‘Bank’ing On Data: The Analytics Behind Banking Innovations

Financial institutions are banking on use of data science and big data analytics in colossal ways. Starting with protecting clients from fraud, reducing customer churn, fostering customer loyalty, to identifying behavioral patterns to increase conversion rates.

IDC estimates nearly $20 billion is invested in data science and big data analytics by financial institutions. This number validates the fact that financial institutions are evolving their innovation efforts, grasping new technologies and thought processes by investing in building data science-backed analytics platforms.

While other banks are yet to realize the potential of AI/ML-based data-driven decision making, Yes bank as a financial institution invests heavily on making data-driven decisions which is evident from the fact that nearly 30% of the banking experience of retail customers & 10% of corporate customers are driven by data analytics, which I learned from my interaction with Sanglaph Bannerjee, EVP Corporate Strategy, Yes Bank.

Yes Bank Datathon Logo

While many banks are yet to realize the inherent capabilities of embedding data science-backed big data analytics into their cultural & decision making process and business operations, there are financial institutions like JP Morgan Chase have built systems that analyze legal documents and extract out critical data points and clauses using machine learning algorithms and opportunities engine to predict and determines a clients best positions in equity deals.

Banks hold a treasure chest of data in form of historical transaction records, legal documents, etc. that has variety, volume, and velocity making it clear candidates for big data analysis. Value is the what they have to derive from the big data in form of insights.

Data science is the solution to analyzing data to find insights. This brings us to the question of what can we gain from investing in data science and machine learning talent:

  1. Uncover spending patterns of customers.
  2. Discern channels of transactions.
  3. Buyer persona building, products cross-selling, leading to better CRO.
  4. Fraud detection & risk management.

Here are few action items banks can leverage right now to kick-start their data science and big data analytics efforts and derive richer insights:

  1. Carve out a budget for building and retaining data science and machine learning talent.
  2. Run machine learning models on big data banks currently possess, churn out continuous analytics report, use that in the decision-making process and business workflows.
  3. Invest in building a modular, microservices architecture oriented data platforms for facilitating data scientists to run their ML models on and analytics teams to automate report generation.

When a financial institution has understood the potential of big data analytics, data science and AI/ML, the next major hurdle comes in deciding what to build with data science. Here are a few products that can be built with AI/ML algorithms:

  1. Natural language understanding chatbots that leverage machine learning models to drive customer journeys and provide an automated digital experience.
  2. Backtesting models that use historical data to validate strategies before rolling them out.
  3. Risk assessment systems and risk prevention systems for prediction and thwarting fraud before they happened.
  4. Document analysis to figure out critical data points and tricky clauses, which can help both corporate as well as retail customers and improve customer retention.

Here are few practical and actionable guiding points to decide on efficient data analytics strategy:

  1. Great analytics is driven by asking the right questions, not necessarily possessing a huge amount of data.
  2. Handling the edge cases make difference significantly and can widen the market gap between you and competition.
  3. Automation beats manual labor anytime, so investing in building data platform is a step towards improving organizational effectiveness and decreasing human friction.
  4. Analytics has to be a team effort and it has to be adopted into business workflows and strategical decisions.

As an effort towards disruption fintech industry and democratize the process of building machine learning models, Yes bank has taken up the initiative of organizing datathon for the community towards building interesting projects by facilitating them with an anonymized dataset of the historical banking data.

Yes Bank Datathon last moment wrap up for presentations

The datathon organized by Yes bank was a 60 days long hackathon styled event concluding on Dec 23, 2018, where 25 teams were selected from 1700 teams. The teams were facilitated with the last 5 years of anonymized transactional data. Yes Bank team was closely working with the teams throughout the span of the event. The teams were expected to come up with innovative analytics models to derive breathtaking insights.

By truly establishing big data analytics as a core business discipline, banks can realize the enormous potential they can leverage against the competition.

PS: I’d be very happy if you could share this with your facebook/twitter network.

PPS: Follow @YesBank on twitter.