IBM asked SMT and one other, much larger company to bid on 80 more units of a particular computer product. The RFQ (request for quote) asked that the overall bid be broken down to show the hourly rate, the parts and materials component in the price, and any charges. for subcontracted services. SMT quoted $1.62 million and supplied the cost breakdown as requested. The second company submitted only one total figure, $5 million, with no cost breakdown. The deci- sion was made to negotiate with SMT.

The IBM negotiating team included two purchasing managers and two cost engineers. One cost engineer had developed manufac- turing cost estimates for every component, working from engineering drawings and cost-data books that he had built up from previous experience and that contained time factors, both setup and run times, for a large variety of operations. He estimated material costs by working both from data supplied by the IBM corporate purchasing staff and from purchasing journals. He visited SMT facilities to see the tooling available so that he would know what processes were being used. He assumed that there would be perfect conditions and trained operators, and he developed cost estimates for the 158th unit (previous orders were for 25, 15, and 38 units). He added 5% for scrap-and-flow loss; 2% for the use of temporary tools, jigs, and fixtures; 5% for quality control; and 9% for purchasing burden. Then, using an 85% learning curve, he backed up his costs to get an estimate for the first unit. He next checked the data on hours and materials for the 25, 15, and 38 units already made and found that his estimate for the first unit was within 4% of actual cost. His check, however, had indicated a 90% learning-curve effect on hours per unit.

Discussion Questions
1. What are the advantages and disadvantages to IBM and SMT from this approach?
2. How does SMT's proposed learning rate compare with that of other industries?
3. What are the limitations of the learning curve in this case?