Forecasting

Business organizations compete more effectively in a global economy when they are able to use information from the past to predict future events. One of the tools is forecasting. A forecast is a statement or prediction about a future event and the act of making this prediction is called forecasting. Forecasts are used in government organizations to predict inflation rates, unemployment rates, welfare payments, gross national product, interest rates, money supplies, and various planning indicators. These forecasts are called economic forecasts. Forecasts that are concerned with rates of technological progress, like introducing new products and services that require new plants, equipment, and technology, are called technological forecasts. Projections for a company's products or services, finances, human resources, and supply chain management are called demand forecasts or microeconomic forecasts.

Forecasts can also be classified with respect to the length of future time horizon covered by the forecast. If a forecast has a time span up to one year but is generally less than three months, it is called a short-range [Page 693]forecast. For example, job scheduling, job assignments, planning purchases, demand planning, sales planning, and production levels planning are examples of short-range forecasts. They are usually implemented by the operations managers. If a forecast has a time span from three months to three years, it is called a medium-range forecast (or intermediate forecast). These types of forecasts are used by middle managers and are implemented in sales planning, budget planning, production, and resource planning. If a forecast has a time span of three years or more, it is called a long-range forecast. These types of forecasts are used by the top level of management or strategic managers when making decisions about locating a new facility, expanding an existing facility, investing in a new venture, introducing a new product or new service, and planning capital expenditures.

The forecasts generated within the firm must be understood and appreciated by its decision makers. It is important to understand the forecasting process and forecasting methods to make proper decisions. In general, there are two types of forecasting methods: quantitative and qualitative. If time series data are available, then quantitative methods will be used. A time series consists of data that are collected, recorded, or observed over chronological moments of time, for example, daily sales of a commodity, weekly sales of Dell computers, quarterly returns of a specific investment, and annual consumer price indices. The time series prediction is based on the assumption that the future is a function of the past. A forecaster is looking at what happened over a period of time in the past and uses past data to predict future.

Time series are analyzed by distinguishing certain patterns: trends, seasonality, cycles, and random or irregular component. Trend represents the gradual increase or decline in time series over a period of time. Seasonality is a pattern of change in the data that repeats itself after a period of weeks, months, quarters, or years. These patterns are easy to observe in sales time series that show customers' demand similarities in each season of the year or in electrical or gas consumption, which is determined by weather and temperature changes. Cycles are the fluctuations around the trend pattern and usually they occur every several years, for example, inflation, weak economy, and real estate prices. Random or irregular patterns are caused by chance or very unusual events, like hurricanes, floods, accidents in the workplace, rapid declines in the stock market, and other hard to predict issues.

...