Forecasting future demand

Because no one can predict the future perfectly, forecasting future demand is necesнsarily inexact. Many economists and statisticians specialize in forecasting market potential. Consequently, as is true for measuring current demand, the best sources for forecasts of market potential are trade groups and services that specialize in this type of information.

Several techniques are available to create sales forecasts. The techniques based on judgment or opinion are clasнsified as "qualitative." The techniques based on statistical analysis of historical data are classified as "quantitative."

The marketer can use a combination of forecasting techniques to gain a broader perspective on the possible range of demand. In assessing the forecast that results from each technique, the marketer takes into account expected changes in the enviнronment, such as changes in the size of the market to be served or in the economy. These are reported in the business press and in trade magazines. Thus, if The Wall Street Journal reports that a recession is apparently ending, the marketer might favor a rosier forecast than if the pace of business is expected to slow.

QUALITATIVE FORECASTING TECHNIQUES The simplest way to arrive at a forecast is to ask "experts" what they predict demand will be. One approach is to seek the outlook of the organization's executives. Such a jury of executive opinion provides insights from people working in a variety of areas, including finance, marketing, and production. The marketing manнager can average the estimates to arrive at a single forecast.

The estimates also can come from the organization's salespeople. They are, after all, the organization members who work most closely with customers. In this case, the forecast would be the sum of the estimates. This approach assumes that salespeople will give unbiased estimates. However, if they think their sales forecasts will be the basis for sales quotas they must meet to earn a bonus, salespeople might be tempted to give pessimistic forecasts.

The marketer also can ask people outside the organization. One possibility is to ask the customers themselves, by conducting a survey of the buying intentions of a sample of the target market. This approach assumes that actual buying patterns will match the stated plans of the survey sample. However, people do not always do exactly what they say they will, so this information can be misleading. Industry experts also forecast industry growth, and marketers can get this information from trade groups and business publications.

Some organizations use a more complex approach to forecasting known as the Delphi technique. In this approach, the marketing department sends a survey to experts inside or outside the organization, asking them to provide a forecast. The results are averaged and sent to the experts along with another questionnaire asking them to review the results and provide another forecast. This process is repeated until the experts reach a consensus.

With the exception of the Delphi technique, these qualitative forecasting techнniques are relatively simple. Especially when experienced people provide the numнbers, they can also be quite accurate. However, inexperience or poor judgment can lead to woefully inaccurate forecasts.

QUANTITATIVE FORECASTING TECHNIQUES To forecast demand less subjectively, marketers use quantitative techniques such as time-series analysis and market tests. Time-series analysis is the use of past data to predict future outcomes. It assumes that demand follows a pattern over time. Thus, it is reliable only if past trends continue into the future.

To use the form of time-series analysis known as trend analysis, the analyst looks for the pattern in the data, and then uses it to project future demand. For example, if sales have risen an average of 5 percent per year over the past several years, the analyst would predict that they will continue to increase at the same rate. Of course, most patterns are not so obvious. Therefore, computer programs that perform computations like regression are readily available to conduct such forms of trend analysis.

A form of time-series analysis that tries to overcome some limitations of trend analysis is exponential smoothing. Exponential smoothing is a form of time-series analysis that gives more weight to more recent data and less to older data by assignнing a weight to each year's data. The sales figure for each year is multiplied by the weight assigned. Thus, to analyze sales data for the past three years, the analyst might multiply the current year's sales by 0.8, sales from a year ago by 0.6, and sales from two years ago by 0.4. Then the analyst completes the forecasting by using trend analysis.

Trend analysis and exponential smoothing both require records of past sales. Therefore, they are not useful for forecasting demand for a new product. Marketers of a new product who want to use a quantitative forecasting technique might use a market test. This involves offering the product in a few test markets, assuming the response will be similar when the product is offered to the total target market. Market tests are expensive but have the advantage of measuring actual customer behavior.

Note that the quantitative techniques are not a substitute for judgment by the marнketer. In fact, they require that the marketer make judgments about many of the figures. For example, to use exponential smoothing, the marketer must decide what weights to assign to each year's data. To use a market test, the marketer must select test markets that he or she believes are representative of the total market. Thus, all of these forecastнing techniques are only as good as the judgment of the marketer using them.

EVALUATING THE FORECAST Before using the numbers generated by a forecast of demand, the marketer should take some time to evaluate the forecast. Because marketing plans are influenced by such numbers, the marketer can avoid costly errors by making sure that the forecast is reasonable. The marketer should review the assumptions and judgments used in preparing the forecast. Upon review, do they still seem reasonable? This is especially important given modern use of computer technology to prepare forecasts. Managers should not be so awed by a computer model or detailed spreadsheet that they fail to judge the assumptions underlying them.

The marketer should also look at the results of the forecast. Do the numbers seem realistic? If not, the marketer should review the estimates, judgments, and computaнtions used to arrive at the results. Perhaps the data came from an unreliable source. When a result seems odd, chances are a mistake was made somewhere. It's less costly and embarrassing to prepare a revised forecast than to have to rework the entire marнketing plan after trying to implement it.

Finally, the marketer should bear in mind that the accuracy of forecasts cannot be guaranteed. Too many uncontrollable factors may make the future impossible to preнdict. For example, sales of cold remedies are at the mercy of the length and severity of the cold and flu season. Good news for consumers is bad news for the makers of these products. In one recent year, the cold and flu season got off to a slow start, causнing sales of over-the-counter cough and cold medicines to fall 9 percent compared to the previous year. Forecasts for these medicines rely only on more predictable trends, such as the growing strength of private-label products ("store brands"). To cope with the remaining uncertainty, marketers should prepare contingency plans and monitor the environment and the reaction of target markets to the organization's efforts. When marketers observe a development that calls for a change, they must modify their plans. Thus, the planning process continues after the last word of the marketing plan has been typed and evaluatedЧeven after the marketing department has begun implementing that plan.