!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> Streamline Training & Documentation: Not Thoroughly Tested, Yet Ready for Prime Time?

Saturday, September 29, 2007

Not Thoroughly Tested, Yet Ready for Prime Time?

A Wall Street Journal article about Arizona State University's experience with installing new enterprise resource planning (ERP) software brought to mind one of the best books I've ever read. That would be The Soul of a New Machine, which earned its author Tracy Kidder the Pulitzer Prize for General Non-Fiction in 1982.

Until I read Kidder's book, it had never occurred to me that there might be an acceptable reason for selling and installing equipment that had not been thoroughly tested and perfected. Kidder set me right. He explains,
In theory, it would be possible to test fully a computer like Eagle [which became the Data General MV/8000 super minicomputer], but it would take literally forever to do so. The veterans in the Eclipse Group [Eclipse was the computer line on which the MV/8000 was based] maintained that most computers never get completely debugged. Typically, they said, a machine gets built and sent to market and in its first year out in public a number of small, and sometimes large, defects in its design crop up and get repaired. As the years go by, the number of bugs declines, but although no flaws in a computer's design might appear for years, defects would probably remain in it — ones so small and occurring only under such peculiar circumstances that they might never show up before the machine became obsolete or simply stopped functioning because of dust in its chips. (p. 184)
ASU — or at least the IT people there — have obviously accepted this idea of taking unperfected systems live because they embarked on their ERP project with the firm view that project deadlines would be rigidly adhered to regardless of what flaws might turn up in the software along the way. The plan was that any and all problems would be managed while pursuing the overall goal of completing the installation on time and within budget.

As reporter Ben Worthen describes it, the attitude at ASU was:
Admit from the start that there will be mistakes; then work through the glitches with users' help. Most companies take their time and don't start using a new computer system until they are convinced almost everything works right; then they are caught off guard when mistakes inevitably happen. Often, the delays allow them to expand the project's scope, which adds cost and can further compound problems.
In effect, users act as beta testers, and their feedback helps home in on problems in need of fixing.

The upside for ASU has been a huge savings in cost — installation, and support over the next five years, are projected to cost $30 million, well below the $70 million that was initially anticipated. The conversion to the new system took 18 months, whereas similar institutions have needed as much as four years to do the job.

The downside is the pain employees have experienced when their paychecks were delayed or issued in the wrong amount. Morale in HR has suffered as department employees have been forced to adjust rapidly to the new system and to correct the mistakes it has produced.

Still, the situation seems to be improving. Worthen reports that "The most recent payroll had a 4% error rate, which is bellow the 6% error rate the old system traditionally had."

###

Labels: