Demand Forecasting in Supply Chain, Simple Moving Average, Weighted Moving Average, Exponential Smoothening Method

Demand Forecasting is a critical process in supply chain management (SCM), as it involves predicting future customer demand to optimize production, inventory, and distribution activities. Accurate demand forecasts enable businesses to meet customer demands on time while minimizing the risks of stockouts or overstocking, both of which can incur significant costs. Demand forecasting is not only vital for inventory management but also plays a central role in capacity planning, budgeting, procurement, and logistics management within the supply chain.

Effective forecasting ensures:

• Inventory Optimization:

Helps maintain optimal inventory levels, preventing excess stock or stock shortages.

• Better Customer Satisfaction:

Ensures that customer demands are met promptly and consistently.

• Improved Resource Utilization:

Helps allocate resources (labor, machinery, materials) efficiently to avoid bottlenecks or idle resources.

• Cost Efficiency:

Reduces carrying and ordering costs, leading to better financial performance.

• Strategic Decision Making:

Provides data for long-term decision-making in production and market expansion strategies.

To perform demand forecasting, several statistical methods are used, depending on the complexity of the data and the type of forecast required. Here are three common forecasting methods:

1. Simple Moving Average (SMA) Method

Simple Moving Average method calculates the forecast by averaging the demand over a fixed number of past periods. Each past period has equal importance in the calculation. It is best used when there is a relatively stable demand without significant trends or seasonal variations.

Advantages:

• Simple to calculate and implement.
• Smooths out short-term fluctuations.

Limitations:

• Does not account for trends or seasonality.
• Assigns equal weight to all past observations, even if some are less relevant than others.

Example:

If the demand over the last 3 months is 100, 120, and 130 units, the simple moving average forecast for the next period would be:

Ft = [100 + 120 + 130] / 3 = 116.67 units

2. Weighted Moving Average (WMA) Method

Weighted Moving Average method is similar to the simple moving average but assigns different weights to each past period, giving more importance to recent periods. This approach is beneficial when the most recent data is more indicative of future demand trends.

Advantages:

• More flexible as it can place more emphasis on recent data.
• Can be adjusted to account for trends or other variations.

Limitations:

• Choosing appropriate weights can be subjective.
• More complex than the simple moving average.

Example:

If the past three months’ demands are 100, 120, and 130 units, and we assign weights of 0.5, 0.3, and 0.2 to these periods, respectively, the weighted moving average forecast is:

Ft = (0.5*130) + (0.3*120) + (0.2*100) = 65 + 36 + 20 = 121 units

3. Exponential Smoothing Method

Exponential Smoothing is a popular forecasting method that applies more weight to recent data but decreases the influence of older data exponentially. It is used to forecast time series data with or without trends and seasonality.

Advantages:

• Simple and adaptable to different types of data (with or without trends/seasonality).
• Places more emphasis on recent observations without ignoring older ones entirely.

Limitations:

• The selection of the smoothing constant α\alphaα is crucial for accurate forecasting.
• Can lag behind when there are significant changes in the demand pattern.

Example:

If last month’s forecast was 110 units and actual demand was 120 units, and α = 0.4, the next forecast is:

Ft = (0.4*120) + (0.6*110) = 48 + 66 = 114 units

Choosing the Right Forecasting Method:

• Simple Moving Average is best for stable demand with no significant trends.
• Weighted Moving Average is ideal when recent data is more relevant and there is some volatility in demand.
• Exponential Smoothing is a versatile method that can handle both stable and fluctuating demand but requires careful selection of the smoothing constant.