What are the different demand forecasting techniques?

Introduction

Demand forecasting is the process of predicting future sales by using historical data to make informed business decisions about everything from inventory planning, and warehousing needs to running promotions and meeting customer expectations. Demand forecasting helps the business estimate the total sales and revenue for a future period of time. The predictions depend upon the past sales pattern and the continuing trend and user behavior in the present. Subsequently, it is not just simply estimating the future demands but combining several techniques for estimating the demand scientifically and objectively.

Before we continue to various techniques to forecast demand, it is necessary to know about the types of forecasts an organization performs.

Types of forecasting

There are several demand forecasting techniques, based on techniques, your forecast may differ. Demand Forecasting can majorly classified based on the level of detailing, the time period considered and the scope of the market considered. Most organizations do multiple demand forecasts to get a better picture out of it.

Here are the major types of forecasting:

1. Active Demand Forecasting

   An Active demand forecasting model takes into consideration market campaigns, market analysis, and research and market expansion plans. This type of forecasting is done for scaling and diversifying business with aggressive growth plans such as startups. As startups have less historical data so they have to make estimations based on external data.

2. Passive Demand Forecasting

   Passive demand forecasting is the simplest way to make predictions using historical data. This type of forecasting is mainly used by well-established organizations with an ample amount of historical data and market experience.

3. Short Term Demand Forecasting

   Short-term demand forecasting will make predictions for the next three to twelve months of time period. It is used to act quickly to changes in customer demand and market behavior. In this type of forecasting, real-time sales data is used to manage just in time inventory.

4. Long Term Demand Forecasting

   Long term demand forecasting is carried out for a longer period of time, from 12 months to 24 months in the future (In certain organizations, it is from 36 to 48 months). This type of forecasting helps in making long-term strategies such as marketing, capital investment, and supply-chain management. It basically focuses on the business growth road map.

5. External Macro Forecasting

   External macro forecasting helps to overlook trends in the broader economy. This projection looks at how those trends will affect your goals. An external macro demand forecast also can offer you direction for a way to satisfy those goals.

6. Internal Business Forecasting

   Internal Business Forecasting deals with internal business operations such as business financing, supply-chain management, cash on hand, and personnel. It basically points towards the areas where the organizations need to increase the capacity in order to meet the expansion goals.

Based on the type of forecasting, required by an organization, various forecasting techniques can be used, which we will discuss in the next segment of the blog.

Demand Forecasting Techniques

There are two methods in which demand forecasting can be done i.e (A) Survey Methods and (B) Statistical Methods.

1. Survey Methods

   1. Market Research

      In the market research technique, consumer-specific survey forms are sent out in tabular format to get insights that an organization can’t get from internal sales. It gives better information about the type of customers and demographic data which will help to target future markets. Market Research helps young companies to know their customer base better.

   2. Sale Force Opinion

      SalesForce Opinion method use data from sales groups to forecast demand. Salespeople of an organization are closest to their customer base, hence can generate valuable information on customer needs, behavior, and feedback and can even give information about the competition in the markets.

   3. Delphi Method

      In Delphi Method, an organization hires a group of external experts. Each Expert generates a forecast based on their market knowledge. After this process forecasts are shared among the experts anonymously, hence experts get influenced by each other’s forecasts. Now the experts are asked again to generate a forecast and this process is repeated until all experts reach a near consensus scenario. The process is intended to permit the experts to expand on one another’s information and assessments.

2. Statistical Methods

   1. Trend Projection

      This is the simplest and most common demand forecasting technique which is used by organizations. Trend Projection uses past sales data to project future sales. This technique can be used by organizations with a sufficient amount of past sales data(typically more than 18 to 24 months). The data is arranged in chronological order to form a time series, time series depicts the past trends based on which future market trends can be predicted.

   2. Barometric Forecasting Technique

      In Barometric Technique, demand is forecasted based on the basis of past events or the events occurring in the present. It is done by analyzing statistical and economic indicators such as saving, investment, and income. This method can be implemented even in the absence of past data. For example, suppose the government plans for a large housing project, this indicates that there would be high demand for construction materials in the future.

   3. Econometric Forecasting Technique

      This technique combines past sales data with the factors that influence the demand to create a mathematical formula to predict future demand. It finds the relation between the dependent variable and the independent variables. If only one factor affects the demand it is known as a single variable demand function or simple regression. Whereas if there are multiple factors affecting the demand, it is known as multiple variable demand function or multiple regression.

      Regression Equation : Y = a + bX , Y is the forecasted demand.

Conclusion

Demand forecasting helps organizations to make smart business decisions. Based on the business requirements, sales data, market research, and economic factors different demand forecasting techniques can be used. It is often an iterative, highly detailed, and expertise-driven process.

Written by

Devansh Gupta

Data Scientist