Companies large and small engage in forecasting their sales in order to decide with the greatest accuracy possible what and how much to build or what and how much to buy. Sales forecasts support operational planning and supply chain management—not marketing or sales efforts. To be sure, sometimes accurate sales forecasts can also lead to intensified selling or its reverse: sales forces fanning out to dampen buyer enthusiasm because the company cannot actually supply the demand foreseen. But such efforts are not central to forecasting. Forecasts are made to project the business into the future accurately in order to avoid costly mistakes.
A classical example of sales forecasting is the determination of "the build." This phrase is used in industries where production falls into one season and sales into another. Thus snowmobiles are manufactured in the summer and sold in late fall and winter. How many sleds to build, how many of the low end and how many of the expensive models—"the build," in other words—must be decided long before the winter season actually arrives, before likely snowfall is predictable, before economic conditions influencing sales then can be known with certainty now. In "bad snow years" demand tends to drop, in "good snow years" it spikes. Snowmobiles are big-ticket items. Overbuilding can hurt the company and its dealer structure for longer than one season; under building leaves substantial money on the table and may lead to a loss of market share if the competitor made a better forecast. Builders of boats and similar "water toys" have the reverse problem: they build in the winter for spring and summer sales. Boat builders, like snowmobile manufacturers, require positive economic times in that they sell high-end recreational products users can do without. Both are influenced by weather. Boat sellers can also be hurt by high prices for fuels.
"The build" serves as an easy illustration of the need for good sales forecasts, but every producer and every merchant faces exactly the same problem in looking ahead. In every case more or less irreversible actions must be taken in advance of actual sales; in every case multiple factors influence future demand which may change in response to yet other factors; in every case bad forecasts may mean being saddled by large inventories or having to turn customers away. Not surprisingly, every business, even the smallest, engages in some kind of sales forecasting. It may be quite instinctive and informal—a gut feel by the owner that more or less should be purchased or made, based simply on experience leading up to the purchasing decision. Sales forecasting is often a process involving contact with the sales channel. In many large companies producing for mass markets, sales forecasting is a very complex, formal, and highly structured activity involving expensive surveys, computer modeling, and statistical analyses.
BASIC TECHNIQUES
Fundamental approaches to forecasting sales rely on 1) looking at the company's own history (internal numbers), on 2) looking at the product's or category's market history (external numbers), 3) soliciting external opinion (channel surveys), and 4) examining other sources of information which indirectly influence the future.
In the first case the company will look at its own past sales and determine a trend, ideally based on units rather than on dollars to eliminate the effect of price changes. If the item is growing at 2 percent a year, the company may feel safe in increasing its production/purchasing by 2 percent for the next period. Such a forecast is typically just the start of a process of review. The company may wish to eliminate the product because its margin is low and decreasing, its warranty service requirements are too great, etc. Alternatively, the company may wish to increase its growth in the category by additional promotional, discount, and sales efforts—and, betting on success, may order above its historical trend projection. Quite complicated formulas are sometimes applied. An example is production of replacement parts for outdated models of a product—in which the forecast is dated on an estimate of the models still remaining "out there" in active use—with the production reduced each year.
The second case, looking at the total market for the item, requires access to data on such sales. If these are available, the company can compare its own performance against the product's growth/decline as a whole and make adjustments accordingly. Suppose the category, e.g., a certain type of garden tool, has been declining as a whole in the gardening field while the company's own sales of that product have been increasing at 5 percent a year. This may mean that the company may have become the last active supplier of a product in its locality thus drawing a segment of the public that still wants the product. Such a finding may lead to energetic stocking up. Conversely, if the company's sales are poor but the product as a whole is making waves, adjustments in price, promotion, display, and the like may justify much more ambitious stocking. In practice it is often very difficult to get objective data on the performance of a specific item for comparison. Similarly, even if overall sales data can be found, it may be very difficult for a merchant to discover why he or she is selling more or less of an item. The merchant's location, clientele, region of operation, and many other factors may influence the result. The small business typically lacks the time and money to go deeply into such a subject unless the product is rather expensive and central, e.g., the business sells farm equipment.
