Forecasting - Qualitative Methods
Forecasting - Qualitative Methods
Time-series and causal-model approaches to forecasting both require an analysis of past data. In contrast, qualitative techniques are based on the subjective judgments of individuals. Of course, the individuals may have access to quantitative information about the past, but the judgments themselves are subjective summaries or distillations of all the knowledge that is relevant to the forecast.
Jury of Executive Opinion
This approach involves combining the judgments of a group of managers about the forecast. The group normally would include a variety of concerned and informed managers representing such functioned areas as marketing, sales, operations, manufacturing, purchasing, accounting, and finance. They each would bring to the forecast a different background, perspective, and set of biases. Often their judgments would be supported by background information that might include past data, economic and industry developments, competitive actions, and relevant news from customers or dis
tributors. For example, the fact that a large retailer has decided to stock a competitor's product or is planning a large promotion might be relevant "news."
This technique has several advantages. First, it tends to be fast and efficient. Second, it tends to be timely in that the forecast is generated on the basis of the most current situation. Third, the knowledge on which the forecast is based is extremely rich. It includes at least potentially all the collective knowledge and experience of the involved managers. Fourth, it has no formal requirements in terms of historical data.
The disadvantages are due to the subjectivity involved, because the information on which the individual forecasts are based could differ from period to period. In one period the marketing manager might have just returned from a trade show or from a sales meeting and have better information than usual. Even the composition of the group might differ. Because the content and background can differ from period to period, it is difficult to make adjustments when forecasts err, even when they err systematically.
Actually, this technique is one of the most widely used. The Conference Board conducted a survey of the use of five major forecasting methods of commercial firms. The results from 161 reporting companies are shown in Table 22-1. The jury of executive opinion is the technique most used in consumer products and service companies and the second-most-used technique in industrial products companies.
Salesforce Estimates
The salesforce estimate involves obtaining the judgments from the sales-force. Usually, each salesperson will provide a forecast for each product or product type by customer. Since many of the decisions based on the forecast—such as production quantities—are product-specific, the forecast usually is needed by product.
Uncertainty can be introduced in several ways. One way is to allow the salespeople to provide a sales estimate range in addition to the single number representing the expected sales level. Another is to have them estimate the probability of each potential sale. The sales forecast then will be the sum of all the potential sales, each weighted by its probability of occurring.
Although the technique is termed a salesforce estimate, it also can be employed in organizations that do not have salesforces. The principle is that those in the organization who are closest to the organization's customers or clients make the individual forecasts, which are then combined into an aggregate forecast. A welfare agency might use caseworkers to forecast their own caseloads, for example. Or branch librarians each might be asked to forecast demand for various categories of books for their branches. The term salesperson, in this context, thus will mean any organizational position that has direct contact with the customers or clients whose behavior is to be forecasted.
Forecasts generated from the salesforce are based on the salesperson's knowledge of the customer or client, which is often extremely complete, sensitive, and current. Like the jury of executive opinion approach, it allows the introduction of subjective judgments drawing on rich experience. The approach can provide detailed information about product lines and customers, which can be used productively in diagnostically identifying problems in the marketing program. For example, if sales are forecasted to be below normal for a certain product line or geographic area, then an investigation might be prompted. Such an investigation based on a forecast could be stimulated many valuable months before historical data provided the same signal.
This approach suffers from subjectivity, as does the jury of executive opinion approach. Furthermore, salespeople often lack relevant information about a company's plans and overall industry trends. In addition, there is a greater tendency for biases to occur. Individual salespeople can be naturally optimistic or pessimistic, and such biases are often more serious when the forecast is linked to their performance measures. For example, compensation often is based in part on quotas, and forecasts sometimes contribute to quota development. In such cases salespeople might have an incentive to forecast low. Salespeople sometimes are criticized for "lost sales," and they may thus fail to report prospects that are not sure things, in order to reduce their lost sales. To counter such tendencies, a salesperson could be given credit only for sales that were listed in the forecast, but, of course, that policy would encourage overly optimistic forecasts.
There are two ways to deal with biases. One is to identify and compensate for them. If one salesperson is, on average, 20 percent low over several reporting periods, the forecast could be increased by 20 percent. Another salesperson could be 30 percent high, and that forecast would be reduced by 30 percent. Another approach is to divorce the forecast from any performance appraisal and ask that it be made as objectively as possible. The problem then might be to motivate the salesperson to be conscientious about the forecast. If the value of the forecast, in terms of the planning of operations and production, is well communicated, the motivation can be generated, but it is nearly always difficult to maintain.
Industrial firms often have relatively few customers and rely heavily on their salespeople to reach those customers. Thus, the salesforce is the logical group on which to rely for forecasts. In contrast, in consumer and service firms advertising plays a much more important role in communicating with the customer and there are usually many more customers involved.
Surveys of Customer Intentions
The third judgment method is a survey of customer intentions. Customers are asked to make their own forecast about their usage and buying intentions. Customer intentions will be based on subjective judgments about future requirements.
All the considerations of developing a valid survey are involved. The product or demand to be forecasted, the forecast period, and the population all must be specified clearly, and the sampling plan needs to be developed. When production or service sales are being forecast, the sampling frame is usually the existing customer or client list, as the bulk of sales or demand usually comes from existing customers. The right person in the customer organization must be contacted, and questions must be addressed with care. As the last chapter indicates, various biases can emerge.
Obviously, the advantage of the approach is that the user-customer has the best information on which to base a forecast. It works best when the customers, or at least the major customers, are few in number. With few people or organizations to contact, the survey sometimes can be conducted quickly by telephone. As Table 22-1 indicates, the greatest use of this technique is in industrial firms, which often are characterized as having a relatively small set of major customers.
Surveys can become expensive and time consuming, of course, and are thus inappropriate for some forecasting tasks. They also are difficult when a sensitive purchase decision is involved and those surveyed may be reluctant to provide information.
Delphi Approach
In the jury of executive opinion a group of people gather and reach a consensus on a forecast. Such an effort is subject to group effects. In particular, judgments might be swayed by the persuasions of some group members who have strong personalities, special interests (a sales manager might be optimistic, to encourage production), or special authority. When long-term forecasts are involved, such effects are particularly likely to occur. On the other hand, the information of the group should lead to a better forecast than that of any one of the individuals.
A method to retain the wisdom of a group while reducing the effect of group pressure is the Delphi approach. In the Delphi approach, group members are asked to make individual judgments about a forecast. These judgments then are compiled and returned to the group members, so that they can compare their own previous judgments with those of the others. They then are given an opportunity to revise their judgments, especially if they differ from the others. They also usually can state why they believe that their judgment is accurate, even if it differs from that of the other group members. After three or four iterations, group members usually reach their conclusions. There still remains group pressure, of course, but it is less overt since the individual is rating anonymously. The technique has been used with thousands of respondents, but more often it involves only from 10 to 40 respondents.
Time-series and causal-model approaches to forecasting both require an analysis of past data. In contrast, qualitative techniques are based on the subjective judgments of individuals. Of course, the individuals may have access to quantitative information about the past, but the judgments themselves are subjective summaries or distillations of all the knowledge that is relevant to the forecast.
Jury of Executive Opinion
This approach involves combining the judgments of a group of managers about the forecast. The group normally would include a variety of concerned and informed managers representing such functioned areas as marketing, sales, operations, manufacturing, purchasing, accounting, and finance. They each would bring to the forecast a different background, perspective, and set of biases. Often their judgments would be supported by background information that might include past data, economic and industry developments, competitive actions, and relevant news from customers or dis
tributors. For example, the fact that a large retailer has decided to stock a competitor's product or is planning a large promotion might be relevant "news."
This technique has several advantages. First, it tends to be fast and efficient. Second, it tends to be timely in that the forecast is generated on the basis of the most current situation. Third, the knowledge on which the forecast is based is extremely rich. It includes at least potentially all the collective knowledge and experience of the involved managers. Fourth, it has no formal requirements in terms of historical data.
The disadvantages are due to the subjectivity involved, because the information on which the individual forecasts are based could differ from period to period. In one period the marketing manager might have just returned from a trade show or from a sales meeting and have better information than usual. Even the composition of the group might differ. Because the content and background can differ from period to period, it is difficult to make adjustments when forecasts err, even when they err systematically.
Actually, this technique is one of the most widely used. The Conference Board conducted a survey of the use of five major forecasting methods of commercial firms. The results from 161 reporting companies are shown in Table 22-1. The jury of executive opinion is the technique most used in consumer products and service companies and the second-most-used technique in industrial products companies.
Salesforce Estimates
The salesforce estimate involves obtaining the judgments from the sales-force. Usually, each salesperson will provide a forecast for each product or product type by customer. Since many of the decisions based on the forecast—such as production quantities—are product-specific, the forecast usually is needed by product.
Uncertainty can be introduced in several ways. One way is to allow the salespeople to provide a sales estimate range in addition to the single number representing the expected sales level. Another is to have them estimate the probability of each potential sale. The sales forecast then will be the sum of all the potential sales, each weighted by its probability of occurring.
Although the technique is termed a salesforce estimate, it also can be employed in organizations that do not have salesforces. The principle is that those in the organization who are closest to the organization's customers or clients make the individual forecasts, which are then combined into an aggregate forecast. A welfare agency might use caseworkers to forecast their own caseloads, for example. Or branch librarians each might be asked to forecast demand for various categories of books for their branches. The term salesperson, in this context, thus will mean any organizational position that has direct contact with the customers or clients whose behavior is to be forecasted.
Forecasts generated from the salesforce are based on the salesperson's knowledge of the customer or client, which is often extremely complete, sensitive, and current. Like the jury of executive opinion approach, it allows the introduction of subjective judgments drawing on rich experience. The approach can provide detailed information about product lines and customers, which can be used productively in diagnostically identifying problems in the marketing program. For example, if sales are forecasted to be below normal for a certain product line or geographic area, then an investigation might be prompted. Such an investigation based on a forecast could be stimulated many valuable months before historical data provided the same signal.
This approach suffers from subjectivity, as does the jury of executive opinion approach. Furthermore, salespeople often lack relevant information about a company's plans and overall industry trends. In addition, there is a greater tendency for biases to occur. Individual salespeople can be naturally optimistic or pessimistic, and such biases are often more serious when the forecast is linked to their performance measures. For example, compensation often is based in part on quotas, and forecasts sometimes contribute to quota development. In such cases salespeople might have an incentive to forecast low. Salespeople sometimes are criticized for "lost sales," and they may thus fail to report prospects that are not sure things, in order to reduce their lost sales. To counter such tendencies, a salesperson could be given credit only for sales that were listed in the forecast, but, of course, that policy would encourage overly optimistic forecasts.
There are two ways to deal with biases. One is to identify and compensate for them. If one salesperson is, on average, 20 percent low over several reporting periods, the forecast could be increased by 20 percent. Another salesperson could be 30 percent high, and that forecast would be reduced by 30 percent. Another approach is to divorce the forecast from any performance appraisal and ask that it be made as objectively as possible. The problem then might be to motivate the salesperson to be conscientious about the forecast. If the value of the forecast, in terms of the planning of operations and production, is well communicated, the motivation can be generated, but it is nearly always difficult to maintain.
Industrial firms often have relatively few customers and rely heavily on their salespeople to reach those customers. Thus, the salesforce is the logical group on which to rely for forecasts. In contrast, in consumer and service firms advertising plays a much more important role in communicating with the customer and there are usually many more customers involved.
Surveys of Customer Intentions
The third judgment method is a survey of customer intentions. Customers are asked to make their own forecast about their usage and buying intentions. Customer intentions will be based on subjective judgments about future requirements.
All the considerations of developing a valid survey are involved. The product or demand to be forecasted, the forecast period, and the population all must be specified clearly, and the sampling plan needs to be developed. When production or service sales are being forecast, the sampling frame is usually the existing customer or client list, as the bulk of sales or demand usually comes from existing customers. The right person in the customer organization must be contacted, and questions must be addressed with care. As the last chapter indicates, various biases can emerge.
Obviously, the advantage of the approach is that the user-customer has the best information on which to base a forecast. It works best when the customers, or at least the major customers, are few in number. With few people or organizations to contact, the survey sometimes can be conducted quickly by telephone. As Table 22-1 indicates, the greatest use of this technique is in industrial firms, which often are characterized as having a relatively small set of major customers.
Surveys can become expensive and time consuming, of course, and are thus inappropriate for some forecasting tasks. They also are difficult when a sensitive purchase decision is involved and those surveyed may be reluctant to provide information.
Delphi Approach
In the jury of executive opinion a group of people gather and reach a consensus on a forecast. Such an effort is subject to group effects. In particular, judgments might be swayed by the persuasions of some group members who have strong personalities, special interests (a sales manager might be optimistic, to encourage production), or special authority. When long-term forecasts are involved, such effects are particularly likely to occur. On the other hand, the information of the group should lead to a better forecast than that of any one of the individuals.
A method to retain the wisdom of a group while reducing the effect of group pressure is the Delphi approach. In the Delphi approach, group members are asked to make individual judgments about a forecast. These judgments then are compiled and returned to the group members, so that they can compare their own previous judgments with those of the others. They then are given an opportunity to revise their judgments, especially if they differ from the others. They also usually can state why they believe that their judgment is accurate, even if it differs from that of the other group members. After three or four iterations, group members usually reach their conclusions. There still remains group pressure, of course, but it is less overt since the individual is rating anonymously. The technique has been used with thousands of respondents, but more often it involves only from 10 to 40 respondents.
Comments
Post a Comment