Demand forecasting is an essential function for virtually every business. While companies have traditionally leaned heavily on historical data to forecast for the future, what happens when that historical data is no longer relevant?

Forecasting challenges created by COVID-19

It’s no secret that the COVID-19 pandemic has had an immense impact on the business world, nor is it expected to affect businesses forever. How it will change, however, remains to be seen. As the World Business Council for Sustainable Development notes about the changing workplace, “the medium and long-term impacts [of COVID-19] are however not all predetermined: much depends on what governments, societies and businesses will do next.”

Much of this uncertainty stems from the fact that the last pandemic, the Spanish Flu, occurred over 100 years ago. We’ve never experienced something like this during the digital era, and there’s no playbook. COVID-19 forced many businesses into new, uncharted territory. Remember how popular the word “unprecedented” was in 2020?

Even now, businesses are performing a balancing act. Everything seems to be in flux.

So, it stands to reason that traditional forecasting methods may not accurately represent the Post-COVID market, let alone the market during the ongoing pandemic. Further, these forecasting methods may remain inflexible, incapable, or otherwise unreliable for demand planning during similar events in the future. A different approach is needed: an approach where historical trends are considered but can be tuned using other relevant factors of the current surrounding market conditions.

How do I achieve more accurate forecasts?

Many organizations already achieve more accurate forecasts than their competitors using traditional methods by leveraging analytics data in their forecasting. In fact, CFO.com reported that 74% of organizations achieved more precise forecasting by leveraging sales planning analytics as far back as 2018.

However, the issue for the Post-COVID world is that while analytics data can help produce more accurate forecasts, the forecasts will likely still rely on the assumption that many of the underlying behavioral and societal trends are the same as before. For example, while analytics data may track the number of interactions a potential customer has with a given item before purchasing, it may not recognize that the customer is working from home and is more likely to order the item online than purchasing it in a brick-and-mortar store.

Situations like those described above may result in accurate forecasts for overall demand but inadvertently result in inefficient supply chain planning (e.g., if stores receive excess stock while warehouses for online orders are understocked). Further, they may fail to account for future unexpected outlier events like COVID-19 because they may be inclined to throw out historical data that doesn’t reflect the new normal or decide to avoid using outlier data in planning after COVID.

As such, organizations will need to do more than rely on sales analytics data for forecasting post-COVID. Leveraging organizational data in conjunction with machine learning and AI, for example, represents powerful tools that can supercharge demand forecasting. By leveraging behavioral, intent, historical, and other data types (such as weather forecasts and social media trends) with AI, organizations can compile and process current and predicted future data to uncover deeper market insights than before. Machine Learning, on the other hand, can enable forecasts to “learn” from results and trend data to provide periodic updates to forecasts based on new findings or sudden shifts. This way, demand forecasts can be continuously improved and fine-tuned as the market shifts with little additional human involvement.

Advanced Demand Forecasting from Neal Analytics

Neal Analytics has a solution for organizations that want to enable more accurate demand forecasting today and prepare for forecasting in the future. Advanced Demand Forecasting from Neal Analytics can leverage all organizational data, including historical, behavioral, intent, social media, and even the predictions of other models (such as weather forecasts) using AI and Machine Learning to deliver a solution that solves for the challenges described above.

Additionally, Advanced Demand Forecasting can automatically respond to new or unexpected shifts in consumer behavior with periodic adjustments to the demand forecast based on these indicators. Best of all, organizations can tailor the periodic updates to satisfy specific organizational requirements. For example, Advanced Demand Forecasting can provide anything from yearly, quarterly, or monthly updates right down to every 30 minutes. It can also leverage historical data, including traditional time series approaches with key indicators, to detect and adjust to drastic market changes caused by future events like the COVID-19 pandemic.

Learn more about Advanced Demand Forecasting

If you’d like to learn more about Advanced Demand Forecasting, read our brochure or contact us.

Additional resources:

• The 4 pillars of an effective and robust forecasting solution
• Advanced Demand Forecasting: A technical overview

This blog was originally published 2/18/2021 and has since been updated.