The Client
A dairy producer creates and ships dairy to a number of retailers nation-wide. A key aspect of their business is coordinating supply production with demand. To do this they perform a monthly sales forecast for the dairy sales for the next month. However, forecasts from the dairy producer’s in-house team were only ever able to achieve an accuracy of 83%, and even this varied month to month. Even a small percentage of error could mean losses of millions in a several billion dollar industry.
The Solution
Difficulties resulting from variations in sales from month to month and the sheer scale of sales data involved led to major issues in trying to create an accurate forecast model using standard AI methods. OneClick.ai’s unique meta-learning approach to the model making process was able to learn from these variations in data and then build a customized forecast model designed to take them, and all relevant sales data, into account. This specialized approach was able to create a forecast model that achieved 94% accuracy, an increase of 11%
The Benefits
Knowing predicted sales month by month will allow the dairy producer to modify their supply to the immediate demands as they change. A task that requires significant planning when your producers are cows. This will prevent any wasted product, cut costs, and allow for greater sales opportunities. As a 10 billion dollar business, an 11% difference in error is a potential 1 billion loss or gain in unsold product or potential sales.
Related Readings
For a step-by-step guide, please read Forecasting.
For other applications of forecasting:
OneClick.ai 