WMAPE: what it is and how to use it in your demand forecasting

WMAPE is one of the key metrics to use when measuring the accuracy of demand forecasting. In this article, we will focus on its use in supply chain demand planning. Learn how to use and calculate WMAPE and why it is important to pay attention to it for your month-to-month operations and planning.

What is WMAPE?

It is a statistical metric that helps measure how accurate a forecast has been. Unlike its sister metric, MAPE, WMAPE weights each error according to the product's sales volume, giving higher importance to top-selling items.

This means that an error in a flagship product that sells 10,000 units carries much more weight than the same percentage error in a product that sells 10 units.

WMAPE Formula

The formula works like this:

1. Top (numerator): you sum the absolute differences between the actuals and your forecasts (the error for each product/period).
2. Bottom (denominator): sum of actual values.
3. You divide one by the other, which gives you a percentage of error.
   

How is WMAPE calculated?

For each product, Datup follows this process:

1. It calculates the absolute difference between what was actually sold (the target or actual demand) and what was forecasted. Absolute means that it doesn't matter if it was over or under; only the magnitude of the error matters.
2. It converts that difference into a percentage by dividing it by the actual sales of the product.
3. It weights that percentage by the actual sales volume of each product. This way, a product that accounts for 30% of total sales contributes proportionally more to the result than one that accounts for 0.5%.
4. The result is expressed as a percentage between 0% and 100%.

WMAPE calculation example

• Sum of errors: 10 + 5 + 50 = 65
• Sum of actuals: 100 + 10 + 1000 = 1110
• WMAPE = 65 / 1110 = 5.9%

Why is it better to use WMAPE instead of MAPE for demand forecasting?

MAPE calculates the error percentage product by product and then takes a simple average. This means a small product (like B, which sells 10 units) has the same weight as a large one (like C, which sells 1000). If product B has a high percentage error, it distorts the total average even though it barely matters to the business.

WMAPE, on the other hand, weights by volume: products that sell more have a greater impact on the final result because you are summing everything together before dividing. That is why it is widely used in demand/sales forecasting, where getting it right for high-revenue products is more important than for those that contribute less.

How do you interpret WMAPE?

On the Datup platform, forecast accuracy is calculated as the complement of WMAPE. In other words: Accuracy = 100% − WMAPE. Thus, a WMAPE of 25% is equivalent to 75% accuracy. Datup's efficiency reports classify this accuracy into three levels: high (greater than 90%), medium (between 70% and 90%), and low (less than 70%).

Why is it important to consider this for demand and inventory planning?

WMAPE translates a statistical concept into a language that connects directly with the business. When the report says "forecast accuracy was 82%," it means that for every 100 units of actual demand, the forecast deviated by an average of 18 units. This allows you to quickly calculate the impact on inventory, service levels, and budget.

Furthermore, Datup calculates this metric not only for the suggested forecast from the demand planning software, but also for collaborative forecasting (the one adjusted by sales and planning teams).

Discover more metrics to improve your supply chain operations:

• GMROI (Gross Margin Return on Investment)
• Cash conversion cycle
• Supply chain KPIs

‍