Application of advanced analytics and machine learning in the banking industry

December 26, 2018
5 mins

Banks have always been custodian of customer data, but they lack the technological and analytical capability to derive value from the data. On the other hand, fintech companies have the analytical capabilities and, thanks to payments services directives, they now have access to valuable data.

Whether it is a bank, non-bank, or fintech, competing in the banking revolution comes down to how efficiently the available data can be used to solve business challenges and better serve the customers.

Hence, leveraging banking data is no longer an ambitious technology project; it is a business imperative. According to the Global Banking Outlook 2018 study conducted by Ernst & Young, 60-80% of the banks are planning to increase investment in data and analytics and 40-60% plan to increase investment in machine learning. While some banks are leading from the front, some are just getting started and a few are still lagging.

It has become evident that the returns on data initiatives can be significant. The pioneers are already reaping the rewards. In this article, Mckinsey highlights a few examples of various banks that employed advanced analytics and machine learning to produce impressive results: “a European bank reduced churn by 25%, a US bank increased revenue by 8% in a few months, and an Asian consumer bank increased the likelihood of upselling by three folds by creating 15,000 micro segments of their customer base.”

The move toward data-driven culture in banks is so palpable that it has given rise to the position of Chief Data Officer in banks.

A recent survey by HackerEarth shows that 35 of the world’s top 50 banks have a dedicated Chief Data Officer. Majority of these positions were created for the first time and within the last five years. A few banks such as Royal Bank of Scotland and Santander have gone a step further and have appointed a Head of Data Innovation. This highlights the need for banks to undertake innovative and imaginative approaches to fuel growth by integrating data into the core business.

Antonio Alvarez, the Head of Data Innovation at Santander UK Technology, in his Linkedin post outlines significant advancements that can be made in a short span and relatively smaller budget.

With a £10M annual budget in Opex and Capex, we have created in less than two years a collaborative environment that enables guided self-service analytics, consolidated management information system, and simplified reporting. Furthermore, we have gone as far as creating real-time applications that deliver insights to customers directly or through channels enabling new marketing strategies and omnichannel experiences.

Impact areas

Analytics and machine learning on their own are mere buzzwords. But when used in conjunction with specific use cases such as open banking applications, security and fraud detection, etc. they can yield significant impact by enhancing customer experience and creating new sources of revenue.  However, in terms of impact the it broadly falls into these four categories:

  • Revenue and growth
  • Profit and productivity maximization
  • Risk management
  • Enhancing customer experience

The following table shows the use-cases and the impact level of data analytics in transaction banking.

Advanced analytics and Machine Learning in trasactional banking

Source: McKinsey

Here are a few examples of banks and other financial services companies that are making their data count.

The Kenyan lending craze

Mobile-based lending backed by data analytics is so lucrative that it has created a micro-lending frenzy in Kenya in the last few years. By gaining access and analyzing data such as user information, user location, peak borrowing time, and transactional data from the Kenya central banks, these companies are able to offer micro-credits on the go. This paved way for people to access untapped and underserved markets offering huge growth opportunities. It should come as no surprise that the leading mobile network operators, such as Vodafone, fintechs, and banking giants, such as Barclays Kenya, have entered the market and are expanding rapidly.

Making sense out 10 million customer calls

TD Ameritrade is an online stock trading and investing firm. In an interview with The Financial Brand, Beaumont Vance, the MD of Analytics Centre of Excellence at TD Ameritrade, explains how his company is using AI to derive value from previously unused data.

Ameritrade gets more than 10 million calls per day from customers. Yet the company was not able to derive any meaningful insights or decisions. Having tried and failed to do things manually via call centre representative, the company used AI software to turn all the calls into data, analyzes voice call and identifies patterns and trends. Using the insights the company was able to make changes to its websites and information it provided. Vance estimates that 1% reduction in call centre volume can cut the costs by $2.8 million.

Read the complete story here>>

Societe Generale’s talent hunt

There is a global shortage of data scientists as 50% of the demands are unmet. Finding skilled data science professionals is even harder for banks as they have to compete with the tech companies. Societe Generale Global Solutions Centre (SG GSC) is a subsidiary of Societe Generale—the French multinational banking and financial services company. The bank took an interesting approach to identify and recruit talented professionals. SG conducted a series of  online machine learning challenges. Over 2000 data science professionals and enthusiasts participated in the challenge to build over 300 predictive models. SG was able to spot and recruit top talent within a very short span.

Read the complete story here>>

Signature Recognition using AI

Axis, one of the largest private-sector banks in India, conducted an AI hackathon to solve two specific business challenges.

  • Signature recognition
  • Table reading and understanding in documents/image

The hackathon received over 3500 participants. The winning idea was an AI model that can recognize signatures with a confidence score of 84.78%.

Signature recognition using AI

more about Axis AI challenge here>>

Implementing a data-driven culture

Here are few important you need to remember before implementing data-driven culture in your organization.

Business-driven data strategy

Business agendas and customer experience should dictate the data analytics strategy rather than other way around. If opening cross-selling is the business goal, then building a next-to-purchase predictive model by segmenting customers, identifying untapped cross-product opportunities would be a good way to proceed. In an interview with McKinsey, Rob Casper, the chief data officer at JPMorgan Chase says,

The best advice I have for senior leaders trying to develop and implement a data culture is to stay very true to the business problem: What is it and how can you solve it? If you simply rely on having huge quantities of data in a data lake, you’re kidding yourself. Volume is not a viable data strategy. The most important objective is to find those business problems and then dedicate your data-management efforts toward them. Solving business problems must be a part of your data strategy.

Efficient data management

Banking data is perhaps the most sensitive information around. Safety and security have to be top priority. It is essential to define basic guidelines such as,

  • Who can access the data?
  • When can the data be accessed?
  • What can and can’t be done with the data?
  • What are the guidelines for transforming or migrating data?
  • What are the mandatory trainings and upskilling required?

Friction-less data access

The decision makers should be able to access to the data with ease. If the sales/marketing team is responsible for new accounts opening, they should have access to the data to create personalized product suggestions.

Measuring ROI

As any other function, the efficiency of your data team should be constantly evaluated. It is important to check the basics such as,

  • How many successful predictive models have been deployed?
  • What is profit or cost saved via the deployed models?

Conclusion

Banks and financial services companies use different approaches to tap the potential of data analytics. Some banks partner with advanced analytics company such as Citibank partnering with Feedzai (a machine learning and AI company) for risk management in payments transactions. Some depend on third-party self service banking analytics software and a majority of the large institutions such as Deutsche Bank have set-up their own operational and excellence centers to enhance their analytical capabilities. However, in the long run, data analytics will be a crucial factor for banks and financial institutions in building competitive advantage.

  •  
    40
    Shares
  • 40
  •  
  •  
  •  
  •  

About the Author

Vivek Siva
Learned what not to do as an entrepreneur. An optimist with a love for business, the outdoors, and movies. Superpower: Ability to smile even when all hell breaks loose. Affiliation: World of witchcraft & wizardry

Want to make innovation the driving force behind your business?

Subscribe to our Innovation blog