A Computerworld article covers some of the problems (and ends with a few happier notes) about a PeopleSoft (Oracle) ERP implementation taking place in ND’s government and education sectors. Although the state agencies sound generally satisfied, the article focuses on North Dakota University System’s unhapiness with the unexpected massive cost and time overruns for getting their system implemented.
Why did they underestimate the costs, which ballooned from the extra time required for the (still) incomplete implementation? The article suggests the lesson to be learned is never embark on a major project like this without employing a full-time project manager (which, surprisingly it sounds like this implementation lacked from the start). But there is something else to learn from the article:
“The academic software modules, particularly a grants and contracts management application, also did not perform as expected and have required extensive customization, said Laura Glatt, vice chancellor of administrative affairs at the Bismarck-based university system.”
I wonder why they did not expect this? Perhaps their original RFI/RFP was not designed to request that information? Did they script some demonstration scenarios for the vendor to show them how the modules would accomplish the sort of functionality they needed? I’d think there could have been some way to prevent this issue–maybe the ghost of a full-time project manager would have thought of that during the selection and evaluation phases.