Sales Forecasting

Sales forecasts directly affect your company's future profits. If you forecast too little, you lose opportunities. If you forecast too much, you run the risk of burdening the company with excessive costs and working capital. Let’s look at what we can learn about sales forecasting from the great minds of the past.

Forecasting has a certain mystique that can discourage decision makers. There are a variety of forecasting methods: from “hunches” or intuitive assessments to pure mathematics, and from very simple techniques to very complex and sophisticated ones. Learning how our predecessors did it before Excel and PC existed can be a good start.

?

I was surprised to find that computers were mentioned in the context of business forecasting already in the late 60s. It was mentioned in a negative context. At that time, computers were referred to as “the biggest villain for decision makers”.” The result of computation was presented in the form of “print-outs” and decision makers often did not know how the input data was processed. It was said that if you put in enough variables, the result looked good and you could claim it was objective and scientific. Back then, this process received the name GIGO – garbage in, garbage out. And I think, now, 60 years later, GIGO is still a possible scenario.

?

Even though it is good to use scientific methods and the computer is not a villain but an obedient helper, forecasting is not an exact science. A good forecast is a blend of scientific methodology and human skills. And the forecaster is the most important factor in the quality of the outcome. What makes a good forecaster? The old sales management literature suggests the following characteristics:

- Deep knowledge of the industry

- Understanding of scientific methodology

- The ability to work with “organized knowledge”

- Can recognize and apply cause-and-effect relationships

- Flexible in his or her assessments???? ????

??

The common mistakes (and I did some of them in the past) of forecasting:

- Complicate it with a wealth of mathematical and statistical formulas and derivations, but also make some simple arithmetical errors, especially in addition and subtraction.

- Don't let others help you with your prediction. Why share the credit when you want it all to yourself? But stick to one path and your own. Don't be fooled into thinking that that it is possible to get information or results using different approaches.

- Leave the predictions to the computer. That way, you don't need any of your own ideas. Just stick with the latest software. You can at least claim that your prediction is objective because you didn't contribute anything.

- Develop forecasting methods that don't rely on cause-and-effect relationships, that involve hard-to-get flawed data, and that cost a lot of time, manpower, and money.

-Make all predictions definitive. Never revise them in light of new information.

?

The basis for any forecast for existing businesses is historical data. It's the same idea as in #TheForgottenArtOfsales – study the past to understand the future. But in the days of Melville W. Mix nephew of William W, Dodge, gathering and organizing information was a very expensive and time-consuming manual labor.

?

Now such data can be few clicks away. But it is still important to make sure that the right type of information is collected and organized or “tabulated” as it was called in the past. “Figures of precious years’ sales are as useful to the manufacturer as harbor soundings and charts to the sea captain.”

?

The collected data was analyzed to identify trends, extrapolate and set benchmark. The historical data is good for all kind of quantitative analyses. For example, you can do regression analyses and analyze relation between number of sales and price, advertising and other variables. But, as Melville W. Mix was saying, “the figures alone do not tell the "why”." With a right set of data you can make a detailed picture of what was happening in the past, but without understanding why it was happening it can be dangerous to use it for predictions of the future results.

?

The information that was used for qualitative analyses:

- information concerning business conditions

- the developments in the trade and among customers

- the agents' methods and the conditions which surround transactions.

- trade publications

- experts opinion

?

The century old approach to sales forecasting is looking surprisingly modern and effective. It was combination of qualitative and quantitative methods, which is allowing to provide balanced perspective. Quantitative approaches use past data to create objective, data-intensive, long-term sales forecasts using mathematical models and statistics. And qualitative methods rely on expert knowledge, market research, and subjective assessments to identify shifts in sales trends and produce short-term forecasts. By combining these methods, a company can leverage the respective strengths: quantitative data serves as a historical basis and qualitative data provides insight into anticipated changes and future patterns, resulting in a more accurate and robust sales forecast.

?

But Mix was making one additional step in preparing his sales pans. He collected and analyzed separately agencies sales data tabulated in easy to analyze cards.

?

The cards contained information about sales, territory, conditions, general facts about agent, selling aids in from of advertising and cost. Mix was using this information to evaluate sales methods and correcting the sales plans by planning to copy some of the most important methods in other territories.

?

In conclusion, once again we can see that #TheForgottenArtOfSales can something to teach us even in the age of technology. Many principles that were used a century ago are still valid today.

?