Introduction

As the competition in markets for products and services continue to become more intense, it is imperative for organizations to improve their attempts to plan for the future (Smith & McIntyre, 1994). One of the tools that organizations use to plan for their futures is the sales forecast. Sales forecasts have become a key component of most firms’ marketing planning process since the 1960s (Morrison, 1996). Recent developments in software and other technologies have improved the accuracy of many of the sales forecasting approaches available to companies today (Chase, 1996).

This non-empirical paper will look at past research as it applies to the roles of salespeople in the development of sales forecasting models that are used by organizations. A review of the role of the salesperson in the marketing mix will be followed by a clarification of the definition of sales forecasting.  A review of sales forecasting models will examine the wide variety of approaches being used by organizations today to predict their sales levels.

The role of salespeople in the forecasting process will be considered. Next, recommendations will be made concerning the use of salespeople in the forecasting process. Finally, limitations of the paper and recommendations for further research will be made.

The Salesperson in the Organization

Salespeople are the front line of marketing in their organizations. They are the individuals who meet with customers on a daily basis. Salespeople often are the face of the company to purchasing agents and consumers (Marks, 1996).

So what is the role of the salesperson, if any, in the sales forecasting process? The answers vary depending upon a number of factors and perspectives. Before attempting to answer that question, the next section will define sales forecasting. After defining forecasting, an examination of different approaches to the sales forecasting procedure will be made.

Definition of Sales Forecasting

The first challenge when exploring the field of sales forecasting is to agree on a definition. Bolton and Chase (1997) report that firms they studied used as many as 31 terms, such as quota, projection, and estimate to describe what they defined as a forecast.

“A sales forecast is an estimate of sales (in dollars or units) that an individual firm expects to achieve during a specific forthcoming time period, in a stated market, and under a proposed marketing plan” (Stanton & Spiro, 1999, p. 392). The key word in that definition, of course, is estimate.

It is an estimate that much of the rest of the organization uses to make decisions. Manufacturing, accounting, shipping and many other functional areas in the organization make decisions based on the sales forecast (Gordon, 1997).

Moon and Mentzer (1998) contend that good sales forecasts are important to providing good customer service. “When demand can be predicted accurately, it can be met in a timely manner, keeping both channel partners and final customers satisfied. Accurate forecasts help a company avoid lost sales or stock-out situations, and prevent customers from going to competitors” (p. 44).

Accurate forecasts can also improve a company’s profits by enabling the firm to more accurately plan its purchases. Transportation costs may be reduced if a firm can more precisely predict what products need to be shipped and when they need to be shipped (Moon & Mentzer, 1998).

Qualitative Forecasting Techniques

Until the 1960s, senior level executives did most of the sales forecasting in organizations using the executive judgement approach. These executives or managers use their past industry experience and knowledge to determine what they thought the company’s sales could or should be (Mentzer & Bienstock, 1998).

This executive judgement method was seen as appropriate and is believed to have worked well in the stable post-war economy of the United States. As the economy has become more competitive and volatile, the executive judgement method has been replaced in some organizations (Mentzer & Bienstock, 1998).

However, as many as 86 percent of firms in a recent study (Herbig, Milewichz, & Golden, 1993) report still using the executive judgement method for forecasting sales. A study of the wholesale industry (Peterson & Minjoon, 1999) reveals that almost all (99.6 percent) of the firms in this study report using managerial judgement as the primary sales forecasting tool.

This continued popularity of the executive judgement model is generally attributed to the fact that it is a simple and inexpensive technique for forecasting sales (Stanton & Spiro, 1999). The executive judgements made can be based on any combination of objective and subjective data that is available to those making the forecasting decisions (Gordon, 1997).

Some firms today utilize other types of qualitative forecasting techniques. Group discussions among members of a forecasting committee bring together divergent viewpoints from different parts of a firm. Pooled individual estimates from different functional areas of an organization are averaged to determine what they think sales figures will be (Paley, 1994).

The Delphi technique has become an increasingly popular qualitative method. Forecasts are submitted in writing to a forecasting team leader who feeds these estimates back to those who submitted them with information about changes in the market place. Eventually a consensus is reached based on the revisions being sent back and forth between the team leader and team members (Paley, 1994).

Another increasingly popular qualitative technique is the composite of sales force opinion. Here the individual sales representatives and/or their sales managers are asked to predict their sales for the coming year. These individual predictions are aggregated to develop the company’s sales forecast (Mentzer & Bienstock, 1998).

The major advantage of using qualitative forecasting techniques is that they have the ability to predict changes that occur in an established sales pattern. Those involved in the day to day operations of the company can anticipate and plan for these changes. This is something quantitative approaches are unable to accomplish (Hogarth & Makridakis, 1981).

Two major problems exist when using qualitative techniques to forecast sales. First is that qualitative forecasts are built on subjective information. The opinions of executives who are predicting future sales are just that–opinions. In addition, the manpower required to collect the data to create qualitative forecasts can be very expensive (Mentzer & Bienstock, 1998).

Quantitative Forecasting Techniques

Over the past 40 years a large and varied array of more sophisticated sales forecasting models have been developed. Mentzer and Kahn (1995) reveal that while many firms are familiar with the quantitative forecasting techniques available, most are still using less sophisticated models. Most of the companies that are using quantitative techniques state that they are satisfied with the results of these methods.

Lawless (1997) argues that it is imperative for firms to employ the quantitative methods available to them in forecasting sales. The combination of an environment that is constantly changing and downsizing in organizations makes reliance on the technology available more crucial. It also creates more credibility for those performing the forecasting function.

Two popular quantitative sales forecasting techniques are time-series analysis and regression analysis. They are both fairly easy to apply (Chase, 1997). Time-series techniques attempt to identify the patterns of the history of actual sales. If a pattern can be established, the forecast can be generated. Time-series relies exclusively on data generated within the organization itself (Mentzer & Bienstock, 1998).

Regression models have been found to be the most effective forecasting technique. Regression models provide a framework that insures the consistency of the sales forecasting process (Chase, 1997).

In regression analysis, a set of variables that are believed by forecasters to best predict the sales of a product, are chosen to generate the forecast. Forecasters are looking for the variables that correlate with the sale of the product. The line that best “fits” the relationship between sales and these other variables is used to do the forecasting (Mentzer & Bienstock, 1998, p. 81).

Both time-series techniques and regression analysis have weaknesses. They both assume that errors in data are independent, normally distributed with a zero mean and a have a constant variance. Often errors in past performance are difficult to detect, making it complicated when building a time-series or regression model (Mentzer & Bienstock, 1998).

Causal forecasting is a model recommended by Lapide (1999). This model is appropriate when the forecast is influenced by controllable, temporary events, such as promotions. It may also be used when uncontrollable, temporary occurrences such as a sudden change in demand or a series of ongoing events influence the quantity of the product consumers demand.

Too much reliance on quantitative models can cause problems. A reasonable combination of quantitative and qualitative techniques is required to optimize the efficiency of the forecasting process. Knowing when to use each type of model improves the accuracy of the model(s) chosen (Moon & Mentzer, 1998).

Bottom-Up Forecasting Techniques

Another decision that a firm typically makes concerning the choice of a sales forecasting model is whether to take a top-down or bottom-up approach to the forecasting process (Stanton & Spiro, 1999). Generally, bottom up approaches to sales forecasting are considered to be more accurate than top-down approaches (Dunn, William, & Spiney, 1971).

These bottom-up methods use information generated by those closest to the consumer to generate the initial data that is used to forecast the sales for an organization. This data can be generated by the salespeople or from data that is collected electronically at the point of sale (Gordon & Morris, 1997).

Critics of bottom-up methods agree that they work well at predicting the sales of individual SKU’s at lower levels. However, most bottom-up procedures do not take into account the overall effects of the economy, seasonal trends, and other variables that can influence the sales of a product (Kahn, 1998).

Top-Down Forecasting Techniques

Top-down forecasting techniques are seen as being more effective forecasting at the macro or aggregate level. These models typically smooth lower level information by accounting for the overall market conditions that can cause variations when the data are added together (Kahn, 1998).

Regression and other forms of correlation analysis were some of the first quantitative methods to be used to forecast sales (Pindyck & Rubenfield, 1976). These are classified as top-down methods as they start with the aggregate sales of a product. Time-series analysis has the ability to smooth out some of the seasonal variations that are generally seen as a weakness of bottom-up forecasting techniques (Kapoor, Madhok, & Wu, 1981).

Hybrid approaches are being developed that employ the advantages of both the bottom-up and top-down techniques of forecasting. While these approaches might produce better results in some situations, they are much more time consuming to implement (Kahn, 1998).

One type of hybrid approach that is gaining some popularity in industrial markets is the simulated test market. Current and potential customers are exposed to a company’s plans for new products and promotional campaigns. The customers’ reactions are then compared to past reactions of consumers to predict future sales. In addition to forecasting sales, the simulated test market can help an organization alter its marketing mix variables for new products (Clancy & Shulman, 1995).

Other Sales Forecasting Issues

Tools need to be designed by organizations to determine the success of their forecasts. Mentzer and Bienstock (1998) suggest that sales forecasting success be measured in three ways. The accuracy of the forecast is the most obvious of these measures. While this is a fairly straightforward concept, interpreting the accuracy of a forecast is more complex. If a forecast is not accurate, does it mean that the forecaster did not do his/her job or did the variables that went into that forecast change?

The forecast should also be evaluated in terms of its cost. The forecasting process incurs training, staffing and computer costs. These need to be weighed against the benefits that the organization perceives it is receiving through the forecasting method being employed (Mentzer & Bienstock, 1998).

Consumer satisfaction is the final area that needs to be measured. A company needs to know if consumers are satisfied with the actions taken by an organization as a result of the forecast it generated (Mentzer & Bienstock, 1998).

Another problem faced by those developing sales forecasts is a lack of quality data. Often past company sales figures that are provided to forecasters are inaccurate. Sometimes salespeople will distort the timing of sales to help them reach their quotas or succeed in sales contests (Geurts & Whitlark, 1996).

Many organizations are unclear about who should be involved in the forecasting process and even more unclear about who should be responsible for monitoring the results (Jain, 1998). Moon and Mentzer (1998) argue the need for a forecasting champion in organizations. This forecasting champion is responsible for the development, implementation and monitoring of the sales forecast for the organization.

Mentzer and Cox (1984) examine the factors that affect the accuracy of sales forecasts. They discovered that providing formal training to those doing the forecasting is the most significant factor in determining forecast accuracy. “The more formal training received, the more accurate the forecast” (p. 153).

Another reason many firms do not do a better job of producing accurate forecasts is their lack of lack of follow-up after the forecast has been made. Companies need to measure the effectiveness of their forecasts and provide feedback to those engaged in the forecasting process. Precise tools need to be developed to track the accuracy of forecasts against actual results (Moon & Mentzer, 1998).

The Salesperson’s Role in the Forecasting Process

As the number of approaches to the forecasting process varies, so do the opinions about if and how salespeople should be involved in the forecasting process. Sales forecasting is a mix of art and science (Keenan, 1995a).

Salespeople can contribute to the process if they can temper their optimism. One of the major concerns relative to the use of salespeople in the forecasting process is that they will predict sales figures that cannot be achieved (Keenan, 1995a).

“Optimism had no place in sales forecasting. Many companies believe that the sales force is too optimistic–perhaps too opportunistic–to be trusted to provide accurate enough input to the sales forecast” (Stack, 1997, p. 64). Salespeople need to be held accountable for the numbers they supply. Sales and marketing managers must insist on realistic numbers that they can depend on from the sales force (Stack 1997).

Another key is for sales and marketing managers to accept the information they receive from their salespeople. One of the reasons salespeople may not take being asked about future sales seriously is that their past efforts have not taken into account during the forecasting process (Stack, 1997).

Not involving salespeople in the forecasting procedure can have negative consequences as well. If salespeople have no input in the numbers that are dictated to them from above, there may be less of a commitment from the sales force to accomplish the forecasted sales figure (Keenan, 1995a).

The type of sales forecasting technique being used may determine the use of salespeople in the process. Bottom-up forecasting methods that start with estimates generated by those dealing with the customers need the kind of information that salespeople can bring to the process (Mentzer & Bienstock, 1998).

In very highly consolidated and less seasonal markets, salespeople are needed to provide information about individual accounts to those doing the forecasting. In more fragmented and seasonal markets, computer modeling for macro-level trend analysis should be a more important tool. Salespeople are still consulted in fragmented markets, but play a less significant role in this type of forecast. The more fragmented the market, the more a company should depend on top-down, macro models (Keenan, 1995).

Moon and Mentzer (1999) recommend the involvement of salespeople when a company has a small number of very large customers who buy sporadically. The salesperson is needed to relate to the company who will or will not buy during the upcoming forecasting period. The buying patterns of these large customers are very difficult to predict using top-down, quantitative forecasting approaches.

Recommendations

Based on the review of forecasting techniques and the role of the salesperson in developing forecasts, the following are recommendations concerning the use of salespeople in the forecasting process. Steps can be taken to include salespeople in the forecasting process in a meaningful way.

First, management must decide to involve salespeople in the forecasting process. Managers whose task it is to develop the forecast must be convinced that it is in their best interest to include salespeople in their plans and to take what the salespeople say about future sales potential seriously.

As part of this inclusion process, salespeople must understand that forecasting is part of their job. Each member of the sales force will be required to create monthly and annual sales forecasts (Moon & Mentzer, 1999).

For salespeople to take this task seriously, they must be motivated. Sales managers need to develop incentives for the salesperson to make accurate predictions about future sales.

Recognition and monetary rewards should be provided to salespeople whose performance matches the forecasts they have developed for their territories (Moon & Mentzer, 1999). A lack of incentives indicates to the sales force that management is not serious about the importance of the salesperson’s involvement in the forecasting operation.

Another important step for management to take to impress upon salespeople the importance of forecasting is to provide training (Mentzer & Bienstock, 1998). If salespeople in the past have not been involved in forecasting or the techniques for creating a forecast have been left up to the salesperson, it is imperative for management to provide training.

Another advantage of training the sales force in the area of forecasting is the improved efficiency that should result. If the entire sales force is exposed to the same training, it should result in more consistent results. If there are differences in results, they should be easier to identify when the manager knows the assumptions from which the salesperson should be working (Mentzer & Bienstock, 1998).

The involvement of salespeople in the forecasting function can also help salespeople better understand their customers and territories. The process of developing a forecast will force them to examine the issues related to their customers’ needs and wants. It will require salespeople to analyze how efficiently they are using their time and talents in attempting to generate sales from existing and potential customers (Keenan, 1995b).

Finally, it is important for managers to separate forecasting from quotas. Salespeople may be reluctant to express their honest opinions about predictions of sales in their territories for fear of having that number attached to a quota. This problem most often surfaces when a salesperson sees a potential increase in demand (Moon & Mentzer, 1999).

Managers need to reassure salespeople that while the forecasts they develop are used to calculate aggregate demand for a product, they are numbers that the salesperson should feel sure that she/he can achieve (Stack, 1997). One way that managers might accomplish this is to have forecasts done in units and quotas done in dollars. This distinction can help keep salespeople from thinking of forecasts and quotas as one concept. Different time periods have also been used to differentiate forecasts and quotas. Some companies use rolling forecasts over longer time periods than are used for quotas (Moon & Mentzer, 1999).

Limitations and Future Research

As the sophistication of techniques for forecasting sales increases, it is important for firms not to forget about the contribution that front line salespeople can make in helping to predict sales. As markets continue to change and competition increases in most markets, companies will attempt to find ways to better estimate future sales of their products.

While arguments are made here for why firms should be using salespeople in the forecasting process, future empirical research is needed to determine what effect salespeople have on the success of the forecasting effort. The empirical research referred to in this paper focuses on a limited number of firms and industries. Salespeople are a valuable resource for an organization. Firms that do not involve salespeople in the forecasting process are not fully utilizing this valuable resource.

References:

Bolton, R. & Chase, C. (1997). Odyssey of a forecaster. Journal of Business Forecasting Methods & Systems, 16, 1, 30-32.

Chase, C. (1997). Integrating market response models in sales forecasting. Journal of Business Forecasting Methods & Systems, 16, 1, 2-3.

Chase, C. (1996). What you need to know when building a sales forecasting system. Journal of Business Forecasting Methods & Systems, 15, 3, 2-3.

Clancy, K. & Shulman, R. (1995). Test for success. Sales & Marketing Management, 147, 10, 111-113.

Dunn, D., William, W. & Spiney, W. (1971). Analysis and prediction of telephone demand in local geographic areas. Bell Journal of Economics and Management Science, 2, 2, 561-576.

Geurts, M. & Whitlark, D. (1996). Improving sales forecasts by improving the input data. Journal of Business Forecasting Methods & Systems, 15, 3, 15-18.

Gordon, R. & Morris, C. (1997). A role for the forecasting function. Journal of Business Forecasting Methods & Systems, 16, 4, 3-7.

Herbig, P., Milewichz, J., & Golden, J. (1993). Forecasting: Who, what, when, where, and how. Journal of Business Forecasting, Summer, 16-21.

Hogarth, R. & Makridakis, S. (1981). Beyond discrete biases: Functional and dysfunctional aspect of judgmental heuristics. Psychology Bulletin, 90, 115-137.

Jain, C. (1998). Quick and easy way to monitor forecasts. Journal of Business Forecasting Methods & Systems, 17, 2, 2-3.

Kahn, K. (1998). Revisiting top-down versus bottom-up forecasting. Journal of Business Forecasting Methods & Systems, 17, 2, 14-18.

Kapoor, S., Madhok, P. & Wu, S. (1981). Modeling and forecasting sales data by time series analysis. Journal of Marketing Research, 18, February, 94-100.

Keenan, W. (1995a). Keeping sales in the loop. Sales & Marketing Management, 147, 6, 34-35.

Keenan, W. (1995b). Numbers racket. Sales & Marketing Management, 147, 5, 64-70.

Lapide, L. (1999). New developments in business forecasting. Journal of Business Forecasting Methods & Systems, 18, 2, 13-14.

Lawless, M. (1997). Ten prescriptions for forecasting success. Journal of Business Forecasting Methods & Systems, 16, 1, 3-5.

Marks, R. (1996). Personal Selling: A Relationship Approach. Boston: Prentice Hall.

Mentzer, J. & Bienstock, C. (1998). Sales Forecasting Management. Thousand Oaks, CA: Sage.

Mentzer, J. & Cox, J. (1984). A model of determinants of achieved forecast accuracy. Journal of Business Logistics, 5, 2, 143-155.

Mentzer, J. & Kahn, K. (1995). Forecasting technique familiarity, satisfaction, usage, and application. Journal of Forecasting, 14, 5, 465-476.

Moon, M. & Mentzer, J. (1999). Improving salesforce forecasting. Journal of Business Forecasting Methods & Systems, 18, 2, 7-12.

Moon, M. & Mentzer, J. (1998). Seven keys to better forecasting. Business Horizons, 41, 5, 44-52.

Morrison, J. (1996). How to use diffusion models in new product forecasting. Journal of Business Forecasting Methods & Systems, 15, 2, 6-9.

Paley, N. (1994). Welcome to the fast lane. Sales and Marketing Management, 146, 8,   2-3.

Peterson, R. & Minjoon, J. (1999). Forecasting in wholesale industry. Journal of Business Forecasting Methods & Systems, 18, 2, 15-17.

Pindyck, R. & Rubenfield, D. (1976). Econometric Models and Economic Forecasts. New York: McGraw Hill.

Smith, S. & McIntyre, S. (1994). A two-stage sales forecasting procedure using discounted least squares. Journal of Marketing Research, 31, 1, 44-56.

Stack, J. (1997). A passion for forecasting. Inc., 19, 16, 37-38.

Stanton, W. & Spiro, R. (1999). Management of a Sales Force. Boston: Irwin/McGraw Hill.

Articles, Marketing, Research
Delphi technique, Quantitative Forecasting, Sales Forecasting, sales forecasting model, sales forecasting process