No Longer a Need to Forecast?
May 2010
With the evolution of tighter supply chains, shorter product cycles and shorter lead times, along with dramatic reductions in inventory throughout the supply chain, one might be tempted to conclude that planning and forecasting are a lot less important today than they were in the pre-Internet world. Don’t you believe it!
In days of yore, mass production ruled the hearts and minds of manufacturers and distributors of any product produced and sold in significant quantities. The mass production mindset says that the key to reducing costs is to increase quantity. Whether manufacturing a product or buying parts and materials, a larger quantity spreads fixed costs over a bigger population, and reduces the unit cost of each member of the group. Thus, we demand and expect volume discounts from suppliers and price our products accordingly, emphasizing and perpetuating the economies-of-scale way of doing business.
The relationship is valid, as far as it goes. When looking strictly at production costs, distributing fixed costs over a larger population will, indeed, result in a smaller unit cost. But we can’t look at anything in isolation; only in context do the true costs and considerations reveal themselves.
Let’s say you are a relatively high-volume producer of a consumer product. You expect to buy a half million of a certain component, let’s call it X, from your supplier this year (based on a forecast, no matter how informal). You approach the supplier, tempt them with the large volume, and negotiate a price that is 30 percent less per unit than they would have charged for an order of, say, a couple dozen. They can live with that price because their fixed costs can be divided by 500,000 rather than by 24. You, the customer, are quite pleased with yourself; you got a great price from a reliable supplier.
Reality strikes. Today is the age of customization. Customers are no longer settling for a one-size-fits-all product. The Internet and e-business have shifted the power from the supplier to the consumer. High-volume mass production is a thing of the past. Today’s market demands mass customization-building and selling custom products and variations in any quantity at mass-production prices.
Success in this high-volume, high-mix environment requires flexibility, not only in your own plant but also down through your supply chain. Suppliers have to be willing and able to provide materials and components very quickly, in any quantity. The cost / price structure, and the very real risk of obsolescence, prohibits large-quantity buys and large inventories of parts.
Under these conditions, then, the supply chain is a short-reaction system. As soon as the demand is recognized, information is passed through the entire supply chain, and all links in the chain react immediately to deliver the product very quickly.
In an ideal world, each link in the chain would be able to react immediately to the demand, produce whatever is needed immediately at an affordable cost, and pass it on to the next step in the process. We do not live in an ideal world. Very few manufacturers have successfully achieved an economic lot size of one. And even if that were the case, transportation considerations and the practicality of producing and shipping one-at-a-time would interfere with customer service objectives. In addition, lead times (for both production and transportation) are not zero. The lead times through the product structure are likely to add up to something exceeding the desired customer shipment timeframe. Face it: zero-zero (zero inventory, zero lead-time) is just not gonna happen.
e-Business, especially through collaboration, can help reduce lead times considerably. Production techniques like flow manufacturing reduce in-plant lead times and make any-quantity/ any-time production practical. But meeting customer demands for almost-instant gratification will still require some inventory somewhere in the supply chain, and will certainly require preparedness that can only be based on sound predictions of demand levels, product variety and market swings.
Fortunately, the same collaboration that reduces lead time and increases flexibility also aids immeasurably in developing more accurate forecasts. Plus the quick response nature of the market means forecasts are for shorter horizons, making them inherently more accurate.
The nature of planning and forecasting is changing, but the need still exists. Collaboration and fast, cheap and ubiquitous communications links with trading partners make this a whole new world, one in which all members of the supply chain pull together to satisfy the end customer.
May 11th, 2010 at 2:22 pm
Hi Dave,
Forecasting and inventory deployment in the supply chain does not get much attention from senior management if they are unaware of cascading negative consequences a poor process has on overall business performance. Many in senior management tend to let the forecasting and inventory deployment process be performed by lower levels without any monitoring nevermind intervention.
Many times over the years, I have told clients that if the time to acquire raw material and to manufacture the product is a cumulative leadtime greater than the customer is willing to wait you have to forecast. The most concerning aspect is that one or more people in the organization are forecasting now and it’s likely to be the wrong person(s), without the right tools and forecasting the wrong way. The result can be lost sales and margin along with increased costs buried in overhead accounts and excessive amounts of compensating inventory.
R. Michael Donovan
May 12th, 2010 at 6:22 am
Actually, we are seeing a positive trend in Demand Management.
Not only the technology has gotten substantially better, but senior exec’s do know the value and we sometimes even see it in their 10k’s and other communications to shareholders how demand management techniques contributed to better performance.
New customer relationship/segmentation models; new relationships between trading partners all managed through demand management activities.
In particular, all this rationalization of supply chains due to globalizations and the economic downturn has been done in a ‘smart way’ by using the Demand/Supply Chain systems companies have.
Our research on 2010 priorities shows Demand Planning the number one priority for this year.
We just got done doing a series of research on this and have been publish extensively on Demand in the Second Decade: http://www.chainlinkresearch.com/research/index.cfm?topic=10 and other reports.
You will also note that sales of software is not too bad, considering this market.
I think the real challenge here is the knowledge and skills to run a better forecast process, something too many people know not enough about.
May 13th, 2010 at 9:11 pm
I agree that with the emphasis on pull systems, cycle time reductions, vendor managed inventories, lot size of one, and lean initiatives that forecasts were deemed less important and fell in to disuse or less use by many companies. Many of them later realized that forecasts and planning are needed regardless of the manufacturing environment. The replenishment triggers may have changed but forecasting and materials planning are still needed to balance our resources and provide information to suppliers.
In my view, last few years have seen a renewed use of forecasts and the development of better forecasts. I believe this is due to primarily to the adoption of formal S&OP. Other factors include schedule-sharing with suppliers, the longer-lead-times associated with global sourcing, and cross-functional-teams participating in the process.
The bigger problem as I see it is not the quality of the forecast, it is what we do or don’t do with the forecast. Many companies use the forecast along with firm customer orders to drive MRP and their Master Production Schedule. Others may us the forecast directly to order raw materials or purchased parts.
Most companies, however, do not use current forecast data to determine ABC inventory classes, order quantities, lot sizes, or safety stock levels. These key replenishment rules are most often set or revised based on past usage, if they are revised at all. Those companies that do revise replenishment triggers and order quantities tend to do it based on past usage data rather than future requirements.
I suggest that these companies are missing a great opportunity to manage their inventories based on what they will need, rather than what they did need.
To my mind, being “demand driven” doesn’t mean lot size of one. Rather, it means making better use of our forecasts and taking a forward-looking approach to managing our inventories and working capital. There are tools available that can help.
May 25th, 2010 at 1:25 am
I agree with the above statements and the fact that forecasts are definitely needed. What was not stated was that forecasts are not a static tool. As demand changes for a variety of reasons (customer demand, new products introduction, political, environmental, etc.) forecasts have to be revised. Suppliers have to be made aware and have to be able to react.
Preceding this action is supplier selection. Ones that actually partner with your company and you with them. Ones committed to reducing setup time and running smaller lots, reducing obsolescence at the component level. Agreements between the purchasing body and the supplier have to be negotiated up front as to how far into the future the supplier can build off the given forecast. The timeframe will vary from SKU to SKU depending on volume, risk and potentially seasonality.
All parties concerned realize that meeting the end customer demand can increase market share for the purchasing body and increased profits for all.