Evaluating vehicles through machine learning: „Phoenix Pricing“ allows our branches to predict a used vehicle’s price precisely.
The complexity of the used car market has been growing for years and thus also the valuation of vehicles in the used car market. Besides the increasing number of model series, other factors including the selection of vehicle features and large number of suppliers make it difficult for dealers to quickly and accurately determine the value of cars. A key success criterion in the used car business is determining a "good price offer".
To support pricing experts in Mercedes-Benz business as best as possible in determining the offer price, the "Phoenix Pricing" project was developed under the management of the Controlling and IT units of Mercedes-Benz Sales Germany. With the help of machine learning, a proposed price is calculated, which delivers the most accurate prediction of the actual sales price on the basis of diverse data records.
This data can be obvious pricing-relevant attributes such as vehicle age and mileage, as well as special features and sales prices from previous years. The more data sources are available, the more precise the predicted price, which is continuously optimized through constant adjustment by the algorithm and integration of other data sources.