Recommender System for a Bank

Client Issue

One of the top CIS banks was interested in increasing sales volume of its retail products, raising effectiveness of cross-sales campaigns and reduction of operating costs for marketing trough advanced analytics.


The bank faced the following issues:

  1. Lack of client profile knowledge for retail bank products
  2. Lack of connectivity in information about clients
  3. Lack of access to data for business decision-making
  4. Lack of effective tools for data extraction, mining and analysis


Main stakeholders for this project were Marketing department, Retail business department and Strategy department.

DAI Solution

Deloitte DAI applied a complex approach on achievement of customer’s goals including the following steps:

  1. Problem framing.

    Deloitte Analytics Institute collected expectations, issues and requirements from business departments and concerned employees. Deloitte Analytics Institute specialists conducted interviews with stakeholders (business, marketing, IT), clarified requirements for the system and managed client’s vision.
  2. Data collection and preprocessing.

    On this stage Deloitte Analytics Institute got all necessary CRM and transactional data for Bank clients “as-is” for subsequent data preprocessing and analysis. Dataset included social-demographical data, client’s purchases history, monthly payments and balances history, information about communication with client. This step included data consistency check between different sources and inconsistencies treatment. This step included handling with duplicates and null values, dictionary forming and transformation, statistics calculation. Detected problems and solutions for them were reported to the client.
  3. Data processing and algorithms development.

    Deloitte Analytics Institute started this stage from data augmentation actions, which resulted in mining of changes in spending speed or income growth, different client’s events triggers. To understand client financial behavior, Deloitte Analytics Institute provided a calculation of aggregated balances and income/outcome transfers for different periods. These features were merged with features from CRM system (social-demographical data, client’s purchases history). A number of machine learning algorithms were applied on these features for predicting best recommendations of products to clients. Following the best performing algorithm was selected based on historical backtesting.
  4. A/B testing.

    In order to proof effectiveness of the designed solution, Deloitte Analytics Institute planned an A/B testing experiment. Three groups were for for the experiment: group formed by recommender algorithm, group formed by social-demographic segmentation and the reference group without recommendations. Bank offered credit cards and credits via call-center call-downs, debit and premium debit cards via SMS broadcasting for different groups of clients. Experiment was completed after a month from the moment when call-center contacted with last client from groups. Results of A/B testing shew significant excellence (in conversion and sales) in group formed by recommender algorithm in comparison to two other groups.

Methods

  • Machine learning
  • Recommender systems
  • A/B testing

Outcomes

Data quality analysis. Deloitte Analytics Institute provide to the client overall status of its data sources, consistency of this sources and ways to solve existing problems with data processing.

Client profiles for key products of Bank retail business.

Recommender system, which can select clients for selling core bank products.


System efficiency for each product:


  1. Credit card: +80% sales growth against random recommendations
  2. Credit: +376% sales growth against random recommendations
  3. Debit card: +50% growth against random recommendations

How our solution affects the Shareholder Value

Shareholder Value

  1. Revenue Growth
    1. Volume
      1. Acquire New Customers
      2. Retain & Grow Current Customers
      3. Leverage Income Generating Assets
    2. Price Realization
      1. Strengthen Pricing
  2. Operating Margin (after taxes)
    1. Selling General & Administrative (SG&A)
      1. Improve Customer Interaction Efficiency
      2. Improve Corporate/ Shared Service Efficiency
    2. Cost of Goods Sold (COGS)
      1. Improve Development & Production Efficiency
      2. Improve Logistics & Service Provision Efficiency
    3. Income Taxes
      1. Improve Income Tax Efficiency
  3. Asset Efficiency
    1. Property, Plant & Equipment (PP&E)
      1. Improve PP&E Efficiency
    2. Inventory
      1. Improve Inventory Efficiency
    3. Receivables & Payables
      1. Improve Receivalables & Payables Efficiency
  4. Expectations
    1. Company Strenghs
      1. Improve Managerial & Governance Effectiveness
      2. Improve Execution Capabilities
    2. External Factors

ROI

One year project ROI is 923%