case study

• Pete Solvik, Ciscos CIO, the last remaining line item of his ERP implementation budget.
• Rewarding performance with cash bonuses
• ERP = Enterprise Resource Planning

HISTORY OF CISCO

• Founded by 2 Stanford computer scientists in 1984, publicly traded in 1990
• Primary product is the router - combo of hardware and software that acts as a traffic cop on the TCP/IP networks that make up the Internet (as well as corporate Intranets)
• 1997 - First year on the Fortune 500, top 5 companies in return on revenues and return on assets
• July 17, 1998 - Market capitalization passed the $100 billion mark
• Don Valentine
  • Partner of Sequoia Capital and vice chairman of the board of Cisco
  • First to invest
  • Reserved right to bring in professional management when he deemed it appropriate
  • Hired John Morgridge as CEO who build a professional management team
    • Founders got pissed and left
    • Free to continue his plans to install an extremely disciplined management structure
    • Maintained a centralized functional organization

HISTORY OF IT AT CISCO

• January 1993
  • Solvik joined Cisco as their companys CIO
  • A $500 mil company running a UNIX-based software package to support its core transaction processing.
    • Functional areas included financial, manufacturing, and order entry systems
• The application didnt provide the degree of redundancy, reliability, and maintainability we needed
  • Outdated couldnt support Ciscos current IT needs
• Werent able to make changes to the application to meet our business needs
  • Application became obsolete, its entire structure became of no use
• Solviks initial inclination to avoid an ERP solution
  • Ciscos strong tradition of standardization
    • All functional areas would be required to use common architecture and databases (the need for being on the same across the company was high/uniformity?)
  • Budgetary decisions on IT expenditures be made by functional areas while IT org reported directly to him
  • Concerns about the types of mega-projects that ERP implementations often become
    • Too big to handle

A DEFINING MOMENT

• Randy Pond, director in manufacturing and eventual co-leader of the project
  • Dilemma facing the functional areas in late 1993
  • Constantly band-aid our existing systems
    • The program itself was unstable, needed constant improvements that made it difficult to use
• Replacing the systems
  • Caused difficulties for functional areas deterioration of Ciscos legacy environment
  • But company sustained an 80% annual growth rate
  • System outages became routine
  • Product shortcomings made coming back from outages more difficult
• January 1994 - Failure of Ciscos Legacy environment
  • Shortcomings cannot be ignored
  • Problem: Unauthorized method for accessing the core application database
    • Trying to overcome the problem stimulated the inability of the system to perform
    • Suppose to have a program that fixed it automatically but it didnt work because software was too out of date(?)
  • Corruption of Ciscos central database Shut down for 2 days
    • System super duper close to total failure
    • Acting independently/freedom to the systems replacement was not going to be enough
  • Ciscos Comeback
    • Carl Redfield - SVP of Manufacturing
      • Got sponsorship from him
      • Start from the manufacturing perspective:
        • Get Order Entry and Financial groups in a single integrated replacement of all the applications; Separate projects took longer time
      • Monolithic IT projects - Unchanging projects become too big
      • Approach:
        • All at once
        • Not going to allow a lot of customization
          • If people want the system to be like their operations instead of retraining people to do things the way the system intended
        • Doable schedule
        • Priority in the company

SELECTING AN ERP PRODUCT

• Needed an integration partner, assist both in selection and implementation of whatever the company chose
• KPMG as integration partner
  • Program manager Mark Lee (previously a director of IT for a company that put in various parts of an ERP system)
  • 20 people
  • Strategy: Build as much knowledge as possible by leveraging the experiences of others
    • Asked for help from other large corps and the Big Six accounting firms
    • Focus selection on what people actually used and continuing to emphasize decision speed
    • Size - an issue in the selection; Put Ciscos fate in a larger company
  • Request for Proposals (RFP) - Sent to vendors, vendors given 2 weeks to respond
  • Selection of Oracle:
    • 1. Product driven strongly by manufacturing; Oracle had better manufacturing capability
    • 2. Promises regarding the long term development of functionality not just one time implementation, provide a growing IT system for a growing company
    • 3. Flexibility of having Oracle close by. Faster solutions to problems.
    • First major implementation of a new release of the Oracle ERP product
      • A totally new ERP system that they dont know if it would even work but if it did it would be very good

GOING TO THE BOARD

• 2 Important questions
  • How much would it cost?
  • How long would it take?
  • Afraid a big project would get out of control and only give okay results
• The process
  • Should take 15 months
  • Want to start at the beginning of Quarter 3 and finish and completely stable for Quarter 4
• Outage in January
  • Biggest customer of current software vendor but the vendor was being bought by another company
    • Left them unclear how their vendor would fit their need/where they stood if the vendor could even support them
    • Reliability, scalability, and modifiability of current applications would not support anticipated future growth
    • 9 months for $15mil
• Cost avoidance is not an appropriate to make the decision
  • Institutionalizing a business model for your organization
• ERP Project was a priority; one of the top 7 goals for that year & every one knew

BUILDING THE IMPLEMENTATION TEAM

• First Act - Extend Cisco's relationship with KPMG through the end of the implementation
  • KPMG's quality service
• Expand core 20 members to 100 - Cross-section of Cisco's business community
  • Those working on the implementation was there on a short-term duration, didn't represent as a career change
  • People chose it because it was a challenge
• 5 "tracks" (process area teams):
  • Cisco info systems leader
  • Cisco business leader
  • Business and IT consultants from either KPMG or Orcle
  • Addtional personnel from the business as team members
  • Managed from a "Project Management Office" - Tom Herbert (Cisco's Business Project Manager) and Mark Lee (KPMG Project manager)
• Executive Steering Committee - Top of the entire project management structure
  • VP of Manufacturing
  • VP of Customer advocacy
  • Corporate Controller
  • Solvik
  • Oracle's senior VP of APplications
  • Partner-in-charge of West Coast Consulting for KPMG
  • Importance placed on the project's success

IMPLEMENTING ORACLE

"Rapid Iterative Prototyping" - broke implementation into a series of phases

"Conference Room Pilots" (CRP) - Purpose to build on previous work to develop a deeper understanding of the software and how it function within the business environment:

CRP0

• Train implementation team and set up techinical environment
• Get team trained on Oracle applications
  • "Immersion" training - "2 x 16 hour classes instead of normal 5 day training classes
  • "Tiger team" - Getting the applications up and running
• Core team met to quickly configure the Oracle package
  • "Locked"/forced
  • Joined by specialists from Oracle and KPMG
  • 3-4 weeks = 80% accurate terms
• Realized - Unable to adhere to one of its early goals: to avoid modification of the ERP software
  • First phase indicated w/o a significant number changes the software would not be able to effectively support the company
  • Changes are substantial

CRP1

• Goal: Each track to make the system work within their specific area
• Team memebers carefully documented the issues they ran across during their modeling and addressed them in weekly 3hr meetings
• Track leaders worked together to resolve issues and push the project forward
• Modeling showed huge numbers of business processes that the software couldn't support
• Modification strategy
• Oracle package would not adequately support the after sales support needs of the company
  • Select a service support package

CRP2 & CRP3

• Before: Systems communicated directly with one another ("point to point")
• New Approach: Communcation would take place via a "data warehouse"
  • Allow all of Cisco's applications to access a single source for their info needs
  • IT did nothing except focusing their energy on the core systems of the company
  • Changed fundamentally all of our underlying data in the company and data warehouse became bridging system between history and future
• Final Goal of CRP2: Begin testing the system, both software and hardware, see how well it would stand up to processing load and transaction volumes required to run Ciscos growing business
• CRP3 Focus, test full system and assessing the companys readiness to go live.
  • Executed these tests by running a full days worth of actual business data and rerunning on an empty day

CUTTING OVER THE ORACLE

Had major day to day challenges that need to be solved quickly to avoid significant impact to the company - Pete Solvik

• Overall business performance plummeted as users attempted to deal with a new system that was pretty unstable
• Primary problem - hardware architecture and sizing
  • Usually would correct by buying more hardware but was messed up due to Ciscos weird contract from the hardware vendor
  • Cisco purchased equipment based on PROMISED CAPABILITY rather than specific configuration
• 2nd Problem - Ability of the software itself to handle the transaction volume required in the Cisco environment
  • Design of the application sucked which put pressure which blew hardware problems out of proportion by inefficiently processing common tasks
  • Mistake - Didnt test the system with a big enough database attached to it
  • During test run, Cisco ran individual processes sequentially as to running it at the same time
  • When processes ran together, the system couldnt handle the amount of work

Strong vendor commitment from Oracle (hardware vendor) and KPMG led to eventual stabilization

AFTER STABILIZATION

• Short lived technical problems

Questions :

1. What is the legacy system problem at CISCO? And how big a problem is?
2. How the specific ERP vendor was selected? Do you think that the way that the team made the decision was right? Why? Why not?
3. The team used a conference room pilot approach. Was this approach appropriate? Why? Why not?

4.What happened at cutover to ERP? Were you surprised by the performance deterioration?