Data & Analytics

00:01 AM
Ross Wainwright
Ross Wainwright
Commentary
50%
50%

5 Best-Practices for Bringing Big Data to Banking

If there is one thing banks do not lack, it is data.

The volume and velocity of data continue to increase exponentially. IDC research indicates that the global volume of data will increase from 130 to 40,000 exabytes by 2020. Banking is one of several vertical industries where this growth will be most pronounced. The new digital consumers are demanding much more from their financial institutions when it comes to quality of service and benefits. To hold onto this audience, banks are heavily investing in integrated channels, CRM, and data collection to better understand and communicate with them. This creates a massive flood of data that needs to be properly managed in order to be effective.

Regulatory requirements such as Dodd-Frank are contributing to more integrated, granular data. Meeting compliance, along with achieving ideal risk and return scenarios, requires heavily detailed financial and transactional data -- more detail than has previously been needed or was possible to process before the arrival of advanced analytical tools.

Big data is the new reality for banks both big and small. They need to efficiently parse through large sets of data for meaningful insight and information that can guide the business. It is this insight that leads to better, faster decisions. Database innovations such as in-memory are revolutionizing how structured and unstructured data can be consolidated and analyzed. For banks, it is powering real-time, personalized offers to customers, identifying fraud, and providing a more granular view of credit and liquidity risks. Most banks have analytical systems in place, but big data takes it to a macro level with deeper insight coming from a multitude of sources.

So how does one adopt a big-data framework? As the number of sources and data variety grow, it becomes much harder to get an accurate read on all the information that is out there. Here are some best-practices for managing big data in financial services:

1. Start with key business problems. Analyzing data in a vacuum will marginalize its potential value. Identify critical business issues upfront, such as cross-selling, fraud detection, and risk management, and explore how the insight associated with those issues can be leveraged. 

2. Collect clean data. In order to be of any value, datasets need to be as clean and accurate as possible. Automated processes can be put into place for filtering out bad data at the point of capture or later in the process. They should be able to detect and eliminate possible errors, duplications, and other inconsistencies, leading to a single, accurate depiction of the truth.   

3. Speed is essential. The value of data can decay quickly with time, requiring a need for real-time accessibility. When deciding on a big-data system, consider the issue of reduced data latency to aid folks dealing with consumers on the front end. Combine this speed with easy-to-use, intuitive software interfaces that make querying and visualizing complex datasets more inviting to the user. This will insure productive usage.

4. Pull a diverse set of data to achieve greater value. To get an accurate reading of a situation you need an all-encompassing view taking into account variables of all shapes and sizes. Look at external unstructured data like social media or websites for customer sentiments, and internal structured sources like risk data and payment traffic. The broader the field, the more likely your reading of the situation will be correct. Don’t dismiss new datasets at face value until you have thoroughly tested them for their potential. 

5. Apply human oversight over automated processes. A human touch is needed to guide and shape the data into useful insight. Appoint a Chief Data Officer who can oversee a bank’s data administration and data mining across the entire organization. He or she should be an equal member of the executive team who can implement strategies and procedures on the enterprise level. Also loop in a broad range of disciplines to contribute as big-data scientists -- their backgrounds and unique perspectives can add to the data. 

Big data needs to be front and center in the IT plans of all financial institutions. As the financial sector becomes increasingly competitive, the value contained within these datasets can determine success or failure. Following these steps will ensure you have a system in place that leverages big data rather than being left paralyzed by it.

Ross Wainwright is global head of financial services industries for SAP. He is responsible for SAP´s end-to-end footprint and ambition in FSI, which today covers more than 17,000 FSI clients. His team's objective is to help customers transform their legacy ... View Full Bio

Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Page 1 / 2   >   >>
lojo1
50%
50%
lojo1,
User Rank: Apprentice
7/17/2014 | 11:23:28 AM
Great article Ross
Really enjoyed what you had to say on big data and banking! Really looking forward to the next installation!

Laurence - Natwest telephone banking
Jonathan_Camhi
50%
50%
Jonathan_Camhi,
User Rank: Author
7/16/2014 | 3:16:34 PM
Re: Big Data best practices
I guess to an extent this would depend on the agility that banks have within their organization to get different teams and departments collaborating together while working off the same clean and accessible data sets.
Becca L
50%
50%
Becca L,
User Rank: Author
7/15/2014 | 8:46:20 PM
Re: Big Data best practices
Ross, great article. The demand for granualr data is now leading to the popularity of data lakes - huge depositories of unstructred and unstructured data - it's completely mindblowing how much information banks are storing and will probably never get around to leveraging.
Ross Wainwright
50%
50%
Ross Wainwright,
User Rank: Apprentice
7/15/2014 | 2:56:37 PM
Re: Big Data best practices
That depends on the roles and responsibilities within the organization. All roles can contribute – the compliance folks regarding data requirements, data scientists / data czar in terms of data architecture, and IT for data validation and correctness of data.
Ross Wainwright
50%
50%
Ross Wainwright,
User Rank: Apprentice
7/15/2014 | 2:30:54 PM
Re: Big Data best practices
Data is still siloed and difficult to sort, filter and present quickly .......however regulators are asking for more detail . Not aggregates but fine grain. In addition, each quarter brings new and changing requests. We have seen a big uptick  in requests for trade surveillance data after the recent libor fix investigations for instance.
KBurger
50%
50%
KBurger,
User Rank: Strategist
7/15/2014 | 2:30:34 PM
Re: Big Data best practices
That makes sense. Does the responsibility for this work fall to the compliance folks, IT, the data scientists/data managers, or who?
Ross Wainwright
50%
50%
Ross Wainwright,
User Rank: Apprentice
7/15/2014 | 2:28:05 PM
Re: Big Data best practices
Thanks for the feedback Kathy. Yes, you are right that banks have this data but a lot of the time it's in the aggregate. We now see the need to hold data at its most granular level so that queries can be replied to in a fully complete form.  Regulators expect banks to hold this data leading to expanding storage needs.

 
HM
50%
50%
HM,
User Rank: Apprentice
7/15/2014 | 11:22:02 AM

Very interesting article Ross! Many uses of big data have a measurable positive impact on outcomes and productivity. Areas such as record linkage, graph analytics deep learning andmachinelearning have demonstrated being critical to help fight crime, reduce fraud, waste and abuse in the tax and healthcare systems, combat identity theft and fraud, and many other aspects that help society as a whole. It is worth mentioning the HPCC Systems open source offering which provides a single platform that is easy to install, manage and code. Their built-in analyticslibraries for Machine Learning and integration tools with Pentaho for great BI capabilities make it easy for users to analyze Big Data. Their free online introductory courses allow for students, academia and other developers to quickly get started. For more info visit: hpccsystems.com
 
andrewboon2739
50%
50%
andrewboon2739,
User Rank: Apprentice
7/15/2014 | 10:27:05 AM
Data analytics is a powerful fraud prevention

Interesting thoughts Ross analytics in banking maybe new but has the potential to be a big differentiator. I work with McGladrey and readers may find this whitepaper on using analytics to detect fraud interesting @ Data analytics is a powerful fraud prevention and policy enforcement tool  http://mcgladrey.com/content/mcgladrey/en_US/what-we-do/services/financial-advisory/forensic-accounting-and-fraud-investigations/data-analytics-is-a-powerful-fraud-prevention-and-policy-enforce.html



Jonathan_Camhi
50%
50%
Jonathan_Camhi,
User Rank: Author
7/14/2014 | 3:46:07 PM
Re: Big Data best practices
Id be surprised if there's that much data that regulators are askign for that the banks haven't already been collecting for a while. But I would think that data is often siloed and maybe difficult to find.
Page 1 / 2   >   >>
Register for Bank Systems & Technology Newsletters
White Papers
Current Issue
Bank Systems & Technology - August 2014
Modern core systems are emerging as the foundations of effective channel integration and customer engagement initiatives.
Slideshows
Video
Bank Systems & Technology Radio