Case Study

Article

Credit: Cloudera (Screenshot from video at www.cloudera.com/more/customers/dbs-bank.html)

Credit: Cloudera (Screenshot from video at www.cloudera.com/more/customers/dbs-bank.html)

How DBS transformed into a data-driven organisation

Big data is driving fundamental transformation across all industries and sectors, and finance is no exception. With the proliferation of mobile devices and rapidly increasing share of digital transactions, financial institutions are striving to make sense of the massive volumes of data being generated and captured to understand their customers and predict their needs, so as to serve them better. Around the globe, they are also under increasing regulatory pressure to improve risk reporting and controlling money laundering and fraud.

As a leading bank in Asia, DBS encounters these issues and hence, it started moving towards becoming a data-driven organisation a few years ago, taking smart decisions based on data and not instincts.

However, the company’s traditional technology stack for supporting advanced analytics was expensive to scale and not flexible enough to support this work.

With Cloudera as a partner, DBS built a central data team and enterprise data hub, enabling DBS to scale out more economically, and experiment more. The agility of the platform allows the bank to explore use cases and iterate easily and quickly, without the need to worry about ROI and build an investment case beforehand.

With the ability to more easily store and analyse billions of events in a modern data platform, DBS can answer questions before they’re asked to more effectively engage customers and deliver better service.

This has enabled DBS staff to experiment more and be on the forefront of innovation when it comes to understanding the customer experience and applying human-centered design to its services.

For instance, machine learning can be used to understand customer sentiments. All calls to the bank’s call centres are recorded. They can be converted to text and then machine learning algorithms can be used on the analytics platform to understand sentiment. Problems can be flagged so that the bank can reach out to the customers.

Ultimately behavioural information and machine learning, in combination with biometrics, could even enable ‘invisible authentication’, where a customer no longer needs to provide many supporting documents or use a physical device for transactions or answer questions like, ‘What is your mother’s maiden name’.

In a video interview with Wee Wu Neo from The Neo Dimension, David Gledhill, Head, Group Technology and Operations, DBS explained that the use of data goes beyond to customers. The transformation to a data-driven organization has significantly improved operations across the organisation.

Data can be used to find out where fraud is happening in the company. To take a specific use case of this type, trade financing is highly prone to fraud. To deal with this, DBS started looking at data other than invoices and transactions to predict the possibility of fraud.

“You look at things like ship movements. If you know the typical movement patterns of goods from one port to another, then anomalous goods movement or timing that doesn’t look like typical timing for that type of transaction or a behavioural shift in importers or exporters or in warehousing, signals where potentially fraudulent trade might be going on,” Mr Gledhill said.

Data analytics can predict the likelihood of a relationship manager quitting within the next three months, so that HR staff can intervene early to retain employees. Data can tell the audit department which branch might have issues and should be audited next.

Operational staff can understand and predict customer flows, ATM load, and call centre volumes using data. In fact, one of the first big data projects DBS embarked upon was figuring out the sequence in which ATMs should be filled. The bank went from hundreds of instances of ATMs running out of cash to single digit numbers.

The bank also moved its financial risk information and data required for regulatory reporting on to the Cloudera platform to simplify reporting.

Mr Gledhill said, “We’ve applied it to a whole range of different use cases and every single one, we see a massive uplift in terms of the base case that we normally do.”

This has also been aided by the huge active worldwide community of Hadoop contributors. It includes not just individuals but also tech giants, such as Netflix, Amazon and Facebook (the platform itself was inspired by technologies created inside Google). So, the platform keeps evolving and improving steadily and DBS can build on the contributions made by this vibrant community.

DBS wanted to make the data analytics capabilities available to everyone in the bank, as opposed to having a separate team of data scientists or little pockets of analytics.

However, the oft-repeated cliché of technology being easy and the ‘people’ aspect being hard was true.

The more difficult part was opening up people’s minds to the possibilities. The first few use cases played a key role in overcoming scepticism. They generated a high level of interest and enthusiasm among different teams within the bank. They began to explore how they could leverage analytics in their area.  

All this improvement in services and operational efficiency has been achieved while reducing costs. 

Mr Gledhill said, “We’ve seen anything in the region of 80% reduction in operating cost in a much shorter build time. The real big benefit lift though is the benefit it provides to the business. If you look at our digitally engaged customers, we see material lift in how much revenue a digital customer brings to the bank.”

This is an ongoing journey and DBS expects Cloudera to help them continue along the path towards deeper, better insights.

Content from Cloudera customer success story on www.cloudera.com and video interview at The Neo Dimension.

Visit site to retreive White Paper:
Download
FB Twitter LinkedIn YouTube