Table of Contents |
In an earlier tutorial, we learned about test marketing, a process for determining how well a product will do based on a limited release. For such an analysis to be successful, the company must already have a production process in place. In the example we used of a new breakfast cereal, the company will already have production of existing cereals, so not much is needed in terms of setting up machines or training workers to put them into production. For a new product or small company, this is not possible.
There is a more significant investment in a completely new product and less ability to forecast demand. Yet demand forecasting is crucial to every aspect of operations, including the facilities layout and design. In this tutorial, we’ll presume that the product design process has been followed and you can identify the audience for the product and the needs the new product meets.
Companies use demand forecasting to help them meet goals, engage in effective financial planning, help with expansion and growth, and can assist in understanding how many employees are needed, and in what roles.
Forecasting future demand is a crucial function for businesses of all sizes. By accurately predicting sales, companies can optimize inventory levels, resource allocation, and production planning. There are two main categories of forecasting methods: passive and active. Passive methods rely solely on historical data, while active methods incorporate external factors like marketing campaigns and economic trends.
The first type of forecasting method is passive forecasting. This method is more frequently used for companies that have been in business for a long time. The method uses a “set it and forget it” approach, meaning that historical data is collected, and projections are based on trends and patterns that have occurred over time.
Passive forecasting techniques are well-suited for products with established sales patterns or for stable market conditions. Passive methods rely solely on historical data, assuming past trends will continue. These methods are well-suited for stable markets with predictable demand patterns. While this approach is useful for companies that have relatively stable sales, for fast-growing companies or for markets that experience a lot of change, an active forecasting method should be used.
Three common examples of passive methods are:
| Method | Description | Example |
|---|---|---|
| Naive Forecast Method | This method assumes future demand will be identical to the most recent period's demand. It's a simple approach, but its accuracy diminishes as the forecast horizon extends. | An ice cream company predicts September sales to be the same amount as August sales based on the naive forecast. Of course, the demand will drop as the summer season ends. |
| Moving Average Technique | Similar to the inventory management technique covered earlier, this technique averages demand data over a specific period (e.g., weekly, monthly) to smooth out fluctuations and provide a more stable forecast. | A clothing store uses use a 3-month moving average to predict upcoming demand, considering the average sales from the previous 3 months. |
| Exponential Smoothing Method | This method assigns weights to past data points, with more recent data receiving greater weight. This approach captures trends and seasonality better than the moving average. | An online retailer might use exponential smoothing to forecast holiday season sales, giving more weight to recent holiday sales data for a more accurate prediction. |
An active forecasting method is the opposite of passive and uses statistical techniques, data analysis, AI, and other computer-generated methods to forecast demand. These methods consider historical sales and other data, like industry growth, to predict demand. This method is used with companies in fast-changing and/or high-growth industries. Active forecasting incorporates additional factors beyond historical data. This approach is ideal for dynamic markets or new businesses with limited historical information.
| Method | Description | Example |
|---|---|---|
| Market Research and Customer Surveys | Both research and surveys provide a path to understanding customer preferences and market trends, allowing for incorporating these insights into forecasts. | A clothing retailer surveys customers about interest in a new product line, informing sales projections for the upcoming season. |
| Sales Force Composite | This leverages the expertise of the sales force and is another active technique. Sales representatives with in-depth customer knowledge can provide valuable forecasts for their specific territories. This method is particularly useful for businesses with geographically dispersed customer bases. | An automaker relies on sales associates to predict the popularity of a new model with their customers. |
| Econometric Models | These models consider the broader economic landscape, including factors like inflation, interest rates, and competitor activity. Econometric models can forecast future demand fluctuations caused by external economic forces by analyzing the relationships between these factors and historical sales data. | A manufacturer of home appliances uses econometric models to predict a rise in demand for energy-efficient appliances due to increasing electricity costs. |
In addition to these methods, companies will forecast short-term and long-term demand. Think of long-term demand as a roadmap to assist with short term forecasting, but obviously, a lot can change in the world that requires the long-term forecasts to be revisited. For example, during the Covid-19 pandemic, companies needed to quickly pivot and change forecasts due to the dynamic nature of that situation. Short term forecasts are focused in one quarter to one year, while long-term forecasts are focused on a year or more. Both methods are important to use, however, especially when focusing on facilities layout and designs.
Companies also engage in internal- and external-level forecasting. Internal forecasting refers to the business operations, such as cash flow, and manufacturing strategy, while external looks at broad market trends.
While you may have identified your final customers, in many cases, the primary customer for a manufacturer is not the end user, but other businesses—stores or distribution centers.
EXAMPLE
Gordon knows that he has two sets of customers. His end user is the cyclist, but his actual sales are to bicycle shops.These businesses will be the best source of information about demand. You may also find good information from similar industries. Though some businesspeople are reluctant to share proprietary information, such as sales volume, others are willing to help start new businesses or launch new products.
Talking to people in your prospective industry can be especially helpful if your proposed product is a service. Say, for example, that you plan to open a pizza parlor. If you visited a few local restaurants and asked owners how many customers they served every day, you’d probably learn enough to estimate the number of pizzas that you’d serve during your first year. Even if the owners weren’t cooperative, you could just hang out and make an informal count of the customers.
You can also learn a lot by talking with potential customers. Ask them how often they buy products similar to the one you want to launch. Where do they buy them and in what quantity? What factors affect demand for them? If you were contemplating a frozen yogurt store in Michigan, it wouldn’t hurt to ask customers coming out of a bakery whether they’d buy frozen yogurt in the winter.
To get some idea of the total market for products like the one you want to launch, you might begin by examining pertinent industry research. For almost any industry, there are associations with websites and newsletters available to the public at no cost.
EXAMPLE
Gordon would likely come across websites for fans of electric bicycles, or articles about trends in e-bikes on sites for cyclists.If you still don’t get a useful answer, try contacting organizations that sell industry data. American Sports Data, for instance, provides demographic information on several fitness activities.
Besides the methods already discussed, companies will gain additional data from both the internal perspective and the external perspective. Examples of internal data for forecasting include:
Gordon is building a new facility to manufacture his bicycles. As he considers demand, he might look at competitors’ sales data to see what patterns exist. Suppose he finds demand is higher in the Southwest part of the United States during fall and spring season, but in summer, demand is higher in the Northwest region of the United States. This data can be valuable in assessing how many bikes should be made. Gordon would also perform other market research to examine the trends associated with e-bikes to help him forecast demand. Another data point he might consider is rising oil prices, which motivates consumers to purchase e-bikes. Given he is uncertain about his new business and actual demand, as historical monthly data is gathered, he will be able to more accurately forecast demand with new data.
Source: This tutorial has been adapted from Saylor Academy and NSCC “Operations Management”. Access for free at https://pressbooks.nscc.ca/operationsmanagement2/. License: Creative Commons Attribution 4.0 International.
