Cisco Systems Inc. : Implementing ERP come ride with us your cowboys: bill atkinson denisa kubricka edmond lui georg wittenburg iman sharif Company Background • Founded in 1984 by two Stanford computer scientists • Became publicly traded in 1990 • Primary product is “router” • By 1997, Cisco was ranked top five companies in return on revenues and ROA in Fortune 500 • In 1998, market capitalization was over $100 billion Markets Cisco was a key infrastructure supplier for the “New Economy” in the mid-90s. • That market went through a period of amazing growth since Cisco formed. • This fast growth rate was directly reflected in Cisco’s sales figures. • The future was looking bright. Company Structure • Three functional divisions: – Order Entry – Finance – Manufacturing • Initial IT Strategy: – Let division take care of themselves. – Overall architecture is shared, enabling sharing of data. History of IT at Cisco UNIX-based software package to support its core transaction processing: – Functional areas supported: financial, manufacturing and order entry systems – Used common architecture and common databases • Growth of Cisco resulted in scalability problems. • Cisco was the largest single costumer of that vendor, resulting in a strategic weakness. Point Blank • Would the software developed for a $300 million company fit the use of a $1 billion company? Point Blank • Why would a multi-million dollar company want to avoid ERP?
A Big Need • Recognized the need for change, but left actions to each functional division: • Thus: – Little progress was made in the year – Each functional area was reluctant to replace the legacy system because of high risk involved – Systems outages became routine – Unauthorized method for accessing the core application database malfunctioned, corrupting Cisco’s central database • Company was shut down for two days Selecting an ERP product • The planning was driven only by timing constraints and panic. There was no business case • Cisco emphasized the need for: – – – – Strong team Strong partners Speedy decision making Getting Executive & Board approval Project Team & Partners • Team: – Know that very best people are needed – Pulled best business & IT people out of their current jobs at Cisco • Partners: – Important that partner could work on the selection as well as implementation of project – KPMG as integration partner – KPMG team of 20 (highly experienced; not “greenies”)
Teams selection strategy • Teams strategy – use experiences of other companies and best practices to accumulate knowledge • Selected five packages within 2 days • After a week of high level evaluation – two packages selected: ORACLE and another major player in the ERP market • 10 days on request for proposals Point Blank • Is it wise to make a decision so quickly? Are there things that should be done to mitigate the risk? Did they do due diligence? Team Selection Strategy Cont’d Oracle & other vendor given two weeks to respond to RFP • Current vendor customers were visited by the team during these two weeks • After response, received a 3-day software demonstration by each vendor (used Cisco’s sample data) • Goal is to show how software meets or does not meet Cisco’s requirements Final Vendor Selection Criteria • Three main criteria used: – Manufacturing capability – Long-term development of functionality of package – Flexibility of Oracle’s being close by (location wise) Other motivations – Oracle’s first release of new ERP product – if Cisco project goes well, favorable product launch of Oracle ERP package • Oracle chosen – team decision, no management approval at this point Time • After 75 days from start of project, major TODOs are: – Negotiations between Oracle & Cisco – Write up a Proposal to Board of Directors • Time and non-interference with annual accounting as main considerations. • Famous last words: – “… there’s no way we’re going to take 15 months to get this done. That’s ridiculous. ” – “Well, can we do it in five months?
That just didn’t seem right. ” – Let’s try nine. Point Blank • Was nine months realistic? Should other criteria have been used to estimate the time, rather than quarters as primarily criteria? Costs • No formal business case for project • Concentrating on system failure as motivation to project start • $15 million budget estimated • Not approached from the justification prospective (no cost/benefit analysis) • Costs: » » » » Software 16% Hardware 32% Headcount 14% System integration 38% Point Blank • How do you think should project costs be estimated for Cisco’s project?
Getting Approval From Board • Met with CEO – comment about ‘jobs lost over much lesser amounts of money’ • Got CEO’s support • Met with Board of Directors – chairman says ‘show me the money’ as first thing • Board approves project • Single largest project ever undertaken by company • CEO makes project priority for Cisco Building implementation team • As not enough time & KPMG performed well during planning phase • KPMG relationship extended for implementation • Extra 80 team members added on from the Cisco’s business community • Five tracks (process area teams) used: Order Entry Track » Manufacturing Track » Finance Track » Sales/Reporting Track » Technology Track Point Blank • Was it worth removing important people from the regular business positions to work on the IT project? How can an IT department in another company convince upper management that this is worthwhile? Steering Committee • High level execs from Cisco, Oracle, and KPMG • Shows commitment and importance of project Point Blank • How important is it to have support of upper management to ensure success?
Implementing Oracle • A development technique known as “rapid iterative prototyping” • Implementation broken into a series of phases called “Conference Room Pilots” (CRPs): – CRP 0 / 1: Build on previous work to develop a deeper understanding of the software and how it functioned – CRP 2 / 3: Implement the ERP system. CRP0 • Training the implementation team and setting up the technical environment • Two parallel efforts: – Training the team in the Oracle applications Normal 5 day training pushed to two 16-hour days! Getting the application up and running by a small “tiger team” CPR0 • • • • • Configuring the Oracle package Hundreds of parameters in the applications Team members “locked” together for two days 1 % effort with 80 percent accuracy Completed one week after the meeting, leading to the realization that changes to the software were needed to support the company effectively Point Blank • Do you think that having 80% accuracy with only 1% effort was just a matter of luck?
Taking into account that typical ERP system configuration takes up to 6 months, not 2 days!! Any risks involved with this approach? CPR1 • Goal of this phase each track makes the system work within its specific area • Details and procedures for completing a process were created • Realized that a lot of business processes were not supported by the software needed modifications • Modifications were classified into: • Green • Yellow • Red, needed to go to the steering committee for approval there were few reds CRP1 30 developers needed for 3 months to modify Oracle • Modifications led to unplanned changes in the project plan and budget • Realized that Oracle could not support the after-sales support needs • Chose a service support package and planned to lunch both packages together Point Blank • Would it have been possible to avoid most of these major changes and the need for a new service package had they spent more time in their initial design and decision making? • Does it help to prioritize the required changes and control scope? CRP2 Continued scope change • Major technical issues • Creation of data warehouse for centralized data communication CRP2 • 100-person IT department started decommitting from other projects • Bore most of the responsibility for the project additions • “IT did nothing else that year” Point Blank • Is it wise to commit all of your resources to a new project, when your existing systems are barely scraping by? CRP3 • Focus on testing the full system • Assess readiness to ‘go live’ • Captured one day’s worth of actual business data and ‘re-running’ it on a Saturday
Point Blank • Would you consider one day of testing adequate (with a subset of data), if you were planning a clean cutover of your entire IT infrastructure? The Aftermath • The new ERP system went live on January 30, 1995, but it took two months before it was operating at a reasonable level of quality. • Problematic areas were hardware architecture and sizing. – Test hat only been run sequentially and with a subset of the real database. • Side-note: How is it possible that no one in the team noticed this before? Cisco Systems (1995) Cisco Systems creates five distinct business units that reflect its major networking product groups — Workgroup, ATM High End, Access, Core and IBM Internetworking. • “While leveraging economies of scale in areas like manufacturing, sales and support, the business units can move quickly in product development and expedite time to market. ” • Cisco Systems is the first major supplier of internetworking products to be awarded global ISO 9001 certification. Cisco Systems (2004) • John Chambers, president and CEO: “Our strong position in the core switching and routing business continues to be complemented by positive momentum in our Advanced Technologies, especially this quarter in storage, security, wireless and IP telephony. ” • Positive Q2 2004 figures: – Q2 Net Sales: $5. 4 Billion (14. 5% increase year over year; 5. 8% increase quarter over quarter) – Q2 Operating Cash Flows: $1. 7 Billion Oracle Corp. (2004) • Third quarter revenues were up 9% to $2. 5 billion while net income grew 11% to $635 million as compared to the third quarter last year. Chairman and CFO Jeff Henley: – “Oracle’s fiscal third quarter was another solid quarter, with new software license revenue growth of 12%, which is identical to last quarter. ” • Oracle CEO Larry Ellison: – “This was a very strong quarter for our database business. ” groupthink Point Blank • Would they be able to do it again? – Which were to key factors to the success of the project? – At which points could it have failed? – Are these one-time events or can we generalize them? Pete Solvik CIO of Cisco Systems • Heads the Internet Business Solutions Group (IBSG) • One of the top 25 unsung heroes of the Net by [email protected] Week Pete’s Tips for CIOs • Make sure that the company’s “business owners” make and fund IT spending decissions. • Use infrastructure as a strategic enabler. • Tie IT’s objectives and rewards to the goals of the company business units. • “Pete Solvik’s Three Tips for CIOs” http://www. voicendata. com/content/top_stories/101010311. asp Thank you for your time!