Why Does GKN Run Lean?

GKN Driveline North America’s (GKN DNA) primary customers are automotive original equipment manufacturers (OEM). An automotive OEM requires its suppliers to ship the right materials, in the right quantity, to the right place at the right time and at a competitive price. That may seem like a basic request, but as supply management staff at the GKN DNA’s Roxboro and Alamance facilities can attest, it takes “lean” operations to be successful in their business.

The driving force behind GKN DNA’s decision to move toward lean was the result of the customers’ trend toward increasingly lean operations. The OEMs want to better meet the demand of the end consumer by reducing their own inventory levels and running efficient operations. As such, they have very stringent delivery requirements. In short, the margin of error regarding on-time delivery in the industry is incredibly small and the associated penalties are quite large. This can be illustrated with some figures from GKN DNA:

Small Margin for Error (1)

• 20-25 outbound trucks per day
• 4 doors for loading outbound production parts

• 45-60 minute time window (specifically scheduled times) for loading trucks

Large Penalties (1)

• A missed delivery generally results in premium outbound freight costs. For example, for a mid-west customer destination, if the load is shipped within approximately 8 hours of the original scheduled time, then it can usually be shipped expedited ground, which costs approximately $1,000-$2,500 per occurrence.

• However, using the same example, if the load is shipped beyond approximately 8 hours of the original scheduled time, then it will require an expedited air shipment, which costs approximately $5,000 to $25,000 per occurrence.

Given the high cost of expediting a shipment, why would GKN DNA do it? According to a presentation by a GKN DNA representative at the SCRC semi-annual meeting in the spring of 2003, one number says it all (1):

• In 2002, the typical OEM assembly line shut-down fee was $41, 978 per minute.

• Not to mention the impact on lost future sales for the offending supplier (which is difficult to quantify).

Obviously, GKN DNA must be responsive enough to maintain a high level of customer satisfaction regarding on-time delivery. But while on-time delivery is one thing, staying competitive on price and quality is another. GKN DNA must ensure that everything it buys and adds value to will be sold to the OEM. In addition, they desire to reduce the amount of working capital held up in inventory. Being “lean” helps here, as well.

GKN DNA typically receives from three days to two weeks worth of “firm” release data from the majority of its OEM customers. This “firm” data, known as an “862” in electronic data interchange (EDI) speak, can actually change as late as 3 hours prior to dispatch.

GKN DNA also receives “830’s,” which can be used for short and medium range planning. For instance, GKN DNA typically gets “non-firm” forecasted release data from the “Big Three” (GM, Ford, and DaimlerChrysler) for about six-months out, and from Toyota and Honda for about two-months out. But there can be substantial variances that far in advance. 830s can vary drastically from actual shipments, both by volume and by mix.

So as a safeguard, GKN DNA compares the OEM’s forecast with one created by their own devices. Currently, the division uses the economic and financial information firm Global Insight, Inc. Global Insight provides GKN DNA with an independent forecast of domestic automotive production on a high level. But data in that form is not currently very useful to production, so a forecaster at GKN DNA headquarters in Auburn Hills, MI must quantify how much of that forecast is likely to be the division’s business.

For this, the forecaster uses a combination of quantitative and qualitative methods which take into account data such as GKN DNA’s historical ship rates and customer-specific model/platform knowledge to arrive at a final GKN DNA volume forecast. Then the data is divided up further into the product mix forecast. The data resulting from this analysis represents GKN DNA’s best estimate, by part number, of future sales to its customers.

That data then goes to the GKN DNA Master Production Scheduler residing at the tier one plant in Roxboro, NC. The scheduler uses the resulting forecast as the source data for the master production schedule (MPS), which is the basis for GKN DNA’s plan for production, inventory, staffing, etc. Once per month, the MPS sets the quantity of each end item to be completed in each week of a rolling 90 day planning horizon. Within the first 60 days of the planning horizon, OEM volume release data varies by an average of around four percent. But there is typically much more variance in the product mix.

Beyond the 60 day horizon, a major driver of forecast variance is caused by imperfections in the source data (from Global Insight). For example, the forecast is based partly on jobs per hour (productivity), rather than total volume. The forecast therefore does not take into account “unplanned” plant shutdowns that frequently occur. Although the forecast does a good job of predicting total sales in the automotive industry, it is not designed to predict the product mix at the driveshaft application level. It also doesn’t handle demand “peaks” or “valleys” very well. In short, it is not intended to look at demand in the detail that GKN requires.

Relying too heavily on inaccurate forecast data could put the GKN DNA in a precarious position. Owning obsolete raw materials – or worse yet, obsolete finished goods – would be a very costly mistake. But since it was unrealistic that the average end consumer was going to tell the OEM what they were going to buy, the forecast data would likely remain imperfect. Thus, GKN DNA decided to move toward “lean” operations. Of course, that doesn’t happen overnight.

Reference:

(1) Greene, G. (July, 2003). Discussion with author.