Learn How to Achieve New Levels of Accuracy with Legion Demand Forecasting

[Thomas Joseph](/content/blog/author/tjoseph/ "View Thomas Joseph’s posts"/index.html) is the Head of Data Science at Legion. He is an expert in Data Science and in designing and deploying large-scale enterprise systems. Thomas was previously VP, Office of CTO at SAP, CTO at TIBCO, and holds a Ph.D. in CS from Cornell.

Our advanced machine learning (ML) based demand forecasting engine is a key pillar and a foundational element of Legion’s Workforce Management (WFM) platform. Legion’s ML-based demand forecasting has several advantages over other approaches:

• Learns more subtle patterns
• Incorporates different types of data
• Scales and produces consistent and accurate forecasts week over week for hundreds of thousands of data sets.

There are a lot of claims around accuracy, but what exactly is an accurate forecast? And why does the way you report and measure it matter? It’s easy to measure accuracy – you get a forecast, wait for the actuals to come in, then compute the difference, and perhaps you express the difference as a percentage. That’s the easy part.

But, it leaves many open questions:

• Over what time period should you measure?
• If you have many data sets, how do you aggregate the results?
• And most importantly, if you have a number representing accuracy, how do you know if it’s good or bad?

Read our latest white paper, Achieve New Levels of Accuracy with Legion Demand Forecasting, which provides insights about demand forecasting approaches and best practices. It describes how Legion WFM applies answers to the questions above and how we measure and report the accuracy of our demand forecasts.

You can download the full white paper now.