A novel approach of recommender systems in retail banking
Who is this presentation for?Non-technical or Business audience
TBC Bank is in transition from a product-centric to a client centric-company. As such, one of the most obvious applications of analytics is in developing personalized and automated next-best product recommendation for clients. Successful applications of recommender engines in retail banking are still rare, but the potential benefits are enormous—TBC Bank’s pilot increased sales conversion rates by 70%.
George Chkadua and Levan Borchkhadze explain how, after considering various collaborative filtering approaches, TBC Bank decided to implement the ALS user-item matrix factorization method and demographic model. There are no explicit ratings available that can be exploited by a recommender system for retail banking, which makes modeling more complicated. The main challenge was to handle implicit data and the nonhomogenous nature of the industry features in contrast to entertainment businesses.
Using transaction data and historical portfolio holdings, TBC Bank overcame the challenge by introducing an acceptable interpretation of input data and corresponding metrics. In particular, it used customers’ active products (cards, loans, deposits, etc.) portfolio alongside a timeframe of usage. Besides the mentioned traditional banking products, it introduced banking services (money transfer, utility payments, POS transactions, etc.) as products in its user-item matrix, which was one of the key points for model development.
Finally, TBC Bank mixed client profitability and lifetime value into the decision system as well: if a client is profitable, the recommender system will freely recommend any product, but if not, recommendations are constrained to profitable products. This way it ensures that its profit goals won’t be neglected.
After multiple online test iterations, TBC Bank obtained surprisingly good results that were well comparable with offline tests. As a result, customer-tailored offers boosted sales and profitability and increased customer satisfaction at the same time.
- A basic understanding of recommender systems
- General knowledge of linear algebra (useful but not required)
What you'll learn
- Learn about recommender systems in fintech, implicit data modeling, and business integration of data science projects
George Chkadua is a data scientist at TBC Bank. His main focus is machine learning and its applications in industries from a mathematics and business perspective. He earned a PhD in mathematics from King’s College London. George has published various articles in peer review journals and has been invited speak on many scientific conferences and seminars.
Levan Borchkhadze is a senior data scientist at TBC Bank, where his main responsibility is to supervise multiple data science projects. He earned BBA and MBA degrees from Georgian American University with a wide variety of working experience in different industries as financial analyst, business process analyst, and ERP systems implementation specialist. Levan earned his master’s degree in big data solutions from Barcelona Technology School.
Leave a Comment or Question
Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?
Join the conversation here (requires login)
Premier Diamond Sponsors
Premier Exhibitor Plus
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires