What in the ad affects how fast a car is sold on Blocket

Sammanfattning: The aim of this thesis is to investigate what in the ad insertion affects how fast a car is sold on Blocket, a Swedish marketplace for second-hand goods. The method, through which the issue was investigated, is the machine learning algorithm random forest. Three models have been created and optimized, each with different datasets. Despite the models’ differences, the results were similar for the three models – their R2 and MAE values indicate that the variables for ad insertion explain little about how long it takes until a car is sold on Blocket. Of the variables in the models, the most important in all three models had to do with the pricing of the car. The results also show that the length of the description, is the second most important variable that the seller can impact herself, of those variables included in this investigation.