This is from a March 1969 article that discusses the Air Force’s implementation of the C/SCPS, which later became known as Earned Value Management:
AFSC [Air Force Systems Command], on the basis of experience in evaluating contractors internal systems, has pinpointed the more important problem areas and situations. The control of undefinitized work, for example, is one such situation. For definitized work a negotiated price provides the parameter, however when work is authorized, but not negotiated, formal contractual changes are not established. In these cases incremental or partial budgets are authorized by the contractor to cover the work, but no baseline is established against which actual costs are compared. Thus when actual costs are incurred with no real indication of value performed, program status is difficult to judge.
“Then there are periodic planning changes,” Teubner explained. “This we call the ‘rubber base line’ problem, which means that retroactive adjustments to schedules and cost estimates are made periodically so that plans equal actuals. The slate is ‘wiped clean’ until another similar ‘recovery’ is necessary. Another problem is caused by borrowing money from future effort to provide ‘get well’ budgets for the immediate and near term problems. This inevitably guarantees overruns and schedule slippages.”
Other areas are: lack of system formality — some contractors were found to have inadequate or non-existent internal cost and schedule control systems; unrealistic cost accounting practices for measuring performance of material planning; ineffective use of available data by managers in identifying variances; and lack of meaningful organizational performance measurement.
That was “USAF Management Specifications Challenges Industry” in Armed Forces Management. I printed it out from microfiche while researching at the Pentagon library during my lunch hour. This kind of stuff is exactly the same things I was dealing with recently on EVM systems and data, except you learn it by talking with others and just looking at enough data to make your own conclusions. When I’d write recommendations on contractor EVMS, I would often include the following — which I thought I “discovered” after seeing hundreds of these reports:
On the problem of establishment of work packages, Teubner noted: “In the engineering area of an R&D contract, work packages tend to be too long and the task descriptions too general. In production, the problem is where to stop subdividing in order to avoid proliferation of unnecessary paperwork.”
EVM continues to force contractors to keep a double set of books — at the government’s expense — and we don’t seem to get much for that expenditure. The 1969 article found that it took nearly a year, under good conditions, for the DOD to audit contractor cost and schedule systems. Still, most (all?) analysts can’t really say how the contract is performing using EVM data when you press them on the issue — the data is just too late, too aggregated, and lacks too much context (though the narratives and variance reports are sometimes helpful, but everyone actually engaged in the project knows those things anyway). Even drilling into the project schedules is hard because they can be gamed pretty easily — you really got to look and trace activities through every schedule release.
EVM makes the most sense on routine hardware efforts, but as intangible qualities of decisions increase, such as in R&D, EVM becomes a rigid encumbrance.
Leave a Reply