Smart: As a technology company, they may have had a better feel for the fragility of their legacy systems, as well as good planning for the new system. They’d helped customers implement complex networking systems in the past, so they had some experience with complex technological projects.
Smart: They knew that extensive customization increases system complexity and development/maintenance costs.
Smart: They knew that they had to do it – this was not an “if” decision, but a “when” decision.
Smart: Handpicked the best and brightest – presented it to them as a challenge, not an assignment.
Smart: PMO empowered to make decisions without getting management approval. This resulted in quick selection of the ERP vendor. “We decided that we should not put Cisco’s future in the hands of a company that was significantly smaller than we were.” Found out what the “big 6” accounting firms and Gartner Group were using /recommending.
Some smart, but partly lucky: Leveraged KPMG, who put their best and brightest on this project in order to build a business installing ERP systems. No “greenies” on the KPMG staff, only very experienced people.
Smart: Steering committee included executives from hardware vendor and Oracle.
Smart/lucky: They negotiated a contract from hardware vendor for the overall system functionality, not for a specific set of hardware. (Lucky that the hardware vendor agreed to this – that vendor lost a lot of $$.)
Lucky: A workaround failed and exposed just how fragile the system was.
Smart: Pete Solvik’s thinking in terms of how big he wanted the company to become, he considered that the current systems would not support Cisco at that size.
Smart: Decided to go against a central business culture – get executive sponsorship to bring Order Entry, Finance, and Manufacturing into the same project – forcing them to all work together rather than independently.
Smart: Established a timeline based on business needs, not just how fast they thought they could get it done.
Smart: Used rapid iterative prototyping via Conference Room Pilot (CRP) implementation phases, moving forward on an implementation/configuration that would give them 80% of the functionality they needed.
Smart: Categorized customization requests to red, yellow, green colors according to “what’s real and what’s not.”
Smart/desparate: “IT did nothing else that year”
Smart: They re-keyed all data, renumbered products, renumbered customers, changed bill-of-materials structure rather than trying to get the new system to work with the old data and data formats.