Advanced Demand Forecasting Part 3: Need-to-Know’s When Considering ADF
Advanced Demand Forecasting is the next-level solution for businesses looking to identify and understand which factors have the greatest contribution to demand so that they may invest with greater accuracy. Often times, demand does not have a simple, straightforward relationship with any singular feature which organizations can increase to up their demand. That’s where Advanced Demand Forecasting steps in.
Advanced Demand Forecasting deploys feature engineering to enable organizations to recognize those inter-woven relationships and recommend when and how to properly adjust variables with the greatest impact on demand. Let’s use an example scenario to show one such inter-woven relationship between features and how reality often presents both internal and external influences.
Retail X’s sales total is influenced by a number of features, one of which is Sale’s Associate (SA) hours. Retail X has multiple stores, but we are going to look at this feature on an individual store level instead of whole-company. At one individual store (Store A), our model shows decreasing sales and one main reason is due to the number of SA hours.
Our model also says that at Store A, one SA hour is worth $100 in sales. The incorrect approach to this scenario is to realize that increasing SA hours directly translates to sales and therefore deciding to increase all of the SA hours at Store A. While the model will does indicate a relationship between SA hours and sales, investing more man-hours or employees into a store doesn’t drive foot traffic or sales for a business. The correct way to read this data is to understand that SA hours are based on daily demand, such as peak store volume times, which correlates to wait times.
Using this example, Retail X will be able to conclude that by increasing employees by N amount at Store A during specific times of the day will increase sales. In the end, we are shown SA hours to be related to two other factors and management should not invest across the board to generate sales.
Now that you understand how features will often influence one another and demand is not dependent on any one variable, it should be apparent that having good data is imperative! Demand regularly fluctuates, especially in the retail and consumer goods industries. To forecast as accurately as possible, here are four key things to remember for Advanced Demand Forecasting to have maximum effectiveness!
- How far into the future Advanced Demand Forecasting is able to predict depends on how stable the demand is. Businesses with higher fluctuations in their demand will produce shorter forecast windows than demand that behaves more regularly year-over-year.
- Predictions are only as good as the data it learns from. Advanced Demand Forecasting relies heavily on internal data, as machine learning requires historical data to develop accurate forecasts.
- Raw data and raw counts are best. Predictions are more accurate when the model can learn from raw or unaggregated data from previous years. Data that is already aggregated will only tell a portion of the whole story or show a percentage of a factor (such as X% of people used this coupon) and not produce meaningful results.
- Daily operational data at the individual store level is necessary. Knowing if promotions are used, how many customers come in, operation hours, staff information, product information, and more all contribute to creating an accurate forecast.
These four points apply to any demand forecasting model, really. So if you have any black boxes or inefficiencies within your demand models, first check to see if the data you’ve been feeding your models are clean and organized. Remember, poor data leads to poor modeling! But if your data is fine, then it is likely your current forecasting solution falls short of your business needs and it is time to consider upgrading!