A common technique for estimating labor hours is to use historical experience from similar work and from prior production runs. Whenever history is used however, it must be adjusted to account for such things as differing conditions, quantities, and improvement curves. When the government evaluates labor hour estimates, it nearly always asserts that contractors show improvement to historical data and patterns. Contractors would do well to also consider improvements to productivity in preparing estimates.
Productivity improvements result from any number of factors. They may be due to the adoption of improved methods and tools or the increased efficiency of the individual worker performing repetitive tasks (e.g. learning curve theory). The amount of improvement per unit of product is generally high during the early part of the production cycle and decreases as production is stabilized, processes are refined and additional experience is gained.
After awhile, the rate of improvement may not be measurable except over a substantial period of time. Where automation is used, the rate of improvement is usually zero. As production nears its end, productivity often decreases. Tooling wears out or the most skilled workers are transferred to other projects. Sometimes, production runs are less than optimal and fixed costs are spread to fewer units.
The government's primary interest in labor productivity is in measuring current productivity and past trends, and determining the causes of past trends so that the likelihood of continuance during the contemplated production period may be assessed. It is important that contractors (or potential contractors) ensure that the government understands all relevant factors in assessing productivity. The government's propensity is to conjure up reasons why contractors should or could be more productive. Sometimes these positions do not consider all relevant facts. Contractors without all the facts, might find themselves agreeing to something during negotiations that could imperil contract performance.