used by many groups of
users
– Grade Opps by Importance
• Actionability = empowerment of employees to be able to ‘act’ on
data
• Materiality = can you save/make $$ with info
• Tactical vs. Strategic
– Strategic = LT = ‘process view’
– Tactical = ST = ‘functional view’
BI Opportunities
• Evaluating Alternatives (cont.)
• Grade Opps by Difficulty
– Cross-Functionality of Design
• Functional Opps = easy = functional view
– Used in one functional area
• Cross-Functionality = Hard = ‘process view’
– Existence & Accessibility of Data
– Complexity of Calculation =
BI Opportunities
• Evaluating Alternatives (cont)
• Rank Opps = BI Scorecard
Level of Effort--low
I = high priority/easy = GO FOR IT
II = low & med priority, easy to do =
CONFIDENT, maybe more homework
II
III = low priority/hard = case by case &
maybe pilot test
IV = high priority/hard = pilot test
I
Business
Priority-----high
III
IV
Case Summary
• Audi = used BI to help improve efficiency of
assembly line/operations
• CompUSA = used BI to help improve day-today store management and operations
• Cascade Design = used BI to help
product/inventory management & to maintain
stable workforce
• DiscoSA = used BI to enhance service to keep
customers loyal
CAN YOU IDENTIFY PROBLEM EACH COMPANY ENCOUNTERED, SOLUTION THAT WAS
IMPLEMENTED, & BENEFITS THAT CAME FROM EACH SOLUTION