Engineering

Update on the State of Demand Forecasting: Navigating a Post-COVID World

Dirk-Jan van Veen
March 31, 2023

Introduction

Demand forecasting has always been a challenging task, but it has become even more complicated as we emerge from the COVID-19 pandemic. The demand patterns we have seen over the past three years, particularly in the home entertainment category, are now drastically different. As a result, both traditional and AI-based time-series models are currently struggling to provide accurate forecasts. However, we expect these models to regain their effectiveness once the world stabilizes and large-scale global disruptions become less frequent.

The Current State of Demand Forecasting: Agility and Adaptability

In the current climate, many companies are turning to real-time data and making adjustments on the fly. No one's demand forecasts are proving to be very accurate, which is understandable given the uncertainty and rapidly changing consumer behavior. In this context, businesses need to be agile and adaptable to stay ahead of the curve.

The Role of AI in Forecasting: A Step Beyond Traditional Algorithms

Based on experience, AI offers only marginal improvements over traditional algorithms, such as the rolling average, when it is based solely on the same time-series data. However, the true potential of AI lies in its ability to process and analyze large amounts of diverse data sources to provide more accurate and robust demand forecasts.

The Future of Demand Forecasting: Incorporating External Data for Unprecedented Accuracy

The key to unlocking the full potential of AI for demand forecasting lies in incorporating external data sources. These can include:

  1. Promotions: Data on promotional activities and their impact on sales can help AI models better understand the relationship between promotions and demand.
  2. Holidays: Factoring in holidays and their influence on consumer behavior can lead to more accurate demand forecasts.
  3. Other sales datasets with similar SKUs: Analyzing sales data from comparable products can provide valuable insights into consumer preferences and trends.
  4. News and social media trends: Keeping track of relevant news and social media trends can help businesses anticipate sudden changes in demand due to events or popular trends.
  5. Economic indicators: Incorporating macroeconomic data can help AI models account for changes in the broader economic landscape, which may affect consumer spending and demand.
  6. Weather data: Weather conditions can have a significant impact on consumer behavior and product demand, making it an important factor to consider in demand forecasting.

Demand Forecasting and Slotting Optimization

A good demand forecast is essential for slotting optimization (what we specialize in at Syncware). We will be more than happy to assist you in achieving more accurate and reliable forecasts for your business. Additionally, by combining AI-driven demand forecasting with slotting optimization, you can not only improve the accuracy of your forecasts but also enhance warehouse efficiency and your operational costs. Get in touch through the form below.

Share this post

Ask Us Anything

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.