Richard Morse, a MIT graduate started National Research Corporation (NRC) as a process development company focused on developing vacuum technology. NRC initially used the technology to produce frozen orange juice concentrates and instant coffee. In 1950 NRC applied this technology for producing of high purity metals and started with tantalum (Ta) production in 1959. NRC went through a series of ownership, with HC Starck AG (Germany) acquiring fifty percent stake and a venture capitalist acquiring the rest in 1976. In 1986 Bayer AG (Germany) and Bayer Corporation US purchased majority of HC Starck. In 1989 HC Starck (HCST) acquired Ta mill and wire production facilities from Fransteel, thus adding to its existing Ta reduction and powder manufacturing facility. This largely increased its share of tantalum metallurgic products market.
II. Situational Analysis:
HC Starck International was manufacturing Ta, an expensive metal used in various industries like aerospace, nuclear reactor, surgical equipments. The largest buyers were the manufacturers of electrolytic capacitors and vacuum furnace parts. The total worldwide consumption of Ta was about 550 tons annually.
HCST used tantalum double salt, Ta Scrap and Ta ingot as a raw material for manufacturing Ta powder, wire, and mill and fabricated products. It sourced the raw materials from mined ore, scrap generated from open market, scrap from its consumers, scrap generated from its own production and Ingot from government reserves.
The raw materials are melted into an ingot and cold forged into a 4inch thick sheet bar, which was the starting material for the metallurgic product division. These sheet bars were rolled into sheets of varying thickness. The entire procedure resulted in production of scarp ranging between 20-30%. Figure 2 and illustrates the generic product hierarchy.
Figure 1: Generic product hierarchy
Though no substitutes were available in the market to replace Ta, delays in supply of the Ta products had often created a fear that customers might switch to Zirconium as a substitute and if it worked well, consumer would never switch back to Ta which was priced four times more than Zirconium. HC Stack Inc also recently faced challenges for other suppliers, in particular the Chinese companies selling Ta wire. HC Starck also experienced difficulties in sourcing Ta scrap, the availability were uncertain and this forced them to maintain huge inventory. Also the market demand for finished goods was very erratic.
Scheduling of Production:
Though new ERP (SAP’s R/3) system was installed at Starck Inc, ordering for raw material, planning and scheduling for production was done manually. Jim Mc Mohan, Supervisor of production control collected sales order and converted them in to production forecast, often for a whole year on month bases and revised it a few times. I addition to raw material ordering Jim also did manually performed shop floor scheduling. SAP system were not used as there were problems with the product recipe pricing (cost) and efforts to fix the problem were taking longer than expected, making the system unreliable.
Orders from consumer were very erratic, to account for this uncertainty HS Starck acquired “blanket orders” and scheduled production of these orders randomly spread, throughout the year. For example if it received a total order for 5000 feet of tubing it produced a standard 20 feet tubing in batches and later sectioned them in to varying sizes based on requirement. But very often sectioning based on order sizes resulted in odd sized pieces generating a huge amount of scrap.
Sales and Marketing:
Sales and marketing team at H C Starck, Inc had consistently achieved sales target, return on assets. The team uploaded its sales figures and forecast through the SAP system, however the production planning continued the use of manual methods, this resulted in ‘information black hole” as the data forecasted was not utilized at the melt shop for planning, physical piece of paper still travelled from sales to production, resulting in longer lead times.
As a way around to this the production planning, sales and operations team instituted a “drumbeat” meet every day to review the shipment that was due in a week time. The meeting managed to keep everyone up dated and expedite delayed orders, but this resulted in most of the jobs being by passes unless they made it to the ‘drumbeat list’.
Lead time and Inventory level:
Inventory was very expensive and H C Starck maintained huge stocks, this was largely due to scarcity of Ta scrap. They also purchased excess to keep it out of the hands of competitors even though the need was not immediate. At any time they had more than adequate to ensure high level of availability.
The average lead time was estimated at 7 weeks. Average manufacturing time was 2 weeks. Analysis showed that longer more than longer time, up to 2 week, was utilized from the point the sales orders were released to the point when stock rooms issued material for processing. This process was still largely manual and SAP R/3 system was not used.
The typical production routing used a made-to-order policy with either 4 inch bar of quater inch sheets as the starting material. Depending on the final product the material passed through multiple process orders before ending up as final product (Figure
Figure 2: Production cycle: Multiple process orders
To expedite on lead time sales team often entered hard orders on probable sales, hoping that the operations team would begin fabrication even before the order is placed and the order would be read by the time customer placed their orders. Sometimes these worked well but often the SAP system required orders have delivery dates and so the sales team made up dates.
III. Problems areas highlighted as in the Case*:
The Metallurgical Products department at HCST was scheduled as a make-to-order job-shop, with customer lead-time performance averaging seven weeks. Can we go in for made to stock kind of arrangement?
Order expediting is the rule rather than the exception, and in fact a daily meeting occurs to enable the expediting.
The plant carries an average of six months inventory, yet few items are sold from stock, or even made in a single production step from stock.
Nearly all work passes through some of the standard gauges of 4″, ¼”, 1/8″, and 0.030″, yet no standard stock is held at these sizes other than a small amount at ¼”, and small left-over pieces at the other gauges.
The Sales group was pressing hard to reduce customer lead-times to fewer than three weeks. The goal seemed attainable since production orders averaged just over two weeks, but something needed to be done to speed the time between when an order was received, and operations began working on producing the final product.
It seemed as though maintaining stocks of some of the standard intermediate sizes would help customer lead times, since end-items could be produced in a single production operation, but which items should be stocked, and at what levels?
Also, not everyone in the organization was convinced that reduction of customer lead-time was a priority – some were more focused on inventory reduction, while others felt that inventory levels were not that important.
Based on the problem statement given in the case, we now work on the strategic alternatives and implementation.
IV. Strategic Alternatives and Implementation
What to stock and how much?
Our product demand analysis shows that for all products being made out of Alloy-A, The top 8 (a class inventory)(20%)products contribute 80% of the sales volume the respective standard deviations and the relative standard deviations percentages are shown as below
*Ref:-Designing and Managing the Supply Chain-Concepts, Strategies and Case Studies (3rd Edition). Page 119-120, Case Wrap Up.
1999 Invoiced Sales – Pounds per month
Disk – 10″
Since the standard deviation for these products is fairly high, a higher level of inventory level is advisable for these series of products, also most of the products are in the form of sheets, which can be modified or value added to arrive at an alternative product, these can form a part of the raw material inventory being kept in the system. Further on Alloy-A, the balance products from 20% (B class inventory) of the customer requirement and are characterised by low standard deviations, hence they can be manufactured as and when an orders come in.
We see a similar pattern for Alloy-B
Here the requirement is much more focussed with the almost 80% of the customer requirement being focussed around top 6 of the products (11%). A similar strategy can be used as above. These products can be made to stock; most of it is in the form of welded tubes hence it would be advisable to place semi-finished stock of plates of alloy-B of the desired thickness so that complete products can be made on short notice. A similar strategy for B class inventory can be adopted for Alloy-B as well
Raw Material Inventory Planning and Wastage Reduction
The corresponding Raw material inventory can be arrived at by calculating the inefficiencies in the system, now it is given that there is wastage of 20% in following stages of manufacturing
There is wastage of 20% in the first stage and wastage of 10% overall in the second stage, hence the total wastage is 28% since the first process serves as a feeder for the second process. We believe that the wastages can be avoided by introducing a minor process change during the tension rolling process, that accounts for 20% of the wastage and it is given that the same is due to the two ends of the Ta sheet touching the cylindrical springs that are essentially there to hold the sheet/bar in place while it is undergoing rolling operation. We suggest that the profile of the Tungsten head be made so that it matches the profile of the “rolling drums” thus allowing additional area of the Ta sheet to be exposed to the rolling operation thus reducing wastages. Depending on the precision with which the Tungsten profile is made the wastage of 20% can at least be bought down to 5%, a saving of 15%.
However, for our current estimates we assume that the wastage is 28% and we plan our raw material inventory accordingly. Currently there is a practice of storing 6 months of raw material which is very high. Various reasons have been given in the case to justify the same, however at the end of the day; inventory does block the company’s working capital. In the model suggested below, based on past sales data and the current lead time of 2.3 weeks for manufacturing which in turn is calculated from the time the raw material is received and goods are packaged ready for dispatch, we suggest that the following inventory be maintained for the two alloys.
Inventory of Raw Material for Alloy A
Inventory of Raw Material for Alloy B
Drum beat meetings and implementation of SAP R/3
The company should continue with the practice of drum beat (DB) meetings as they are a good means to reduce distortions in communications, identifying urgencies and expediting resolution of pressing issues. However, the current purpose of DB meetings is to communicate production and sales urgencies which should not be the case. SAP R/3 should be immediately implemented for the production planning stage and the process of issuing multiple production orders should be done away with by homologation of input parts.
D. Multiple Product Orders:
The problem of issuance of multiple production orders will also be minimized once we focus on key SKUs as suggested in the inventory rationalization model. One more purpose of the DB meetings could be to iron out continuous issues with the implementation of SAP for PPC (Production, planning and control). However the discussion on planning should be kept out of preview consciously by those attending the meeting.
E. Cost Savings from raw material inventory Reduction:
H C Starck maintains large quantities of raw material inventory. The total cost saving from reducing inventory levels and making the production a just in time (JIT) process would reduce the cost as shown below. The company can begin by hiving off the raw material procurement division, and develop the same as a centre of excellence, supplying quality raw material.
Avg. piece weight
$ 25.00/ hrs
Change over time
Labour cost ($/pound)
$ 0.67 (1)
Inclusive of 10% on Overtime
Raw material and finished goods price:
Average price (in $)
»¿Tantalite ore (contained pentoxide basis)
$ 35 – 45
$ 135 – 240
$ 180 – 250
$ 100 – 150
Average Material Cost ($) (2)
Total Cost of raw material and work in process Inventory
Break Down and Finished
6 months Inventory
Cost of Material (2)
Total Cost (3)
Average cost of work in progress inventory (WIP)
Raw material cost component for WIP
Labour cost (1*4)
Total (WIP) (5)
Total savings (3-5)_
Hence the total saving form reduction in raw material inventory would be $32.89 million.
F. Cost saving from Safety Stock
Assuming that the fixed cost per service order (SO) is $2000 and a service level of 97%, cost of holding the safety stock for 7 and 3 weeks lead time for the top 10% of product line for alloy A are shown below: