The importance of demand forecasting with long leads
William Atkinson -- Purchasing, 4/1/2004 2:00:00 AM
Forecasting demand is important to any purchasing organization, perhaps none more so than those buying from Chinese suppliers with longer than normal leadtimes. Lighting manufacturer Quoizel Inc. of Goose Creek, S.C. is one of those buying organizations, relying heavily on overseas supplier for low-cost parts. With long leadtimes and more than 1,500 home lighting retailers on its distribution list, Quoizel needs to be able to forecast demand accurately to ensure smooth flow of finished product.
If Quoizel manufactured its products domestically, forecasting might not be as important because there would be a shorter leadtime to alter supply to customers. But as Russell Read, purchasing manager at Quoizel explains, the lighting industry moved to a Mexican supply base first in the 1980s and then to China more recently, which now makes 90% of lights sold in U.S. market. In terms of total lighting exports, China shipped $1.3 billion in 1997, $4 billion in 2001, and is expected to ship $9 billion by 2005. Quoizel has been doing business with China for about 10 years, primarily for cost savings.
But the competitive advantages to buying lighting from China can be offset by extended leadtimes. "It takes more than three months to get product from China, which compounds the problems of forecasting," says Read. With leadtimes that long, demand forecasts take on an even higher significance. Adding to the challenge, according to Read, is the lack of manufacturing and communication sophistication in China's lighting industry. Manufacturing and quality systems are often antiquated at Chinese suppliers, says Read, with some still using routing tickets and white boards to track progress.
"The Chinese throw manpower at everything rather than considering automation and technology," he explains. "Chinese suppliers have limited control over their subcontractors and suppliers in terms of quality. Instead of trying to understand sources of quality problems, they just prefer to rework things."
The fluctuating leadtimes throw off capacity plans as well. Most Chinese lighting manufacturers measure capacity by how many containers they can ship in a month but don't understand capacity planning at the component level, Read says. "One of our supplier's leadtimes jumped from 45 days to over 100 days."
About a year ago, Quoizel initiated some research on forecasting, using a relatively sophisticated Access database for sales and inventory. However, it didn't offer much in the way of forecasting capability.
"We developed a forecasting tool on our own, where we did a rolling three-month average, with buyers going in to do adjustments," Read says. After about a year, though, the company realized it wasn't going to get any better with the in-house system. To improve its forecasting, something more sophisticated would be needed that could address the variations in the types of lighting the company sold. Quoizel eventually decided that a tool from Silvon Software was the best suited to solve its forecasting quandary.
Silvon's Stratum enterprise performance management (EPM) system analyzes and manages key metrics in sales, customer purchasing trends, and actual profitability of products and customers. Users can also use the operational planning and analysis application to plan, build, procure, and manage inventory and assess supplier performance in terms of on-time delivery, fill rates, and other performance metrics.
"The Silvon system helps us make better decisions related to the whole supply chain," explains Read.
Quoizel is implementing the Silvon system now and inputting data. There are typically four peak selling points in a year for the company, two of which are tied to a lighting show in Dallas (January and June) and the Silvon tool will also be used to help drive new product introductions. In addition, the company does two "tab" inserts each year (Spring and Fall) to generate sales for its customers' stores and plans on using the tool to help forecast how much product to purchase.
"We need to begin planning for this about four months before the merchants place these tabs in their local newspapers," states Read. Silvon has the capabilities to do modeling at the product level, which helps Quoizel understand patterns and trends and stay ahead of the ordering.
"There's nothing worse than selling a new item to a customer and then making him wait three or four months to get it," emphasizes Read.
The company's current system uses a static load for its supplier leadtime, making it difficult to measure what supplier leadtimes are and incorporate the variance between what they promise to Quoizel and what they actually do in terms of delivery. "The system doesn't allow us to incorporate this into our forecast," Read adds. Once the company gets this metric built into the Silvon system, it will be able to leverage the forecast to compensate for suppliers' actual delivery times.
As no forecast is 100% accurate, Silvon's system also determines the variation between Quoizel's actual bookings and its forecasts to help the company modify and improve its forecasts, which would be impossible with its existing tools. With Silvon's demand forecasting analytics, Quoizel will be able to create and adjust forecasts more quickly and accurately. In the future, the company will also use the system to bring point-of-sale data for its forecasts.
10 Tips for Business Travel in China
11/03/2008Most read on Purchasing.com
09/16/2009Ideas for a New/Old Purchasing Job in China
06/26/2007Economic woes hit China's metals industry
02/11/2009
Now Playing: 