An increasing number of companies have begun to forecast demand using predictive modeling. Many of these businesses have mastered the art of gathering and utilizing data from their own business activities in these forecasts. This is good news – after all, the best predictor of future demand is past demand, right?
Yes and no. While it is great that companies are understanding the importance of data and taking the initiative to build predictive models, this approach leaves a big gap in grasping and forecasting customer demand. Using only internal data in predictive models is equivalent to assuming that nothing other than internal business operations will affect consumers’ choices. Given the events of the past year and a half, it should be very clear that this is not the case.
By using external data in addition to internal data, however, it is possible to create a reliable demand forecast that takes all drivers of customer behavior into consideration. Let’s break down why.
Components of an Accurate Demand Forecast
A demand forecast is an assessment of future customer demand created through a predictive analysis of data. You can read this blog post for a more detailed overview about the basics of demand forecasting if you’re new to this concept.
Internal data is the data that you collect from your business. POS sales data, inventory levels, online sales, payroll figures, and customer service feedback are all examples of internal data that will help you determine what future demand may look like, and plan accordingly.
External data covers factors that affect demand but that are not directly related to your operations and that you typically have no influence on. These are factors like weather, economic conditions, holidays, or, like we have witnessed over the past year, a global pandemic. Given the impact some of these factors can have on a businesses’ sales, it is becoming increasingly important to include external data in any demand forecast.
At its core, a predictive model is simply an automated mathematical process running data. But that means that the predictive model can only be as good as the data that powers it; using clean, relevant data is critically important to achieving an accurate demand forecast. Therefore, both internal and external data must be collected regularly, and each applied in the same way and at the same granularity.
External Data’s Role in Demand Forecasting
Additional external data that applies to many businesses includes local disposable income levels, unemployment figures, and industry growth. Although simplified, the following examples illustrate why including data on relevant external factors will take a demand forecast to a whole new level.
If a downtown storefront business with primarily walk-in customers tends to experience more business on sunny, warm days compared to cold, snowy days, placing weather in the demand forecast will help managers to schedule the proper staff required to handle increased customer traffic for the sunny week ahead.
In a college town like Ann Arbor, businesses tend to be quite busy on Saturdays with a home football game. And many businesses in a city with an annual festival will have higher demand during the time of the festival. Knowing local event data and the effect it has on demand will allow businesses to plan not only their staffing, but also their inventory to accommodate these spikes in the future.
As already indicated, local, state, national, and global economic conditions are also important forms of external data. During the COVID-19 pandemic, this became extremely clear. Global production stalled, unemployment rose, and some demand came to a standstill, while other demand skyrocketed. If a store was only relying on their sales data from 2019 to predict demand in 2020, they would have been very far off base. And while the pandemic as a whole wasn’t predictable, once the extent of it became clear, businesses utilizing AI-powered demand prediction were able to pivot and adjust to the new conditions far better and far quicker than their competition.
These are just a few basic examples of the role external data plays in demand forecasting. Today, any company offering AI-powered demand forecasting software should include internal and external data in their prediction models. Incorporating both factors into a model enables the forecast to be sensitive to the business and the state of the world it operates in.
Author: Anna Schultz
Anna is Marketing Coordinator at RXA, a Growth Marketing Intelligence company fueled by data science and applied artificial intelligence, and parent company of Weave Workforce, a workforce optimization software. In her role, Anna helps companies meaningfully activate the benefits of data analysis through RXA’s GMI platform and applications. In today’s world, all companies collect data – it’s what you do with it that makes a difference!
