Forecasting Customer Demand: Lucky or Lousy Predictions Set Inventory Investment Levels

By Bruce E. Jacobs

Customers never buy what the medical device manufacturer forecasts. They never buy the forecasted volumes, the forecasted products or within the forecasted time periods. So why forecast if the forecast is nothing better than a “lucky or lousy” prediction of future demand? If you were able to consistently develop an accurate forecast of customer demand, wouldn’t you use that ability to predict the outcome of more profitable endeavors such as horse races, the lottery, professional athletic events, Friday-night bingo or any potentially profitable event?

The future is never what it appeared to be in the forecast, and customers never buy what the forecast predicted. More important, in a competitive business environment, the recent past does not depict the near future in terms of a competitor’s actions, a customer’s actions, demand volatility and world events, for example—not to mention the effect of the economic climate, prime rate, stock market performance, housing market, healthcare costs, oil prices and consumer price index. Today’s business environment changes at a faster pace than in the past and will continue to change at an even faster pace in the future, adding higher levels of uncertainty to the predictability of customer demand and the development of an accurate forecast.

But, even with the inaccuracy, volatility of demand and the ever-changing dynamics of the business climate, customer demand still must be forecasted. Forecasting is essential when it comes to planning manufacturing capacity, having the raw materials for production, identifying the resources required to manufacture the product and having the appropriate level of finished goods inventory available to fill customer orders when they are received. Manufacturing companies, however, get into trouble when they rely totally on a forecast to establish inventory targets, determine production requirements and lot sizes as well as purchase parts and raw materials. This is indicative of low inventory turns and increasing investment in inventory. Even though sales volumes are increasing, inventory levels continue to rise. These factors are built into the forecasting process and practices.

To make matters worse, there may be multiple forecasts being developed and used. The sales organization develops a forecast of product volumes and revenue. The materials organization develops its own forecasts of product volumes because it knows better and because sales’ forecasts are always wrong. Customers may provide a product forecast of the quantities they will need, but their forecasts can’t be relied on because the volumes change weekly. Couple multiple forecasts with lucky or lousy predictions, and it shouldn’t surprise executives when inventory investments rise and product shortages increase, forcing the factory to break into production schedules to produce the products customers are ordering that are out of stock. This results in higher manufacturing costs incurred from production schedule break-ins and enormous expediting of materials and product to meet orders. The end result is a total landed cost structure that is too high and a supply chain that is too slow to respond because it is being bull-whipped into a just-in-time mode of operation by default, not by design.

To break this degrading cycle, most companies will attempt to raise inventory levels using the “peanut-butter approach”—ie, they add a little more inventory for every product and focus on developing a more accurate forecast by either purchasing a forecasting model or putting more resources around the forecasting process. The solution seems simple—just develop an accurate forecast. If it were that simple, though, the investments companies have made in state-of-the-art forecasting models would have alleviated the problem a long time ago.

Forecasting and Lead Times

The primary purpose of forecasting is to have product available when the customer wants it, which means finished goods inventory and manufacturing product to an inventory target are based on forecasted customer demand. In its simplest form, finished goods inventory is required when the lead time to manufacture and deliver product is greater than the lead time the customer allows for the product to be ordered and delivered; therefore, a forecast is required. But if the lead time can be reduced, significantly less inventory is required because the product supply chain’s capabilities and responsiveness are improved. Fewer customer order shortages occur, and back orders and delayed shipments waiting for a product are reduced. The greater the lead-time reduction, the greater the flexibility and responsiveness of the product supply chain.

Most manufacturing lead times can be reduced from 60% to 90%, which increases throughput capabilities. In every manufacturing operation, the total days of inventory requirements are directly proportionate to the total days of manufacturing lead time. Eliminating a day of inventory in the product supply chain by reducing lead time can provide order-of-magnitude economic gains. As lead time is reduced further, eliminating more days of inventory is possible, which begins to improve the total cost structure and availability of product.

When lead times improve, the supply chain becomes more capable of producing product closer to actual customer demand, which reduces dependency on forecasts—more importantly, accurate forecasts. Moving to a pull-based demand replenishment model begins to come into play as daily visibility of actual customer orders establishes the production schedule for manufacturing operations, which signals replenishment to a lower finished goods inventory target. As the manufacturing enterprise develops the skills and capabilities to replenish products at the rate of actual customer demand, inventory levels can be lowered. The forecast is adjusted for actual demand, which realigns the forecast, and a decision is made to either manufacture more product or fill orders from inventory and lower the inventory target.

With the ability of the product supply chain to produce product closer to the customer’s actual rate of demand, the product supply chain begins to deload, operating with a shorter lead time and lower inventory level. Replenishment to inventory targets and the requirement to pre-build inventory because of known peaks in actual demand or capacity limitations are determined based on regular monitoring of the forecast and actual customer demand. With a shorter lead time, the product supply chain can react to demand much quicker, and the pre-build of inventory can be delayed longer with tighter demand management and reviews of the forecast and actual customer demand.

Evaluating Progress

Weekly demand management and communication (DM&C) meetings are held to review actual customer demand and the forecast. Any known changes in actual demand and the forecast also are identified and communicated.  The meetings result in forecast, customer demand and inventory target agreements among DM&C meeting participants who represent sales and marketing, manufacturing operations and materials management. Representatives from these business functions are focused on achieving the highest level of customer service for the product, and on manufacturing the lowest total level of inventory. The sales and marketing function is responsible for creating product demand, forecasting volumes and communicating with customers about their demand for products. The manufacturing and materials functions are responsible for demand fulfillment of products when it is needed. The sales and manufacturing functions are jointly responsible for establishing and agreeing to the inventory targets and meeting the customer service requirements. All customer service issues related to product availability are reviewed at the weekly DM&C meeting, where forecast accuracy, customer demand and inventory performance are measured and reviewed.

When pre-build inventory targets are required because of capacity limitations, volatility of demand or seasonal peaks in demand, the pre-build inventory target is established, and forecasted and actual demand are monitored closely to make adjustment as required. Having an accurate forecast may never be achievable, but improving product availability and customer service with lower inventory levels can be achieved with improved lead times and DM&C management processes, thus lowering the dependency on lucky or lousy forecasts.

Bruce E. Jacobs is a principal with BKD, LLP, a CPA and advisory firm, and a member of BKD Manufacturing & Distribution Group, providing solutions for the management and financial needs of medical device businesses. Bruce has more than 26 years of experience in manufacturing and distribution strategies. Contact the author at [email protected].