Title: Mark C' Clark
1Case StudyRapid Re- implementation One Path to
Release 11i
- Mark C. Clark
- O2Works LLC
- NCOAUG Winter 2003
2Intended Audience
- Project Managers, Team Leaders
- Functional or Technical Team Members
- Any Groups considering their alternatives in
getting to Release 11i
3Goal of the Presentation
- Situation/Background of Actual Company
- Some of Their Alternative Paths to 11i
- Challenges
- Deciding Factors
- The Choice
- Approach
- Team Composition
- Challenges and Obstacles
- Lessons Learned
4What We Dont have Time to Cover
- The right or wrongs of re-implement versus
upgrade - A group methodology for making the decision
- Specific Recommendations on re-implement versus
upgrade - Past Presentations Available online -
5The Case A Real Company
- Retailer - Approx 500 Million in Revenue
- Catalog Sales Direct to Consumer
- Retail Outlets around the United States
- Implemented in 1996 with help of their hardware
vendor - AP, GL and Check Clearing Features
- Release 10.7 Character with Some ADI
6Assessment- 3 Day Crunch Session Determine
Challenges
- Technical series of automated scripts run
against production rapidly paint a picture of
where we are - Functional Pre-Meetings to dig up departmental
issues with current system, face to face sessions
with stakeholders, working sessions on some new
features - Managerial Goals of the business what should
the system be doing to meet those goals
7Their Challenges
- Hardware and DB combination de-supported
- Entirely New Mgmt team started in July 2002
- Calendar and Chart of Accounts could not meet
reporting requirements of the new team - Process Improvements required
8Challenges (continued)
- Very Limited Budget available
- Small staff and Limited time available from
Finance, Accounting and IS Staff - Wanted to keep History (okay some history)
- Dozens of off the shelf and home grown software
packages with duplicity in functionality - Some non-standard customizations SQL Scripts
and other run from Batch programs launched from
users desktops - Important Later
9Challenges (continued)
- Legacy and other systems all had old chart of
accounts information hard-coded - Customizations not documented Mainly reports
- The infamous Second Production instance
- Microsoft Terminal Servers
- ADI was recently rolled out
- Live by End December 2002 (less than 3 months)
- Parts of the system did not function as desired
10Had to Consider Many Alternatives- The Main Ones
- A Fresh Install and 2 Upgrades of History with
Eventual Merge/Conversion to New Install - B - Upgrade Use an off the shelf changer and
a little less consulting help later merge data
in second production - C - Re-Implement and selectively convert data
Save 10.7 instances
11These Factors were a wash
- Staying on 10.7 unsupported not an option
- Inbound Interfaces would be addressed under all
scenarios (its a wash) - Extensive Integration Testing needed under all
scenarios - Other rewrites such as de-supported AP Interface
and FSG reports under all scenarios
12Considering the Alternatives
13Alternative A Pros
- 2 Upgrades, install new, then Convert
- You get all your historical data in historical
format - You get all your historical data in new COA and
Calendar Format too - Can turn over most of the work to outsiders
except for testing
14Alternative A - Cons
- 2 Upgrades, install new, then Convert
- 3 Instances to maintain
- Reporting across History Instances
- A Lot of work
- We are writing conversion code
- some history not wanted
- Could not make December 2002 live date
15Alternative B -Pros
- 1 Upgrade then Crystallize, then Calendar
Change, then consolidation - Meet the go live date requirements
- Can Consolidate instances after conversion to
alleviate workload - Can rely on outside software/service providers to
do the work
16Alternative B - Cons
- 1 Upgrade then Crystallize, then Calendar
Change, then consolidation - Custom non-Oracle Supported changes occurring
behind the scenes (COA and Calendar Change) - Difficult to determine source of error
(upgrade, Crystallize, or Calendar related) - COT Software license and labor
17Alternative C - Pros
- New Install Selectively Convert, and save
history instances as is - All desired data can be made available in proper
format/calendar - Least labor intensive
- Can still use outside resources to augment
- Old Data still available for inquiry, 1099,
settle disputes, etc.
18Alternative C - Cons
- New Install Selectively Convert, and save
history instances as is - Requires installation and setup of a new system
- The staff remembered how horrible installation of
10.7 had been - Extra migration effort (conversion)
- Not all data in the same place
19What was picked and Why?
- Alternative C
- No Custom Worries (Oracle would be the only
culprit) - Offered fastest, clearest, and simplest and
proven approach - No AP Conversion - Do 1099 Reporting from Old
instance - Cheapest perceived route
- History could die a slow death or
- Migrate to DW
20Now we needed to Execute Quickly! Overview
(Six Weeks)
- Install 11i Vision Nav and New Features Basics
- Rapid Review of 10.7 Setups
- Incorporate challenges/issues documented from
Assessment - Consultants Suggest and Perform Setups in 11i
- Conference Room Pilot Core Processes
- Feedback and adjust
- Integration Testing Instance Most Processes
- Production Cutover
21More Detail -System DBA
- 10.7 7.3.4 RDBMS non server partitioned
- DG Aviion
- Moving to IBM AIX
- Access apps via telnet sessions
- Default passwords
- Minimal usage sometimes very slow
22More Detail GL Functional
- Finance Provide Old to New COA Mapping in Excel
- Determine Amount of History to Convert
- Pick Top Few FSG Reports to Rebuild with new
COA - Produce Budgets in New 4-4-5 Calendar
- Reconciliation and Allocation schemes
23More Detail GL Technical
- Build of the Black Box COA Conversion
- Mapping Table
- Forms
- PL/SQL Package to be called
- Loader Utility (Excel old to New)
- All registered with a custom application
- ADI Install and Configuration
- Modify existing inbound to use Black Box
24(No Transcript)
25More Detail -AP Functional
- Address some of issues and low hanging fruit
improvements - Pay out all 1099 vendors in 2002
- 1099 Reports out of history instances
- Identify/Flag Employees in Supplier Database
- Cutover Planning for New System (no conversion of
open invoices or payments)
26More Detail AP Technical
- Supplier Conversion
- Employee Type Suppliers flagged as such
creation of employee records too - Printer Driver Issues Vendor Labels and Check
Printing Standard Evergreen Short - Check Recon Cash Management Work with Bank
- New Use of Payables Open Interface
27The Final Outcome
- GL History Converted for 1 year only
- All Suppliers converted cleanup project with
sites pending - Cutover on weekend prior to January 1 (New Fiscal
year start) - AP Used system immediately new invoices and
payout old invoices through this year - GL after close of the year on old system
- Old System still available to support close
- History instances will be finally copied to new
system very soon
28Critical Success Factors
- Rapid 10.7 Review and 11i Recommendations
- No Re-Engineering 4 Hour Rule Phase II
- Consultants all experienced Functional with
Technical Skills - Onsite Squeaky Wheel Approach to Issue
Resolution - Customer Flexibility on Conversion of History
GL and AP - Consulting Team that has worked together before
- CFO as Project Sponsor
29Critical Success Factors (cont)
- Black Box for COA Conversion
- Customer addressed their inbound interfaces
modified by no re-writes - Time was the common enemy
- VPN Access for remote support
- Small teams
- A plan of action was developed before execution
began
30What could we have done better?
- More testing time from end users
- Better communication plan
- Assumption from Bank File format BAI2 for check
clearing - Better qualify need for history
- Plan for using offsite developers in advance
- Logistics taken care of VPN, Logins, Workspace,
heat!
31Our Team Structure
- Project Manager/Customer Liason
- Project Strategist and QA
- Prime Functional Specialist
- Supporting Functional Specialist
- 2 Developers (Offsite) Cross Module
- Team Leads and Sponsors from customer side
- DBA with experience on 10.7 and 11i
32Gotchas and Roadblocks
- Supplier Conversion Leading/Trailing Spaces,
Validation better in 11i - Unix Install done on the cheap initially
- No backup plans or capabilities for test systems
- Focusing on the small details rather than big
picture - Centralized Issue tracking system unreliable
- Skepticism of IS staff and the old guard
- Customer Staff already worked 7 days/week 12 hour
days
33Enablers What Helped
- Templates and Pre-Existing Code
- Experience of team and their flexibility
- Small number of participants
- Data Loader
- ADI to point to history and production instances
- The Mission of the CFO and his deadline
- Scope Control in place
34Future Phases
- AP and GL Improvements
- Internet Expenses
- BIS and OFA
- Purchasing and I-Procurement
- Oracle Projects
- A large number of other modules in the Oracle
E-business Suite
35Post Mortem Costs and Timeframe
- Start Early October
- Unit Test done End October
- Integration Test Complete Thanksgiving
- Cutover to Production Dec 30 2002
- Approx cost with Travel expenses 225K
36For More Information Re-implement vs Upgrade
- OAUG www.oaug.org
- List Server
- Paper Database
- O2Works Website www.o2works.com follow link
to KnowledgeWorks - Email Author mclark_at_o2works.com
37THANK YOU
- Mark C. Clark
- Managing Consultant
- mclark_at_o2works.com
- O2Works LLC
- Dallas - Chicago
38QA
39About the Presenter
- 13 Years Consulting Oracle Applications
- North Central OAUG Co-Founder
- South Central OAUG Board of Directors
- North Americas GEO Coordinator for OAUG
- Past Several Years specializing in strategy and
issues related to moving to latest releases of
Oracle Applications - Founded and managed two nationally focused
boutique consulting firms dedicated to Oracle
Applications