Task

One of the main channels for selling loan products, which ensures bank’s profitability, is the call center. We needed to analyze and point out the factors that affect incoming calls, in order to determine which communication channels should be used to increase the incoming call flow and how to maximize the return on each hryvnia invested.

Solution

The work process consisted of 3 stages: problem analysis, modeling and data analysis, drafting recommendations and their implementation.

First stage was dedicated to collecting and systematizing all available data on the client and its partners, comparing them with business indicators of previous advertising campaigns, as well as socio-economic factors of country's development, open data on consumer confidence in Ukraine and loan products usage dynamics.

Thus, we applied the methods of mathematical analysis and forecasting to the key metric "Incoming calls to the call center". We also analyzed parameters that affected conversion from media activity to calls, to applications and to sales consequently, and added them to the model.

Based on this approach, we found the solution and created a submodel for daily and weekly business tasks monitoring. 

Result

Through our ongoing systematic work with data, we have been able to achieve such results and determine the impact of each factor by developing recommendations for maximizing the use of media activity.

Using machine learning technology has affected the 2018 advertising campaign results:

·  We pointed out factors that affected the calls and managed the main ones.

·  We figured out the most effective communication channels mix to increase an incoming call flow to call center.

During the very first phase of the campaign, we received an 8% higher response to each TV rating, compared to the previous advertising campaign, while the overall growth potential after following the recommendation was 58%.  This means that no matter the amount of investment, we can receive a higher response, as well as the required numbers at a lower budget, in the future.