Customer churn fully under control
We work with churn models in the retail, banking, services and telecommunications domains (in particular predicting churn of fixed internet connections of a telecom operator).
An effective customer churn management process is a good tool to increase the overall profitability of the company. To achieve this goal, it is necessary to understand the value of the customer, their preferences and motivations, as well as the correct cost of retention.
The churn model is the first model that defines the first set of attributes that then form the data base.
The implementation and final state of the attributes is dependent on the exact state of the current data in the DWH. Factors that will affect the final form are:
- Data availability: existence of data in the data warehouse in the required volume, for a sufficient number of customers.
- Data quality: is the data complete? Are there data gaps for some periods? Does the data have a high number of missing values, incorrect formats, etc.?
- Data reliability: how late is the data updated in DWH?
- History: sufficient history at the required temporal granularity + historicization of changes.
- Consolidation capability: linking of individual data objects – tables using unique keys.