StatConsulting as a company was ranked in the 5th place (8th in team ranking) in the largest international Data Mining contest - KDD Cup 2009, which took place in April 2009.

The 2009 contest was devoted to Customer Relationship Management (CRM) in the telecom sector. The task was to generate three effective prognoses estimating the customers’ propensity to switch providers (Churn), buy a new service or product (Appetency) and buy add-ons proposed to them to make the sale more profitable (Up-selling). The data for the analysis was provided by the French cellular phone operator Orange.

As usual the contest generated a lot of attention among people interested in Data Mining with either business or academic background. There were 465 registered participants from the whole world, who submitted 4,291 valid solutions.

Distinctive for this contest is the FAST TRACK path. It constituted the biggest challenge, because to deliver the results a relatively large set of data (15 thousand columns and 100 thousand rows) had to be analyzed in a short period of time (3 predictions in 5 days from the moment the data was made available). Participation in KDD Cup 2009 was a test not only of the participants’ skills, but also of the effectiveness of the employed analytical tools. To carry out the contest tasks StatConsulting used its own AdvancedMiner system developed by the company for many years. AdvancedMiner is a cutting-edge analytical tool supporting full range of tasks necessary for the construction of Data Mining models and data analysis.

The solution developed by the StatConsulting team was ranked very high in the contest. The difference between the results submitted by our team and the top-ranked solution by IBM Research was merely 1%. We owe this success to the knowledge and experience gained in numerous analytical projects as well as to the effectiveness of the AdvancedMiner system.