Creation of sales/demand forecasting model based on machine learning and big data. We have used all possible variables as weather, economic situation, demographics, promotions, competition, price sensitivity) to create model that forecast the demand for products with much better results.
We make sales forecasting model for distribution and auto spare part company.
The difference between classic ERP extrapolation prediction and machine learning forecast
- Classic ERP systems use for forecast extrapolation of past numbers
- This becomes an issue when some variables as weather is significantly changed
- Profio engine adds the variables in the forecasting
- This makes it more accurate (even by 40%)
How it works
- We run the data and variables through statistical modelling each day.
- We then choose the best model for each item.
- The system learns by itself from the past. It improves each day.
This has these implications for the company
- Increase of Service Level up to 99%
- Elimination of the stock-outs
- Inventory decrease by 25%
- They increase the sales by increasing the availability of goods on shelves for clients
- Big Data variables
- Neuro Networks
- Machine learning
- Statistical Modeling
Other possible clients
- The ideal client is a company with more than 700-1000 items in the warehouse.
- Spare part distribution
- Retail stores chains (food, drugstores, pharmacies)
- Pharmacy distribution
- Food production
- Generally distribution companies
McKinsey predicts machine learning will reduce supply chain forecasting errors by 50% and reduce lost sales by 65% with better product availability.