Sales Forecasting: methods, models and how to create an accurate forecast

Sales are typically the primary source of financial resources for businesses, essential for operations and growth. Therefore, sales forecasting, much like cost planning, is a crucial element of company management. So, what are the methods for effective sales forecasting? Forecasting is the result of applying scientific methods to predict events, phenomena, or trends that are expected to occur in the future. In the context of sales, forecasting involves determining its most approximate value at a given time (e.g., a year), as well as estimating its dynamics in subsequent periods, such as individual quarters or months.

Identifying the Predictable in the Unpredictable: Analyzing Vast Amounts of Information

It should be remembered that every forecast is merely a prediction of future events, so it is obvious that determining their occurrence is always subject to some risk – greater or lesser, depending on the type of event. While some events are almost certain or predetermined (e.g., the fact that the sun rises and sets every day), even for these, at least theoretically, we cannot be sure whether they will actually happen the next day, a month, or a year from now. Of course, events, circumstances, and factors influencing sales are not as constant as astronomical laws. On the contrary, they show enormous variability. The number of variables that directly and indirectly determine sales is truly significant, and a large portion of them are difficult to predict and often depend on a random factor. Therefore, in practice, one may encounter the statement that there is no absolutely effective method for determining future sales, i.e., one that provides an infallible and certain forecast. Moreover, there is never a guarantee that even the most reliable forecasts will prove accurate. Yet, in planning resources, employment, production, or investments, it is always better to have some sales forecast than to have none at all… Sales forecasting is a complex and relatively difficult field, requiring solid knowledge of mathematics, statistics, and econometrics. Since sales depend on many variables, the basic principle in planning them is to analyze historical data. It is worth adding here that the information typically available to a company can be divided into ‘hard’ data, describing specific facts, events, or numbers, and ‘soft’ data, based on individual employee experiences, or even their intuition.

Variables Shaping the Sales Forecast

What variables can influence sales, and which of them should be considered in the sales forecasting process? They can be divided into internal and external factors.

Internal factors influencing sales levels

prices of offered products and services,
quality parameters defining the competitiveness of offered products and services,
quality of marketing materials and communication in the sales process,
quality of customer service or the quality of customer experience at every stage of cooperation with the company and at every touchpoint (so-called customer experience),
motivation level of sales and marketing employees and their job satisfaction within the company,
various types of promotions, their alignment with specific market segments, and their impact on customer purchasing decisions,
ethics in operations at various organizational levels,
type of active sales and marketing activities undertaken towards potential and existing customers, and the frequency of these activities,
quality of information management in sales, customer relationship management, and the sales funnel,
quality of communication conducted by the company with its environment outside the sales area,
speed of response to inquiries, order fulfillment time, and delivery of products and services to customers,
quantity, quality, and productivity of available resources within the company,
available distribution network and partners influencing sales levels,
scope of operations and communication reach in the sales process (degree of utilization of various tools).

External factors influencing sales levels

quality parameters of products and services offered by competitors,
competitors’ activities in sales, distribution, promotion, communication,
prices of competitive products and services,
quality of customer service and support from competitors,
customer opinion about the company’s products and services and those of its competitors (strength of individual brands),
customer loyalty, their attachment to a given brand, purchasing habits,
macro conditions – legal, political, and economic – affecting demand, pricing, production and delivery processes, operating costs,
any changes in customer trends that affect the scale of demand – growth or contraction of market segments.

To the above factors, a range of archival information and data describing actual past sales results should be added. Analyzing this information and examining correlations between various parameters allows conclusions to be drawn about estimated sales levels in future periods.

Sales forecast results from the company’s business model

The list of potential parameters that influence sales levels is very extensive. However, it should be added that not all of them need to be considered when formulating sales forecasts. Much depends on the business model in which the company operates, its markets, the nature of its offering, and the complexity of the sales process itself. As an analogy, one can cite the weather forecasting process, which from a scientific point of view is more complicated than predicting sales levels. Companies involved in numerical weather modeling constantly analyze even several hundred different meteorological parameters, using powerful supercomputers and continuously improved algorithms. As a result of this work, various types of weather forecasts are generated – different parameters are analyzed for photovoltaic farms, others for wind power plants, still others for heating plants or entities involved in transport and logistics. Similarly, in sales – depending on the company’s operating model, the market, and its segments, appropriate parameters must be selected for analysis, and sales are forecasted based on them. The correct method for developing the forecast must also be chosen.

Sales Forecasting Methods

There are various methods for formulating sales forecasts, chosen based on both the company’s business model and the data available for analysis. The most popular methods include:

Time Series Analysis and Forecasting

Identifies and examines trends in past periods to determine approximate sales values in the future;

Utilization of Leading Indicators

Identifies and analyzes parameters whose fluctuations precede specific changes in sales levels;

Application of Analogy and Inductive Forecasting

The application of analogy involves using data from the sale of similar products or services to forecast the sales of a product or service. The inductive method uses certain regularities occurring in a narrower scope (e.g., within the sales of one product) to forecast sales in a broader context (e.g., for an entire product family);

Heuristic Methods

Based on the experience, knowledge, and intuition of individuals involved in formulating the forecast. This working model is particularly recommended in situations of insufficient or lack of historical hard data, or the inability to detect sales trends. A variant of the heuristic method is the Delphi method, which involves aggregating information from experts involved in shaping the sales forecast. Information is collected in stages using anonymous surveys. After all information is gathered, the position most consistent with the opinions of the majority of respondents is identified;

Econometric Method

Utilizes econometric equations that include forecasted dependent variables and independent variables. These methods are based on adopted assumptions that state both the random factor and the detected relationships between independent and dependent variables are constant over time.

Market Research Method

Based on information obtained from research conducted on groups of potential customers who directly or indirectly indicate their willingness to purchase a specific product or service;

Simulation Method

Uses data obtained through experimental sales, e.g., in periodically opened selected points or through periodically launched online sales channels.

Who should develop sales forecasts?

People with appropriate competencies and knowledge in sales, marketing, and finance should be involved in sales forecasting. It is beneficial when an interdisciplinary team from these three areas works on the forecast, whereas a common mistake is to leave the responsibility for preparing the sales forecast to only one department – e.g., finance.

Sales Forecasting Software

Probably the most commonly used IT tool by companies for developing sales plans is Excel. However, when a forecast requires more sophisticated and advanced analysis of archival data, a Business Intelligence solution, such as Qlik Sense, can be very helpful, enabling multidimensional data analysis and visualization without the need to involve IT specialists. Read more: Qlik Sense: a tool for business analysis and data modeling

Accurate Forecasts = Better Sales

International studies conducted among sales managers showed that only 45% of them are confident in the accuracy and reliability of their sales forecasts. A similar percentage (47%) claim that the main reason for inaccurate forecasts is low data quality or incomplete data (information comes from the Gartner report “State of Sales Operations Survey”, February 2020). These numbers clearly indicate how much companies need data standardization for actual sales results, and at the same time, how helpful IT systems for data analysis and inference can be in utilizing these resources.