After a million dollars of investment, the ASTEC system was born.
The California State Auditors discovered, after a year of operation of the new system, that the new operational cost was $3.60 per document, as opposed to the previous manual system’s cost of 50 cents. This meant over seven times more for automation, plus the initial investment.
There were no other redeeming attributes to justify this cost difference, so the new system was destroyed and replaced by the old one.
My source of this intriguing reverse-automation tale, Montgomery Phister, later provided me with a post-mortem consultant’s report on the ASTEC project. This showed that by killing the automated system the State saved about one million dollars per year. The operational cost went back to the original 50 cents per document. Detail were originally published in Phister’s Data Processing Technology and Economics, Digital Press, Maynard, Mass.
The ASTEC system is a glaring example of how the specification of a solution (in this case, and ‘automated system’) became a goal in itself. The proof that it was regarded as a goal was that nobody reacted to the failure during the first year of operation. The misunderstood goal had been ‘successfully’ achieved. The real goals, among which were the control of public expenditure, were never clearly stated, never designed into the system, and never monitored for real success by the people who implemented the system.
What about your project? Would it stand up to the test of meeting the real objectives of your user?
4.9 An exercise on your current project documentation
it is an instructive exercise to find a piece of project documentation, and check the following questions : Are the functional and attribute requirements clearly separated from each other? Are suggested solutions kept entirely separate from the requirements, or are they simply mixed in with the real management goals in such a way that people might get confused?