This blog moving to Medium

I prefer Medium's platform, moving this blog to the following going forward: https://medium.com/@terigrossheim

Considerations for Data Analytics in Financial Services

While financial services organizations have been early adopters of data analytics, they also have unique challenges. The need for compliance in an industry filled with regulation, it’s critical for financial services organizations to possess accurate and timely data. A mistake or decision made on inaccurate data can cost financial services organizations time, money, and their reputation. This post will explore various unique situations and challenges to financial services organizations with data analytics.

Chief Data Officer Trends

Banking and insurance are among the first industries to take on the Chief Data Officer (CDO) role. Like government, these industries have overarching regulatory requirements, risk-management practices, and a practical knowledge of the importance of data in business and reporting to industry regulators. CDOs exist in a world overcome by data, along with regulation and risk, likely functioning in a complex organization. While the role of CDO continues to evolve, there is an ongoing battle between a CDO’s alignment with IT versus business.  Since the use and management of data belong to both business and IT, CDOs should make note that one is not exclusive of the other. It’s truly a team effort.

Auditing

Internal auditors at financial services organizations also benefit from data analytics. Business processes that pose financial, regulatory, contractual, and fraud-related risks are often audited, as a result auditors must be prepared to provide assurance within a dedicated timeframe. While there are dedicated tools for data analysis for auditors, there are a few capabilities that transcend tools and technologies:

• ETL (Extract, Transform, Load) – In order to obtain data from various (or even single) data sources, ETL is a set of functions that obtain data from sources such as an ERP, CRMs, and databases such as SQL Server.
• Data quality ensures the dataset is valid and complete, does not contain inconsistencies, and matches the data dictionary provided by the owner.
• Sampling a set of features that allow for selecting subsets of data based on questions that need to be answered (eg. How many insurance claims were created last month?)
• Query-based analysis via data sketch, histogram, and Benford diagram
• Lastly, reporting and workflow that supports auditors the ability to make decisions and draw conclusions. Traditional reporting, dashboards, and audit trails fall into this area.

Big Data

Financial services organizations have also been early adopters of big data. Big data within financial organizations is also not immune to regulation, security, and privacy, which all continue to be priorities. In some countries, information barriers are required between certain types of businesses in banking and insurance. With big data, organizations need to be aware of deterministic versus probabilistic modeling. For instance, in probabilistic model analysts should be aware of false positives, such as a financial transaction following a certain path that money launderers follow. While the transaction might follow a certain pattern, it might be for a different reason.

Privacy concerns around big data have also been under the microscope in the media in recent months. The credit and insurance organizations possess volumes of data but need to continue compliance with the Fair Credit Reporting Act, where these organizations do have permission to collect personal data. Financial services organizations must continue to be sensitive to these regulations as they move forward with big data initiatives. A recent World Economic Forum reports shows promise in possible future standards, where codes are assigned to an individual’s preference regarding how his/her data will be used.

Disruption within the Financial Industry

Speaking about personal data, there is discussion regarding the financial industry using personal data for new sources of revenue. Personal data bank (PDBs) is a concept where an entity will collect, protect, and monetize personal data and ensure that this data is entrusted to individuals. In theory, PDBs will operate with personal data similar to bank operations with customer money. According to Maverick Research by Gartner, PDBs can potentially create a new market and with the absence of government regulation, first mover advantage can take hold and as a result, have PayPal-like success. 

References:

Big Data Is Opening Doors, but Maybe Too Many. New York Times, March 2013: http://www.nytimes.com/2013/03/24/technology/big-data-and-a-renewed-debate-over-privacy.html

Big Data Analytics Requires An Ethical Code of Conduct. Gartner ID G00256399. November 2013

CIO Advisory: The Chief Data Officer Trend Gains Momentum. Gartner ID G00254672. January 2014

Maverick Research: Put Your Data in the Bank, Get Dividends. Gartner ID G00264262.  September 2014

Technology Overview: Data Analysis for Auditors Can Lower Audit Costs and Detect Fraud. Gartner ID G00247280. April 2013