Client profiling algorithm design driving commercial strategy

Client Needs & Objectives

  • The objective of this study is to construct a prototype that allows to refine the knowledge of the current client database and to adapt the commercial offer when needed depending on the client profile
  • Depending on the portfolio composition, a decision-making algorithm could be constructed (on the basis of advanced statistical techniques on one hand and on a job validation on the other hand) allowing to target and calibrate the commercial offer depending on the profile of new clients
  • The objective of the prototype was to give an initial overview of the optimization possibilities which could be offered by analyzing data and leveraging on the capabilities of Big Data

Our Approach

1. Definition of the variable to apply:

  • The variable to explain corresponds to the target variable i.e the algorithm output
  • Selection of the explanatory variables
  • These are the input variables needed to get the target variable

2. Descriptive study and correlation matrix:

  • The descriptive study allows to understand the data individually and compare them

3. Algorithm construction by iteration:

  • The iterative process allows to guarantee the optimal calibration of the algorithm for an improved precision

4. Validation of model and benchmark:

  • Statistic validation of the model with the backing of experts makes it a robust model. If possible, the benchmark and backtesting of the model could be tested

Client Benefits & Main Results

Construction of an algorithm which allows to target and calibrate the commercial offer depending on the profile of new clients

Optimisation of the sales process