Given a large set of customer attributes, the task in the challenge was to estimate the following target variables:

• Churn probability: This is the likelihood that a customer will switch providers. The churn rate is also known as the attrition rate or the participant turnover rate, and is a measure used to find the number of individuals, objects, terms, or items moving into or out of a given collection, over a given time period. The term is heavily used in industries that are driven by customers and use subscriber-based models; for example, the cell phone industry and cable TV operators.
• Appetency probability: This is the propensity to buy a service or product.
• Upselling probability: This is the likelihood that a customer will buy an add-on or upgrade. Upselling implies selling something in addition to what the customer is already using. Consider it like the value-added services that are provided by most cell phone operators. Using sales techniques, salesmen try to make customers opt for value-added services, which will bring more revenue. Many times, customers are not aware of other options, and the salesmen convince them to use or consider those options.

The challenge was to beat the in-house system developed by Orange Labs. This was an opportunity for the participants to prove that they could handle a large database, including heterogeneous, noisy data, and unbalanced class distributions.