Business leaders need some type of prediction to make the decisions needed to move their companies forward. For this reason, business forecasting – the part art, part science process of predicting the future for a particular business or market – is an essential task, regardless of a company’s size or market.
What Is Forecasting in Business?
Business forecasting uses data, analytics, and managers’ experience to make predictions that relate to a business’s needs and plans. In other words, a business forecast creates a picture of a future state, usually at a particular level of economic activity, such as for an entire industry, a specific company, or even an individual product. It often includes an element of
financial forecasting, such as projecting revenue and profit.
A business forecast is usually supported by a business analysis, which can be any examination of data or relationships. Typically, business analyses explore potential changes in a business, whether in operating processes, new-product development, or pricing.
To see how business forecasting and business analysis are related, consider a business forecast that predicts customer preference for electric or hybrid vehicles. This might influence a business’s product and marketing initiatives. A possible business analysis that supports that forecast could include accumulated sales data by car model across the industry over the past several years, illuminating the electric vehicle growth rate.
Why Business Forecasting Matters
Business forecasting helps a company plan for the future, both in times of growth and during challenges. It’s important during periods of economic growth to help a company position itself for the potential additional revenue, profit, or market share. It’s also important when times are tough, providing insights that urge a business to take preventive measures and prepare to react more nimbly if certain negative conditions result.
Forecasting methods in business can help extend management’s line of sight, which can better inform both short-term and long-term decision-making. A good business forecast should minimize surprises like a shift in customer demand, a disruption in a supply chain, or even a cash crunch from an unexpected expense.
Business Forecasting: How to Do It
There are two approaches to business forecasting – qualitative and quantitative. Whether you use one, the other, or both depends on the goal of the forecast and the type of available data.
Qualitative business forecasting is a way to develop a view of the future using descriptive qualities or characteristics. It’s more useful in scenarios that aren’t conducive to numerical data trending, like trying to anticipate changes in customer preferences or opinions. Quantitative forecasting analyzes hard data. It applies various statistical methods to historical data to develop projections for the future.
Qualitative forecasts use information gleaned from surveys, narratives, observations, or experience, often either from customers or industry experts and opinion leaders. Since qualitative forecasts tend to be visionary and can be somewhat subjective, they are most useful when a business is facing a new, undefined market opportunity. Focus groups are an example of market research that can help a company get a feel for potential future demand through exploring customer behavior, like buying habits and favorite product features. The Delphi Method is a qualitative approach that uses an iterative question-answer-estimate process to gain consensus about trends and outcomes from thought leaders. Qualitative business forecasts are commonly developed as a first step in business planning before creating quantitative forecasts, although some companies use qualitative forecasting alone.
Quantitative business forecasts predict future metrics from past results using statistical methods. A key characteristic of quantitative forecasts is that they are measurable. Quantitative methods are used in the development of financial forecasts, which are an important part of business forecasting. Financial forecasts translate the overall business forecast into pro forma financial statements that show what the company’s financial performance would be like should the forecast come true, including income statements, cash flow statements, and balance sheets.
Quantitative Methods for Business Forecasting
Quantitative business forecasting can take many forms, depending on the business’s situation and needs, including demand forecasting, startup forecasting, and financial forecasting. Data quality is a primary factor affecting the accuracy of quantitative forecasts, but unforeseen wild cards can cause even the most thorough forecast to become inaccurate. It’s important to frequently compare forecasts with actual results.
Common quantitative methods used in business forecasts fall into two categories:
Time series analysis and projection uses patterns established from historical data to statistically extrapolate future values. Time series methods require significant amounts of historical data and assume that the patterns will continue without change. Straight line and moving averages are two time-series forecasting techniques. Straight line is the simpler of the two, since it merely extends the upward or downward slope of a straight line, as if on a graph, to determine forecast values.
For example, if product sales have increased by 4% each year, the straight line method would simply forecast another 4% increase in successive periods.
Moving average is slightly more complex: a constant recalculation of the average rates of change over time. It also weighs recent data more heavily than older data. Moving averages are typically calculated using shorter periods and therefore better reflect short-term changes, such as seasonal sales.
Causal methods predict outcomes by understanding the relationships between independent and dependent variables. It’s the most sophisticated type of forecasting model and can incorporate external information, such as economic, competitive, and socioeconomic data.
Regression analysis is an example of causal analysis. In regression analysis, you change one variable and then predict changes in a second variable, using a formula based on the second’s relationship to the first. That relationship is established by historical data.
Multiple regression analysis takes this further, using the historical relationship of multiple independent variables to calculate a forecasted dependent variable. For example, a record company might collect data on the number of times a song is played on the radio in a day (an independent variable) and the number of times customers stream that song on one of the streaming platforms (dependent variable). Over time, a relationship can be calculated between the two. The relationship can be used to predict future streams for various different levels of radio play. If a second independent variable is introduced – say, number of public appearances by the band – a multiple regression analysis can be run to predict the anticipated number of customer streams for a particular song when radio play and public appearances are at different levels.
A good business forecast should minimize surprises like a shift in customer demand, a disruption in a supply chain,or even a cash crunch from an unexpected expense.
Which Business Forecast Method Is Best?
When embarking on a forecast, it’s important to select the method – qualitative, time series, or causal – that best fits the business’s scenario.
Determine the best method by asking the following questions:
- How much data is available?
- What is the nature of the entity being forecasted?
- How will the forecast be used?
- How relevant is past data?
- What’s the budget for data collection and labor costs to do the forecast?
- How quickly does the forecast need to be completed?
- What is the tolerable error level, given the risks they introduce?
To illustrate how these questions help identify the best forecasting method, consider the following scenarios:
Scenario 1: A company is creating a sales forecast for a new, innovative product (answering question 3) in an emerging market. The answers to questions 1, 2, and 4 might suggest using qualitative methods, since there would be little data for an untested product in an undeveloped market where any past data might not be relevant.
Scenario 2: A large consumer product company is developing a sales forecast for a new product, which is a brand extension of a well-established product. They might answer those three questions in the opposite way, since they have significant data for a well-known type of product that is relevant to the new product. These answers suggest using the causal method to develop a forecast.
The level of investment needed to launch the brand extension might help the consumer product company determine the answers to questions 5 and 7. If the investment to bring the brand extension to market is high, the tolerable error for the sales forecast might be low. Such a case likely calls for a causal approach, since it's less subjective than a qualitative one. Further, the causal approach might justify a higher budget for data collection and labor.
In contrast, if the level of investment was minimal, or if competitive pressure required a fast time to market (question 6), then qualitative methods might work.
Business forecasting and analysis provide company leaders with a crucial context for decision-making to better lead their companies into the future. Understanding the different business forecasting types and techniques can help ensure that the resulting forecast is the best fit for the situation at hand.
Photo: Getty Images