DuPonts Best APO GoLives Focus on Data

1
DuPonts Best APO Go-Lives Focus on Data

• Dinesh Vandayar
• Helen Collison

2
Agenda

• Introduction
• A typical Data Implementation
• Data collection- The old way
• Why old data methodology does not play well with
  APO
• New Data methodology
• Future State
• Conclusions

3
Introduction

• DuPont
• Infinite Strategies

4
DuPont- APO Implementations

• 5 Separate Business are live today using SAP APO
• 3 Business are scheduled to go-live by Oct 2004
• All implementations were Big Bang
• R/3, APO and BW all were implemented at the same
  time

5
Big Bang- Maybe that is the problem

• We already have productive R/3 systemDoes this
  apply to me?
• My APO implementation is stand alone Does this
  apply to me?

6
Data- What priority!!

• Be Honest!!!
• Where do you rank data collection and cleansing
  among other project priorities?
• Everybody gets it.. Data is Important!!!
• Always Underestimated
• Never placed on par with design and configuration
  activity

7
Data- What is the big deal?

• Our Experience shows that the Single Greatest
  Risk for an APO Project to be Unsuccessful is
  Lack of DATA Readiness

8
Typical Issues at Go-Live

• Performance Related Issues
• SNP Optimizer or PPDS Planning runs do not
  complete in the allotted time
• It Worked during Realization and Testing- But it
  does not work now
• Variability in data used during testing and real
  production data

9
(No Transcript)
10
Problem Statement
11
A typical Data Implementation

• Large Team of people usually by work stream
• Collecting Data in spread sheets one object at a
  time
• Enormous Coordination effort between work streams
  on common objects
• Typically takes about 1 year to complete
  collecting data
• Moving Target

12
Project Timeline- Where is the Disconnect?

• Project Plan
• Blue Print Phase
• Business process to SAP application mapping
• Prototyping
• Realization Phase
• Configuration Activity
• Unit Testing
• Testing Phase
• Integration Testing

• Data Plan
• Data Team Assembled
• Design Data Templates for each master data object
• Train the unfortunate few on the design and
  collection of data using the templates
• Data Collection
• Parallel Activity to realization Phase- Data is
  being collected one object at a time
• Mock Loads
• Typically a small of real data is available
  during testing phase

13
Issues with the old methodology

• Real Data is not typically available during blue
  print/ Realization phase
• Available data does not represent real business
  process
• Typically created by configuration team
• Inflexibility in adapting to changes

14
Old Methodology- Why it does not play well with
APO?

• APO is all about DATA
• Very little configuration especially Supply
  Planning
• Planning is all about data design
• Data volume is crucial to validate design
• Design/Models is very optimistic with Managed
  Data set
• Testing and validation of design/supply chain
  model is an Iterative process

15
Old Methodology- Why it does not play well with
APO?

• R/3
• Configuration drives data design
• Availability of real data at realization phase is
  not as critical for R/3 as it is for APO

16
Our Solution
17
New Data Philosophy

• APO Realization- Only with Real data
• Data in development should represent production
  and should be a collaborative effort of all work
  streams
• Data set representing true supply chain should be
  available during Realization phase
• Actual Materials, BOMs, Recipes, Resources etc..

18
What Does Data Representing a Supply Chain Mean?

• Data representing the Supply Chain.. What does
  that mean?
• Capture Master Data as per flow of materials
  From Procurement of Raw materials to
  Manufacturing and shipping Finished Product to
  customer
• A complete set of data representing all supply
  chain scenarios that is relevant to business is
  required at the beginning of Realization Phase-
  Preferably including the most complex scenario
• If there are 4 Major Supply Chain Scenarios- 4
  complete sets of data representing each scenario
  should be available during the start of
  realization
• Data could be added to existing scenarios thru
  the progression of the realization phase
  culminating with a production set of data before
  the end of realization

19
Obvious Advantages

• Improved Overall Quality of the Development
  process
• Cross work stream issues come to light sooner

20
How do we get real data so early in the game?

• You just said it takes a year to collect and
  cleanse data
• Does it mean that data collection and cleansing
  has to start before the project is sanctioned?

21
It is all about..

• Rules

22
What do you mean by Rules?

• Data objects for any business or industry sector
  can be classified using a certain number of KEYS

23
What do you mean by Rules?

• For Example,
• Materials can be classified using the following
  KEYS
• Location
• Financial Hierarchy
• Planning Strategy
• Source of Manufacture
• Once Keys are identified- Design RULES to
  populate master data object

24
What do you mean by Rules?
25
Rules Based Approach

• Can we do RULES for any DATA Object?
• The Answer is YES
• We have successfully applied Rules based
  approach to the following objects
• Product Master
• PPMs- PPDS and SNP
• Resources
• Transportation Lanes

26
Advantages

• Rules Based Method allows to
• Make real data available during the beginning of
  Realization Phase
• Production Volume of data available during end of
  realization phase
• Planning design validated using real data-
  Quantity and Quality
• Provided flexibility to change the rules easily
• On-going maintenance is enormously simplified
• Integrated with CIF

27
Future State

• Can we apply Rules based approach to non-APO
  data objects?
• We believe we can
• Where can we apply?
• Customer master
• Vendor Master
• Material Master
• Recipes etc.

28
Future State

• Imagine.
• No 250 plus field template
• No complex load templates
• No Conversion Programs
• Not having to fret about design changes affecting
  data collection
• Ease of On-going maintenance

29
Conclusions

• Success of APO implementation is directly related
  to EARLY Data readiness
• Generating master data using RULES enables data
  to be available EARLY

30
Questions Answers

• dineshv_at_infinitestrategies.com

31
Insert session code
Session Code