Sales Forecasting ‘The purpose of a sales forecast is to enable effective planning of business processes, The sales forecast is used by several other functional departments: (f) the manufacturing or production department: for setting up production capacity and planning production; (ii) the finance department: for raising cash forinvestment and ‘operations as well as for profit planning; (if) the purchase department: for planning their purchases; and (iv) the human resource department: for manpower planning. Thus, the sales forecast has a role asa forerunner to all planning activities in.an organisation. Accuracy of the sales forecast is important because all functions base their plans on such forecast, tis the responsibility of sales and :narketing managers to prepare the sales Forecast ‘Types of Sales Forecast ‘The term “forecast” is generally used ta describe a prediction for.a future period, such asa weather forecast. A sales forecasts the estimated number of units or monetary value of sales. for a specific future time period based on an assumed marketing environment and a proposed marketing plan, Firms refer to sales Forecast by defining three factors: (I) product level, (i) geographic area, and (i) time period. These factors are further subdivided into subfactors. as shown in Fig. 5.2. Thus, a total of 6 x3 x 5 = 90 different types of sales (or demand) forecast can be made, one foreach combination of product level, time period, and geographic area, Figure 5.2 ‘Types of Sales Forecasting [> Alcota Sales f—> Industry Sales —+ iy, apt om Product Line Sales |-—® Product Variant (orm) Sales ‘—» Product Item Sales: — Long Range (erm) Shes ot» Poet ls eum argo rem) Forecast L—» short Range (Term) -—> Woria }-—-» Nation (India USA) Region (North’South) ‘Geographic * area }—> Territory (Branch/Distict) ‘— CustomerBasic Terms Used in Forecasting Before we discuss approaches and methods of forecasting, it is necessary to define and understand some of the basic terms used in forecasting sales or demand. These terms are closely related, used loosely and therefore, cause misunderstanding among academicians and managers. Market potential It is the best possible (or maximum) estimated sales of a given product or service for the entire industry in a given market for a specific period of time. For example, the market potential for personal computers (PCs) in India for the year 2005-06 is estimated to be 4 million numbers. Market potential is also called as industry sales forecast. The following four parts must be included for a complete definition of market potential as well as sales potential and sales forecast. 1. Item marketed, such as a product, or service. 2. Sales estimated in units or value (Rupees or Dollars). 3, Description of the market by a geographical area, or type of customer, or both. 4. A specific time period, such as a particular year. Market forecast It is the expected industry sales of a given product or service at one specific level of industry marketing expenditure, ina given market, fora specific period of time. Market forecast is also called as market size. For instance, market size of talcum powder in the organised sector in India, for the calendar year 2004, was Rs, 700 crore (Rs, 7000 million). Sales potential (or company seles potential) It is the best possible (or maximum) estimated sales of a given product or service for a company in a given geographic area for a specific period of time. Sales potential is also defined as the maximum share (or percentage) of market potential that is expected to be achieved by a company. For instance, sales potential of ICICI-Prudential is expected to be close to five per cent of the gross premium collection of life insurance industry in India in coming years. Soles forecast (or company sales forecast) Itis the estimated company sales of a given product or service, under a proposed marketing plan, in a given market, for a specific period of time. There is a relationship between the company’s sales forecast and proposed marketing expenditure (or marketing plan). A company may make a sales forecast for an entire product line (like detergent from Hindustan lever) or a product item (such as wheel brand). Sales potential and sales forecast are usually not the same. Sales potential is what a company would achieve under ideal conditions. Typically, the sales forecast is less than the sales potential, for different reasons, such as limited production facilities, and inadequate financial resources. Seles budget Itis the estimate of expected sales volume in units or revenues from the company’s products and services and the selling expenses. Sales budgets are generally set slightly lower than the sales forecast, in order to avcid excessive risk. Sales budget is used for making purchasing, production, manpower and cash flow decisions. The sales budget goes into complete details of expected sales of each product item. We shall discuss more on sales budget subsequently in this chapter.Sales quota Itis a sales goal (or a performance goal) set for a marketing unit for a specific period of time. The marketing unit may be a salesperson, a branch, a region, a dealer, or a distributor. A company ‘management sets sales juotas on the basis of the company sales forecast. More on sales quotas will be discussed later in this chapter. FORECASTING APPROACHES Two basic approaches of forecasting, as shown in Figs 34 and 35, are: (1) Topdown (or Break- down) approach, and (2) Bottom-up (or Build-up) approach. Top-down/Break-down sporoach In this approach, typiczlly the company sales forecast is developed at the business unit (or SBU) level, by using suitable forecasting methods. The head of sales/'marketing then breaks down the company sales forecast into region, district, territory, salesperson, and individual customer sales quotas. The steps followed in this approach are shown in Fig. 3.4. cealiBO., Sales and Distribution Management Forecast Relevant External Environmental a [7 Company Sales Potential = industry Sales Forecast (step 2) ag] owiia Companys Share of Total Industry Sales in Percentage r wets ao ~~ Sales/Marketing Manager's Forecasts for Regions, ‘step 5 we gs oe TFRFA Development of Top-down Approach ‘The key steps in Fig. 3.4 are step 2 and step 3. Instep 2, some of the methods used for forecasting industry sales (or market potential) are: (i) delphi method, and (ii) regression analysis, which will be described subsequently in this chapter. In step 3, for estimating the company’s share of the total industry sales, a number of factors must be considered, including the company’s current market share, target customers and their perceptions about the company’s performance on key factors like quality, service, and price in comparison with major competitors, and the company’s relationship with major customers. With respect to step 4, the company sales forecast is usually lower than the company sales potential due to insufficient funds, increase in competition, or shortage of raw material. The last step of breaking down the company sales forecast to different regions and territories is done based on market potential of different geographical areas. For this, {wo major methods are available: (1) market build-up method, primarily used by business marketers, and (2) multiple-factor index method, mainly used by consumer marketers.Market-Buildup Method ‘The first step in this method is to identify existing and potential business buyers in the geographical territory. The second step is to find out their potential purchases of the product under study. The final step is to add-up the business potential of all the buying firms to obtain a fairly accurate estimate of market potential for the product or service for a specific ‘geographic territory. Multiple-Factor Index Method For the consumer marketing company, the estimation of market potential for a geographical area is done not by identifying individuals and households, because they are very large in numbers. This method first identifies the factors that influence the sales of a product or service. Generally, there are more than one sales factors, such as population and income that influence sales. These factors are given certain weights, corresponding to the degree of sales opportunity. Let us take an example of a company, manufacturing and marketing detergents all over India. The company wants to find out the market potential of detergents in all cities, including Bangalore. The major factors that influence sales of detergents are population, personal income and retail sales, which are given weights of 0.4, 0.3, and 0.3 respectively. Suppose, Bangalore has 0.7 per cent of India's population, 1 per cent of India’s disposable personal income, and 0.9 per cent of Indias retail sales. The multi-factor buying index for Bangalore would be: 04 (0.7) +0.3 (1) + 0.3 (0.9) = 0.85 Based on the Indian detergent industry forecast of Rs 55000 million for the year 2005-06, the market potential for detergents in Bangalore would be: 0.85 per cent of Rs 55000 million, i.e. equal to Rs, 467 million for the year 2005-06. Botton-up/Build-up approach Bottom-up or build - up approach starts with the company’s area or branch managers asking its salespersons to estimate or forecast the sales in their respective territories, as shown in Fig. 3.5. Combined into Company Sales Forecast Step 4 eee rss [combined io RegionatZonal Sas Forecasts | step eniaiaiaia r Combined into Area/Branch Sales Forecasts | Step2 a & ‘Salespersons’ Sales Forecast of Individual Customers | Step 1 TBE, Development of Bottom-up or Build-up Forecasting ApproachSalespersons are given guidance by their respective area/branch sales managers on how to get information from the existing and potential customers on the estimated purchases of thecompany’s products services for the specified future time period. Each area or branch manager then adds the sales forecasts received from the salespersons, modifies the same wherever needed, and sends the combined sales forecast figure for each product in units and value to his superior, that is, regional or zonal sales manager. Each regional/zonal manager adds the sales forecast received from the area or branch managers, modifies the same, if needed, and sends the regional sales forecast to the sales/ marketing head. The head of sales /marketing repeats the process and presents his proposal to the CEO for discussion, modification and approval of the company’ sales forecast. ‘Some companies use both top-down and bottom-up approaches, inorder to increase their confidence in the sales forecast. However, comparison of the two approaches indicate that top- down approach needs less time and cost, as it uses data on forecast from secondary sources like planning commission report or market forecasts, and indicators by Centre for Industrial and Economic Research, New Delhi. Bottom-up approach is very accurate for short-term forecast (upto ‘one year), as it is based on primary data collection, which means more cost and time. Firms may develop sales forecast internally or buy forecast from outside sources such as marketing research companies, or specialised forecasting firms like Centre for Industrial and Economic Research. We Sales Forecasting Methods As shown in Fig. 3.6, sales forecasting methods or techniques can be classified as: (1) qualitative methods, and (2) quantitative methods. + [_Exeaative Opinion [+] Delphi Method + + [ Bates Force Composite ‘Survey of Buyers’ Intentions: "Test Narketing ‘Qualitative Methods ——>{ Moving Averages —+[[ Exponential Smoothing Mothode ‘Quanttative +[_Decomposiion 7] | [ive Ratio Method | —+[ Regression Analysis |__,F conemetric AnalyssExecutive Opinion Method ‘This is the oldest, simplest, and the most widely used method, A study of one hundred and fifty companies found that 86 per cent used executive opinion method to forecast sales, The method involves getting the views of top executives of the company regarding future sales, The sales forecasts are made either by taking the average of all the executives individual opinions or through discussions among the executives. Some executives’ opinion may be supported by the use of forecasting methods, and other executives may form their opinion based on experience, judgement, and intuition, ‘The advantages of this method are: (i) forecasting can be done quickly and easily, (Hi) less expensive than other methods, and (ili) very popular, particularly among small and medium-sized companies. However, there are some disadvantages: (a) un: (b) subjective, and (©) difficult tobreak-down the forecast into subunits (ike regions, branches) of the organisation. Delphi Method ‘This method is similar to the executive opinion method, and was developed by Rand Corporation during the late 1940s. The difference is that the members of expert panel do not meet or discuss in a committee. The procedure includes selection of panel of experts from within and outside the organisation. A coordinator asks each expert separately to make a forecast on some matter, Each member of the expert panel submits histher forecast anonymously. The coordinator summarises these forecasts into-a report that is sent panel member. The experts are then asked to make another prediction seperately on the same matter, with the knowledge of the forecasts of the other experts on the panel. This process is repeated until the panel of experts arrives at some consensus. The basic belief in this method is that experts, without any pressure or influence will develop a more accurate prediction of the future. ‘The advantages of this method are: (1) objective forecast that is accurate, (ii) useful for technology, new product, and industry sales forecast, and (iii) both long and short-term forecasting possible. However, the disadvantages are: (a) difficulty getting a panel of experts, (b) longer time for getting consensus, and (c) break- down of forecast into products or territories is not possible. Sales-force Composite Method ‘This method involves salespeople so estimate their future sales. It is an example of bottom-up approach and is, also called a “grass-roots” approach. Each salesperson estimates in his/her territory how much quantity or value existing and potential customers will buy of each of the company’s products and/or services. This method is often used by industrial or business marketing companies Sales representatives make the sales estimate in consultation with customers and sales supervisors and/or based on their experience and intuition. The company sales forecast is made up (composite) of all the salespersons’ sales forecast, ‘The advantages of this method are: (i) forecasting is done by salespeople who are closest to the market and have better insight into sales trends than any other group in the company, (il) detailed sales estimate broken down by customer, product, sales representative and territory are possible, and (ii) involvement of sales people The disadvantages include: (a) sales forecast are often pessimistic or optimistic, as salespeople are not trained in forecasting, (b) if sales forecast are used to set sales quotas, which are linked to incentive schemes, salespeople may deliberately underestimate the demand, and (¢) many salespersons are not interested in sales forecasting, andl prefer to spend time in the field meeting customers. To encourage better forecasting, companies are adopting some of the following modifi earlier practices:sailes; the company assumptions 1. Salespeople are given information on their past forecast with actu; s, strengths, on the business outlook during the forecasting period, majorcompetitors’ strategies, ta and weaknesses, and the company’s proposed marketing plan. ‘The salesperson and the immediate sales supervisor work out the sales estimates independently. Then they meet and reconcile any differences in the figures. The same process is carried out between the sales supervisors and the territory manager, and later on, between the territory managers and the sales manager. 3. Conducting training programmes for salespeople on the process and methods used for fore- casting sales, Survey of Buyers’ Intentions Method ‘This method is sometimes called as market research (or market survey). Itincludes asking existing and potential customers about their likely purchases of the company’s product and services for the forecast period. For instance, a question like the following is asked: Do you intend to buy a refrigerator within the next six months? 00 (notatall) 02 Glight possibi 04 —— (fairpossibility) 06 (good possibility) 08 (high possibility 410 (certainty) ‘The above is called a purchase probability scale, The customers are also asked other questions, such as product quality, features, price and service, which are alla par of the questionnaire, The information collected from buyers helps the company to make effective decisions not only in sales and marketing areas, but also on production, research and development. Several research organisations conduct periodic surveys of consumer buying intentions. They combine various bits of information into a consumer confidence measure. One such survey is done every quarter since December 2002 in India by Economic Times, which measures the consumer optimism through consumer confidence index (CC), Even though this method is grouped under qualitative methods, it can be classified as a qu ‘method, where the respondents are selected based on probability sampling techniques, and analysis is done ‘with multivariate statistical tools. ‘The advantages of this method are: (i) useful in forecasting sales for industrial products, consumer durables, and new products, (il) it also gives customers’ reasons for buying or not buying; (ili) relatively inexpensive and fast, when only a few customers are involved (for example business buyers’ survey). The disadvantages are: (a) sometimes buyers are unwilling to reveal their plans, (b) buyers are somet (©) expensive and time-consuming in consumer non durable markets where consumers are very large in number. tative Test Marketing Method ‘This method is useful for forecasting sales for a new product, which has no historical (or previous) sales figures. It can also be used for estimating sales for an established product in 2 new territory. Major methods Used for consumer-product market testing include: (1) full-blown test markets, (2) controlled test marketing, and (3) simulated test marketingFull-blown test marketing _ In this method, the company chooses a few (two to six) representative cities in full promotion campaign is introduced, similar to what would be done in national marketing. The duration of the market test varies from a few months to one year, depending on the repurchase period of the new product. Buyer surveys are carried out to get information about consumer attitude, usage and satisfaction towards the new product. Ifthe test markets show high trial and repurchase rates, the product should he launched nationally; if they show a high trial rate and slow repurchase rate, the new produet should be redesigned or dropped: if they show a low trial rate and a high repurchase rate, the product is acceptable, but more consumers should try it; if both trial and repurchase rates are low, the new product should be dropped permanently. Controlled test marketing The company with the new product hires a research firm and gets a panel of stores at specified geographic locations. The research firm delivers the new product to the panel of stores, arranges promotionsat the stores, and measures the sales of the new product. The research firm also interviews sample consumers to get their perceptions on the new product. Both full test marketing ane controlled test marketing expose the new product to the consumers. Simulated test marketing In this method, about 30-40 consumers (or shoppers) are selected, based on their brand familiarity and preferences in a particular product category, such as baby care and soft drink products. These shoppers are shown commercials or print advertisements of well-known products and also the new product, without any specific mention, These consumers are given asmall amount of money and asked to buy any items in.a store. The researcher of the company notes how many consumers buy the new product and compéting products. These consumers are interviewed to find reasons for buying or not buying, and later, after usage of the new product, satisfaction levels and repurchase intentions. This method tends to give quite accurate results, For industrial-product market testing, the methods used are alpha testing (testing within the company), and beta testing (with outside customers) for high cost and new technology products and services. For example, Infosys sent its banking software product, Finnacle, to the experts for a thorough check-up to see it was fit for the multi-billion dollar US market. The supplier of the new product/service will ask the customers about their purchase intentions and other information after the beta testing. Another method used for introducing new industrial product is participating in the industry trade-shows. ‘The main advantages of these methods are: (a) their usefulness for forecasting the sales of new or modified products, and (b) in deciding whether a company should go a head for a national launch of a new product, without spending a huge amount. The disadvantages are: (a) for some of the methods like fold-blown test market and controlled test marketing, where there are possibilities of the information on new products going to ‘competitors, there are chances of spoilage of the test marketing, and (b) if repurchase period is long, particularly for consumer durables, itis difficult for the company to wait to measure test results, In such cases, the company decides to inteoduce a new product in a small geographical arca and subsequently ‘roll on’ to-other areas, in a planned manner. Moving Average Method This is a relatively simple method that develops a company forecast by calculating the average company sales for previous years. As shown below, a company's sales forecast was worked out by calculating moving averages for three to six year time periods. The formula used is:Table 5.20 An Example of Moving Average Method (Figures of sales are in Rs Million) ‘When a forecast is developed for the next period, the sales in the oldest period is dropped from the average and is replaced by sales in the newest period; hence the name “moving averages”. If the company operates in astable environment a short two or three year moving average may be most useful. Moreover, if a firm is in an industry which exhibits cyclical variations, the moving avenge should use data, equal to the length of a cycle or a longer averaging period. The advantages of this method are: (a) relatively simple method, (b) easy 10 calculate, and (c) widely used for short-term and medium-term sales forecasts. The disadvantages are: (a) unable to predict a downtum or upturn in the market, (b) cannot predict tong-term sales forecast accurately, and (©) historical data is needed. Exponential Smoothing Method ‘This method is closely related to the moving averages method for sales forecasting. Using the exponential smoothing equation, the forecaster can allow sales in certain periods to influence the sales forecast more than sales in other periods. The formula for the exponential smoothing method is as follows: F=ah +0-OR ‘The forecaster decides the value of the smoothing constant 4 based on: (a) review of sales data (b) knowledge and observation about the conditions in the forecasted period and conditions in previous period, and (c) intuition. For instance, a smoothing constant A with a high value of 0.7 or 0.8 allows most recent periods of actual sales to influence sales forecast more than sales in earlier periods. Whereas a smoothing ‘constant with a low value of 0.2 or 0.3 allows earlier periods of actual sales to influence forecasted sales more than sales in recent periods. The smoothing constant 4 can range between O and 1. In the example above (Table 5.21), the forecasted sales for the year 2008 can be calculated by using the ‘exponential smoothing method, with smoothing constant 2 =0.2. The sales forecast forthe year 2008 would be 0.2(956)+ 0.8(869.83)=Rs 887.06. ‘The advantages of this method are: (a) simple to operate, (b)forecast’s knowledge or intuition can be used. in forecasting, (c) useful method when sales data have a trend era seasonal pattern, (d) immediate response to an upturn or downtum in sales, and (e) used by many firms. The disadvantages are: (a) smoothing constant is somewhat arbitrary, and (b) long-term and new product forecasting are not possible.Table 5.21, An Example of Exponential Smoothing Method - —_,- (Figure of sales are in Rs Million) Exponential smoothing can be improved by including a trend component. This method is known as double exponential smoothing. Decomposition Method In this method the company’s previous periods’ sales data is broken down (or decomposed) into four major components, such as trend, cycle, seasonal, and erratic events. These components are then recombined to produce the sales forecast. Let us consider this method continuing the example above. Assume that various analysis have broken down (or decomposed) the previous sates data into the following components: a growth of 3 per cent in sales due to the development in technology, capital formation and population (trend component), increased terrorist activities are expected so reduce sales by 5 percent (erratic events component), a 10 per cent reduction in sales is expected due to a recession in demand (cyclic component), and the sales in the third quarter of the year are expected to go up by 15 per cent due to festive season, as compared to other three quarters (seasonal component). ‘The forecaster would combine the different components, as follows, in order to forecast sales for 2008. ‘Suppose that the total sales in 2007 was Rs 956 million. The trend component shows that 2008 sales would be Rs 985 million (= 1.03(956)). The sales are reduced due to introduction of erratic event component to Rs 936 million (= 0.95(985)). The sales forecast changes further due to cyclic component of recession to Rs 842 million (= 0.9(936)). Thus, the annual sales forecast for 2008 is Rs 842 million. The quarterly sales forecast would be Rs 210 million (= 0.25(842)), if the scasonal component is not considered. The seasonal component (due to, for instance, festive season) in third quarter would suggest 15 per cent increase in sales forecast, that is Rs 242 million (= 1.15(210)) forthe third quarter and consistent sales forecast of Rs 200 million (=(842-242)/3) each for the other three quarters. From the above example, we notice that trend, cyclic and erratic events are included in the calculation of annual sales forecast, while the seasonal component is used for forecasting sales for less than a year, like quarterly or monthly sales forecast. The major advantage of this method is that it is conceptually a sound ‘method. However, the disadvantages of this method are: (a) difficult and complex statistical methods are needed to breakdown sales data into various components, and (b) historical data is needed.Naive/Ratio Method Naive or ratio method sa time series method of forecasting, which is based on the assumption that what happen- ed in the immediate past will continue to happen in the immediate future. The simple formula used is as follows: Let us continue with the above example, and forecast the sales for the year 2008. The actual sales of the ‘current year (2007) is Rs 956 million and that actual sales of last year (2006) was Rs 948 million. The next year (2008) sales forecast would be Rs 964 million (= (956)(956)48). ‘The advantages of the method are: (a) simple tocalculate, (b) requires less data, and (¢) accuracy is good for short-term forecast. However, the method also has certain disadvantages such as: (a) tcan not be used for forecasting sales for long-term periods and new products, and (b) accuracy of sales forecast would be less, if past sales fluctuate considerably. Regression anslysis This isa statistical forecasting method that is used to predict sales, called as dependent variable ‘Y’. ‘The company then identifies causal (cause and effect) relationship between the company sales and the independent variables (or factors), which influence the sales. If there is only one independent variable (x), say promotional expenditure, itis plotted on a graph of paired data of past sales and promotional expenditure. This is called linear(or simple) regression, with only one independent variable, and different pattern of relationships, as shown in Fig. 3.6. Curviinear Linas Poti Linear Negative a Roletonsop © promotional Exenditure ~ Sees Gata HRB] Relationships in Rogression Analysis ‘To show the relationship between paired data (x, y) called a scatter diagram, the simple method is to draw a freehand line on the plotted points, keeping in mind to minimise the distances of all the points from the line. This trend line can be extended to project sales in future periods. A more accurate method of finding the “best-fitting line” is to use the least-squares formula for a straight line, y=a+ bx, where ais the intercept of the line on y axis, and bis the slope or trend of the line. In practice, the company sales are influenced by several independent variables, such as price, promotional expenditure, and population. To forecast the effect of several independent variables on the company sales, the method used is called Multiple Regression Analysis. The availability of computer software forecasting packages such as Statistical Analysis System, (GAS), and Statistical Package for the Social Sciences (SPSS), has increased the usage of regression analysis in many companies. The advantages of regression analysis are: (a) high forecasting accuracy, ifrelationships between variables are stable, (b) objective method, and (c) can predict turning points of the company’s sales. The disadvantages are: (a) technically complex, (b) can be expensive and time consuming, and (¢) use of computer and software packages are essential.Econometric anelysis In this method, many regression equations are built to forecast industry sales, general economic conditions, or future events. The procedure followed is as follows: To find out which factors or variables influence sales and the relationships between sales and these factors as well as the interrelationships between the factors, develop a number of regression equations representing these relationships. A forecast is prepared by solving these equations ona computer. The major advantage of this method is that accurate forecast of economic conditions and industry sales are possible. The disadvantage is thata large volume of data is required representing, the various factors.