Manufacturer of Super Alloys Uses SmartForecasts to Set Accurate Stocking Levels and Reduce Inventories by 20%

“Classical inventory control approaches have not worked well here,” according to Charles Sponaugle, vice president of business planning for Haynes International. “Our products are very specialized.”

Haynes International is a major manufacturer of super alloys made of nickel and cobalt (and not to be confused with the underwear people). Its products are used in high temperature applications such as flying gas turbines used to power jet aircraft as well as in corrosive chemical applications. While similar in strength to carbon steel at room temperature, Haynes’ alloys are as much as 1000 times stronger at 1,800 degrees Fahrenheit.

Haynes operates in a specialized international market where there are only a few competitors. The products it makes are relatively expensive—100 times more expensive than carbon steel. A number of products they make are hard to find; for some, Haynes is the only supplier in the world. To service this market, the company maintains four warehouses in the U.S. and three in Europe. While the company stocks about 2000 items, many are related to each other and are often sold in concert to satisfy a single customer order.

The Challenge
According to Sponaugle, many of the products Haynes sells do not follow normal demand patterns. While the old inventory management system was good at forecasting fast-moving items, Haynes had difficulty accurately forecasting a significant portion of its inventory that demonstrated intermittent, slow-moving demand. The old system was very cumbersome to use, and was only able to give planners limited data to which they could just react. It didn’t really help them plan. Rather, this situation resulted in cyclical swings in Haynes’ inventory that Sponaugle describes as “like a see-saw.”

Haynes makes all the products it sells, and lead times can fluctuate significantly. The typical manufacturing lead time for any given item can be from 8 to 16 weeks. However, for standard products (i.e., not custom mill work), Haynes’ customers are often able to accommodate a delivery time of 2 to 3 weeks, allowing Haynes to supply the item from incoming stock. This delivery flexibility enables Haynes to run lean, setting minimum stocking levels without jeopardizing business. In fact, this flexibility enables Haynes to operate at a fill rate of only fifty percent, accepting back orders on many of its standard items without hurting customer satisfaction.

However, large customers account for a significant percentage of Haynes’ business, but their orders are not consistent or predictable. One customer may have 200 items identified as needed over the course of a year. However, each order usually has a completely different and random set of items. Over the course of a year, the demand history, at the SKU level, for this customer is extremely intermittent.

The Solution
Y2K was a catalyst for Haynes to overhaul its manufacturing information systems. Along with a new IBM AS/400 computer system, the company invested in new software including SmartForecasts Enterprise with its patented intermittent demand forecasting technology. This capability of SmartForecasts was an important factor in Haynes’ purchase decision.

To come up with its forecast, Haynes combines a statistical forecast with information from its customers. Haynes generates the statistical forecast with SmartForecasts using demand history downloaded from the new AS/400 system. The company then adjusts the statistical forecast employing information on customer usage obtained from the sales force. The consensus forecast is then used by the company’s finished goods system where current inventory is compared to the forecast, and this process triggers a replenishment order, if needed.

The Results
“The biggest advantage of using SmartForecasts,” according to Sponaugle, “is that it gives us an accurate forecast! Other forecasting packages can give us a projection, but with Smart’s intermittent forecasting module we get usable results. The forecast shows us whether our stocking levels are set right to meet our inventory targets. As a result, our planning process is running smoother.”

In little more than a year, Haynes has reduced its inventory by over 20 percent while maintaining customer service levels and improving the allocation of manufacturing assets.

“The optimum inventory level for us is the minimum amount necessary to conduct the business appropriate for each service center location,” continues Sponaugle, “SmartForecasts with its intermittent demand forecasting capability provides a realistic estimate of how much material, in aggregate, we need to efficiently satisfy customer demand.”

“SmartForecasts helps us understand our demand and its relationship to the future. It picks up subtleties that can easily be missed or overlooked,” says Sponaugle. “For example, we now have a clear idea about how lumpy our demand can be and how many pounds of an alloy and form we need to achieve a specific service level.”

Knowing the company’s requirements and how to meet them with the help of SmartForecasts enables Sponaugle and Haynes to keep on target—meeting sales goals while minimizing inventory costs.

Back to Success Stories