Return to the Table of Contents.

7.1 Forecasting.

In a completely flexible production system where supplies, labor, and capacity can adjust immediately to demand and where goods and services are produced only when a customer is ready to buy, a production forecast isn’t needed. Agility, lean manufacturing, and pull systems all tell the IE to work toward that ideal, but as long as the lead time on any input to production is greater than zero, some planning and thus some forecast are needed.

The goal of forecasting is a prediction of aggregate demand (that is, demand for all products in a broad grouping) over a forecast horizon.

Operations management would be much easier if the level of demand for an organization’s products or services never changed. Then the forecast for today would be the “same as yesterday.” If customers have been purchasing the products or services of an organization, what causes the level of demand to change? Why doesn’t the demand stay the same?

• Seasonality. Most consumers buy Valentines in February, Halloween costumes in October, and more of almost everything for gift giving in December. An organization that produces a seasonal product or service will have such regular -- and somewhat predictable -- changes in demand. For example, during its busiest period from Thanksgiving through December, UPS adds leased aircraft to the fleet of planes it owns (Armacost et al, page 16).
• Long term trends. Over a time period of years or even decades, consumers tastes have changed. Consumers tend now to buy larger houses, more prepared foods, and more casual clothing than in the past. Demographics, such as the aging of the baby boomer generation, can affect demand. The demand for school buildings has gone up and down with changes in the number of children of school age.
• Economic cycles. When the whole economy shrinks or grows, demands for all products are affected to a greater or lesser degree. Demand for some products and services tend to be immune to economic cycles (for example, gaming and luxury yachts), some tend to be demanded more when the economy is doing well (for example, air travel and automobiles), and others are counter cyclical, that is, the demand increases when the economy is not doing well (for example, products related to do-it-yourself and home repair and applications to graduate school). Based on historical data on the economy and on demand for its products and services, an organization can use economic forecasts to help forecast demand.
• Product life cycle. Most products have a life cycle involving stages of introduction, growth, maturity, and decline. This web page shows some examples of life cycles for products that are still used (although more current data would probably show a decline in the VCR). Note that this life cycle is for the product as a whole, not for any one particular item. Fashion items and fads have very short life cycles -- do you remember pet rocks? An organization can watch the demand for its product and services and gain some knowledge of where each product or service is in its life cycle.
• Marketing. An organization spends money on marketing in order to increase demand, so forecasts should reflect the predicted effects of marketing campaigns.
• Change in pricing. A planned temporary or permanent change in prices will affect demand and should be reflected in the forecast. Temporary discounts may not increase demand, but only shift it over time, and thus can be used to smooth a seasonal demand pattern.
• Product innovations. As the saying goes, build a better mousetrap and the world will beat a path to your door. When they do, your production system will have to adjust to the new demand.

Reread the above list again, and notice that some of the factors are more under the control of the organization than are others. For example, an organization can decide on a marketing plan or pricing structure, but it can’t move the date of Valentine's Day or Halloween. However, an organization can insulate itself from some of the factors by choosing its mix of products and services. For example, Current Catalog sells paper products related to: New Year's, Valentine's Day, St. Patrick's Day, Easter, Mother's Day, Father's Day, Graduation, the Fourth of July, Thanksgiving, Halloween, Christmas, and Hannakuh, as well as other nonseasonal items (for example for birthdays and anniversaries). They always have some products for the current season, whatever the time of year. An organization can strive to have a mix of products and services where the demand for each goes up or down, but the overall level of demand is fairly level.

Most advice about how to maintain an organization’s demand for its goods and services over the long run focuses on innovation. Even some of the oldest companies do not make the identical products and services they did even five years ago. For example, Coca Cola was founded in 1886 and now sells over 400 brands worldwide. In April 2006 Coca Cola and Godiva Chocolatier announced the launch of a new line of "premium blended indulgent beverages called Godiva Belgian Blends."

“New and improved” is a common phrase in marketing, but even the best companies make mistakes. Coca Cola introduced a new formula for Coke (New Coke) on April 23, 1985, and returned to the old formula on July 11, 1985, after consumers protested very loudly.

An organization’s competitors are also developing new products and their actions, especially their innovations, will affect the demand of all organizations producing similar products and services. Many say: innovate or die. All our previous discussion about the need to have a focus on the customer applies here.

By considering all its product lines, the place of each in the life cycle, and plans for new products and services, an organization can use an aggregated (or bottom up) approach to create a forecast for the overall level of demand for the products and services of the organization. Good forecasts build on quantitative analysis and on judgment.

[Add more on quantitative techniques. Do example with seasonality. Have HW problem with seasonality and trend.]

The above discussion assumes that we make only one forecast, but you will learn in later courses that a more sophisticated approach involves the use of scenarios and probability to make forecasts and plans.