What Is Demand Forecasting?
Demand forecasts estimate future demand for products and services, which helps inform business decisions. Demand forecasts include granular data, historical sales data, questionnaires, and more. You can improve their accuracy by incorporating real-time information, advanced analytics, machine learning, and data science.
Accurate demand forecasts help you with inventory management, capacity planning, product demand, and resource allocation. Accurate forecasts ensure you can order appropriate SKUs, stock enough products, and set reasonable pricing to meet customer demand without facing supply shortages.
This guide will provide an overview of the different types of demand forecasting techniques and the advantages and disadvantages of each method.
Demand Forecasting Methods, Models, and Techniques
There are three main types of demand forecasting processes:
1. Qualitative and Quantitative
2. Time-Series Analysis/Projections
3. Causal Models
Each leverages different data and has varying degrees of demand forecast accuracy and forecast error rates.
Qualitative forecasting is a type of demand forecasting that relies on subjective factors such as customer opinions and market trends. You would often use it when little or no historical data was available.
Qualitative forecasting is less accurate than quantitative forecasting but often faster and cheaper to implement. Qualitative forecasts combine, or aggregate, as much data and information as possible to produce unbiased estimates.
Qualitative Forecasting Pros and Cons
• Can be used for new products
• Focuses on customer sentiment and perceived needs
• Quick to develop and produce results
• Subjective nature may lead to inaccurate results
• Difficult to compare to historical data or other forecast methods
• Time-consuming to collect data from consumers
Qualitative Methods and Forecasting Models
The Delphi method is a structured technique for gathering expert opinions or estimates. It is mainly used in forecasting and has been shown to produce more accurate results than individual estimates, especially when there are considerable disagreements among experts. The basic principle of the Delphi method is to avoid biased estimations of probability and value by aggregating the judgments of a group of experts.
Market research uses a formal approach to forecast market demand. It's used to test hypotheses about future demand, which is helpful for emerging and new markets. These forecasts often use case studies, research, focus groups, customer experience reports, and more.
Panel consensus is a process used to gather data and opinions from a group of experts. You could use it to generate ideas, identify trends, and make decisions. Panelists are chosen based on their knowledge and expertise in the subject matter. The panel typically meets to discuss, then votes on the issue. The consensus is then used to make a decision or recommendation.
The visionary forecast technique is used to generate ideas about the future. This technique involves a group of experts who share their visions of the future. The visions are then analyzed and compiled into a report, which is used to make decisions about the future.
The historical analogy technique is used to generate ideas about the future. This technique involves looking at past events and trends to identify patterns that may occur in the future.
Quantitative forecasting is often more accurate than qualitative forecasting, but it can be more expensive and time-consuming to implement. It uses historical data, statistics, and other objective metrics. This type of forecasting is often used for inventory planning, short-term and long-term sales forecasting, and supply chain management optimization.
Quantitative Forecasting Pros and Cons
• More objective and unbiased
• Able to detect patterns and trends
• Better for long-term demand forecasting
• More expensive to produce when using manual processes
• Difficult to account for external factors and newer information
• Can miss out on big picture ideas and trends
Quantitative Methods and Forecasting Models
Naive forecast methods use past data to project future demand. Because of that, you won't be able to account for new demand influences, such as trends and events.
Seasonal forecasting uses seasonal historical data, such as actual sales, to forecast future seasonal demand. Seasonality forecasts can include event, monthly, quarterly, and more projections. Sometimes the data can be skewed by major economic factors, such as downturns, increased consumer spending, or SKU availability.
Revenue Run Rate
Revenue run rates forecast end-of-year results, such as projecting profit margins, but can be inaccurate during times of volatility. These forecasts are based on past performance and don't include current trends, patterns, or other economic fluctuations.
Historical Growth Rate
Historical growth rates help you measure the growth rate in a particular metric by measuring its historical data. This data can help you project future growth if you expect performance to remain consistent. Volatility and trends can affect forecast accuracy.
Linear regression is a detailed process that uses a set of data points to identify a cause-and-effect relationship. It does this by comparing how a specific variable affects an outcome. For example, sales calls compared to sales conversions. Once you establish a relationship between the data points, you can use it to forecast expected outcomes. However, this method requires that you use a variable that significantly affects the outcome. As the saying goes, correlation doesn't prove causation.
Time-series analyses use data from many years to predict potential demand. However, the data must produce precise, reliable, and stable relationships and trends.
The trends gathered from past data are used to assess whether current trends in consumer demand are in line, behind, or ahead of where you might expect them to be.
Other factors can affect the accuracy of time series analysis, including economic disruption and seasonal variations. However, you can use additional statistical techniques to account for these variations in data and analysis.
Time Series Analysis Pros and Cons
• Relies on past data to establish trend
• Can identify long-term relationships in data
• Generally less expensive than other methods
• Data needs to be of high quality
• Data must follow clear and consistent patterns
• Only works if the future closely resembles the past
Time Series Analysis Methods and Forecasting Models
The moving average technique helps you identify trends in data. The process involves taking a series of data points and calculating the average, which is then plotted on a graph. The direction of the moving average identifies the trend.
The exponential smoothing technique uses a similar process to moving average but assumes that the most recent data point is the best estimate of the current trend. This technique assigns exponentially decreasing weights to data points as they get older. The weight assigned to a particular data point depends on the value of a parameter. You can use the exponential smoothing technique with or without seasonality.
The Box-Jenkins technique uses a similar process to the exponential smoothing technique but instead assigns errors, or an error rate, to older data points.
The X-11 technique uses a similar process to exponential smoothing and Box-Jenkins but instead uses multiple parameters, including seasonality, trends, and irregularity. This technique works well with medium-range forecasting.
This trend projection technique identifies future trends in a given dataset by extrapolating past data. Trend projection can help predict future demand for a product or service, future sales figures, or any other numerical data.
Causal models are the most refined of the forecasting tools. They're best suited for long-term forecasts. Causal models require you to meticulously analyze historical data until you can identify clear relationships between two data points or factors.
Causal models further highlight causal relationships within the data and can incorporate information such as other analyses and location data. Because of this, you will most likely continue to update causal models as new information is received.
Causal Model Pros and Cons
• Can be used for long-term forecasting
• Can identify relationships between data points
• Generally more accurate than other methods
• Data must be of high quality
• Data analysis can be time-consuming
• Requires access to historical data
Causal Methods and Forecasting Models
Regression is a powerful statistical technique you can use to predict future values of a response variable based on known values of predictor variables. In regression analysis, relationships between variables are quantified using a best-fit line (or regression line), which captures the linear relationship between the predictor and response variables.
The econometric model technique is a statistical approach to modeling economic data. You can use it to predict future economic activity, measure the impact of economic policy, and understand the relationships between different variables in the economy. Econometric models are usually estimated using historical data.
Intention-to-Buy + Anticipation Surveys
Intention-to-buy surveys are an essential tool for measuring consumer behavior. Using past sales information, you can make more informed marketing and product decisions by gauging consumers' future purchase intentions. You can also use these types of surveys to identify potential problems or areas of opportunity within the company.
The input-output model is a quantitative economic technique used to estimate the ripple effects of changes in one economic sector on other sectors. The input-output model is based on the transactions between firms within an economy as captured by an input-output table.
Life Cycle Analysis
The life cycle analysis is a tool used to assess a product's or service's environmental impact. The life cycle analysis looks at the entire life cycle of a product or service, from the raw materials used to produce it to the waste produced after disposal. The life cycle analysis can be used to identify the most environmentally friendly way to produce a product or service.
Real-World Examples of Demand Planning and Forecasting
Demand planning and forecasting is a critical part of any business. It allows you to anticipate future retail and ecommerce needs and ensure you have the resources to meet them. In this section, we will discuss some real-world examples of demand planning and forecasting.
Supply chain leaders use demand planning and forecasting to ensure the procurement and delivery of goods. The goal is to maintain inventory levels to meet customer needs without losing money due to surplus supply. Many factors can affect demand needs, especially shifts in economic and consumer behavior, changes in the labor force, natural disasters, global events, and more. The most important thing when forecasting demand is to use the most suitable method for each situation.
Coca-Cola uses forecasting across their organization to create customized reports for each store, predict which flavors to use in products, and anticipate which machine components will require maintenance.
Office of Finance
The office of finance is responsible for an organization's financial planning and forecasting. These responsibilities include developing budgets, preparing financial statements, and managing cash flow.
Chick-fil-A uses forecasting to build risk scores and create reports for senior management.
Retailers use demand forecasting to determine how much product to order from suppliers and how to price their products. Forecasting demand can help retailers avoid stockouts and overstock and manage their inventory costs.
Rent-A-Center uses forecasting to understand customer needs and adjust marketing promotions based on customer profiles. The customer segmentation allows them to optimize merchandise to match local demand instead of using the same assortments in every store.
Machine learning is a type of artificial intelligence sometimes used to increase demand forecasting accuracy. Machine learning algorithms learn from data and improve over time, making them well-suited for forecasting, as they can handle complex data sets and make predictions about future demand.
7-Eleven uses machine learning and AI for demand forecasting to get insight into same-day reporting, promotions, seasonality, and out-of-stock reports. The forecasts help them manage inventory for thousands of products across 9,000+ stores.
Demand Planning and Forecasting Software
There is no one-size-fits-all piece of software for demand planning and forecasting. Organizations often use multiple products to predict demand. Here's a look at a few of them.
Spreadsheets are the most popular type of software used in demand planning and forecasting. Around 75 percent of supply chain departments rely on spreadsheets for decision-making.
Benefits of Spreadsheets for Demand Planning Forecasting
• Can perform regression analysis and time-series analysis
• Allow for minimal coding and scripting
• Can be somewhat automated using other software
Coding Programs and Software
Coding programs, like R and Python, are often used by data scientists for advanced analytics. Coding works for various demand planning and forecasting tasks, including time-series analysis and machine learning.
Benefits of Coding Programs for Demand Planning Forecasting
• Can perform advanced analytics
• Can automate processes
• Can handle large datasets
• Can be customized to specific needs
Visualization software, like Tableau and Qlik, is used to create graphical representations of data. Visualizations can be helpful in demand planning and forecasting by making complex insights and data easier to understand.
Benefits of Visualization Software for Demand Planning Forecasting
• Can make complex data sets easier to understand
• Can be used to create dashboards and reports
• Can be customized to specific needs
BI and Reporting Software
Business intelligence (BI) and reporting software, like SAP Business Objects and Oracle BI, are used to create reports and dashboards. Reports and dashboards can provide deeper visibility into data.
Benefits of BI and Reporting Software for Demand Planning Forecasting
• Can provide visibility into data
• Can be used to create reports and dashboards
• Can be customized to specific needs
Geospatial Analytics Software
Geospatial analytics software, like Esri ArcGIS and MapInfo Professional, analyzes geographical data, providing insights into customer behavior and ideal retail locations.
Benefits of Geospatial Analytics Software for Demand Planning Forecasting
• Can provide insights into customer behavior
• Can be used to create heat maps
• Can be customized to specific needs
Analytics Automation Software"
Analytics automation software, like Alteryx, is used to automate data analysis. Analytics automation reduces the time, effort, and costs of preparing and analyzing data to generate demand forecasts. High-end platforms include features and resources for data prep and blend, analysis, advanced analysis, machine learning, AI, data science, geospatial analysis, data governance, and more.
Benefits of Analytics Automation Software for Demand Planning Forecasting
• Can reduce the time and effort required to analyze data
• Can be used to automate processes
• Can be customized to specific needs
• Quickly scale operations to business needs
Demand forecasting is a process by which a company attempts to estimate future demand for its products or services. There are various methods of forecasting demand, and each has advantages and disadvantages. The most important thing when forecasting demand is to use the best method for each situation.