Background: A client in the business of remanufacturing very large electrical transformers contract with us to improve productivity among his management staff. Typical Lead Times for remanufacture is close to one year, with much of that being procurement of bushings from Germany, and other special parts. Takt Time for these transformers – the rate at which a transformer is loaded onto a rail car and shipped – is about one week.
Situation: The rather lean management staff, charged with the engineering design, the sourcing of a myriad number of components, and the oversight of the labor, was asked to switch their attention to providing a budgetary estimate and perhaps firm price quote for a new opportunity about once per week. The time and effort required to accurately estimate the labor man-hours, which could run as high as 10,000, kept the staff from attending to the production work they knew was important. Delays, poor quality and deteriorating morale on the shop floor was a result, as was occasionally forgotten scope elements in a quotation.
Analysis: The detailed process of estimating man-hour cost of the work and the data the team used to do that was observed over the course of several weeks. Attempts were made to correlate the data the team used, including readily available nameplate data, to actual man-hours from 29 prior jobs. Regression analysis showed a high correlation (.82 to .97) between "core and coil weight," a value found on the unit nameplate, and total man-hours. Surprisingly, the results provided by the regression equations more accurately predicted the actual man-hours than the efforts of the management team did.
Improvement: Regression equations for "small" and "large" units within the two major transformer designs were introduced as a pilot to compare against more labor intensive efforts, with the Plant Manager having the final voice in the man-hours quoted.
Result:The team slowly began to trust the regression equations more, as more data was gained. Though the long method has not yet been replaced for firm price quotes, it is done more quickly now as team members worry less about omitting part of the scope, and use the regression equation to temper their own estimate. The plant manager carries a sheet with the regression equations and uses them to challenge the team's estimates as well as to provide instantaneous budgetary estimates.