FintechAdvanced Analytics & AI

Prediction and visualization of credit card consumption

The credit department of a Fintech company, brings up the need to encourage consumption through credit cards, to reduce their Churn Rate (dropout rate).

Industry: Fintech
Solution: Advanced Analytics & AI

Book a meeting

Success case: prediction and visualization for credit card consumption in a fintech industry

01 / 03

The challenge

The credit department of one of our clients, a company in the Banking/Fintech industry, brings up the need to encourage consumption through credit cards, to obtain higher income, and achieve better customer retention, reducing the Churn Rate (dropout rate).

They have tried to do this with different strategies, but when trying, the following problems arose:

There was no easy way to track credit card usage. The information was distributed and access to it was difficult. Therefore, the trading strategies were not personalized and were made without taking into account the available information or based on incomplete information.
Even if they had such consolidated information, they found it difficult to identify pattern and trends that would demonstrate the probability of customer Churn.

02 / 03

What Tekne did

Objectives

Therefore, it was necessary to achieve a better use of the data from their systems, and for this they aspired to:

Have their data in an orderly and efficient way to be able to establish priorities, draw conclusions, detect patterns, predict future behavior and build business strategies based on it.
Develop predictive models with the objective of measuring the customer Churn probability, as well as their reactivation.

Proposed solution

We worked in tandem with the company in search of the best strategy to solve its problems in terms of access, visualization of its data and prediction of the behavior of its client portfolio.
After various meetings, we agreed to implement two functional predictive models that, through a standardized input of previously selected variables (Ex: Quantity of Pending Installs, Affinity Group, Purchase Limit, etc) proceeded to store the resulting predictive scores for later analysis and use in the decision-making process of the commercial department of the company.
Dashboards were also developed, with indicators and visualizations agreed upon with the company for a global vision of their credit card consumption, as well as the presentation of the results obtained by the predictive models that were implemented.
Our work continues by advising and collaborating with the company, with the objective of establishing good practices in data management and facilitating the fast implementation of technological tools for data management.

Architecture

Model's Diagram

03 / 03

Results

After wotking with us, the company now has:

A standard application model for the detection of customer Churn (not consumption).
A standard application model for the detection of customer consumption recovery.
Predictive score for the analysis of customer consumption behavior and the establishment of profiles according to behavior.
Predictive score for carrying out campaigns aimed at customers with a higher probability of Churn.
Significant reduction of the Churn Rate.
Dashboards that make it possible to know the situation of the company in relation to the consumption with credit card of its clients and their probability of Churn and reactivation.