In part one of this series, we looked at how data science (extracting useful information and knowledge from data) is perceived within the operational structure of financial technology organizations, based on the results of a survey of Georgia-based financial technology companies. We learned the following:
- Financial services and fintech organizations are in the early adopter stage with data science.
- Many of the initial data analytics programs have not yielded the results executive management expected. As a result, they are cautious about investing in new programs, especially with ever-increasing pressures on the bottom line.
- Employees’ perceptions, then, are that management is not committed to data science as a way of doing business, and not driving the innovative analysis of data into the cultural fabric of the financial services organizations.
We will now focus on the respondents’ view of the technology foundation that currently supports data science and what investments are needed to support data analytics and big-data programs. I’ll examine the issues surrounding data privacy/security and its impact in data analysis. And I will look at the plans of these organizations to adopt a data-driven model by 2020.
Let’s start with the technology infrastructure. For most financial services organizations, 60 to 65 percent of IT budgets is dedicated to maintaining legacy systems -- the core transaction processing systems. After general operating and personnel costs are deducted, there is generally less than 10 percent of the budget left over for new systems development, the source of new products and services. Changing these percentages is one of the biggest problems fintech organizations have in changing the course of the battleship.
On one hand, you would think that with the existing compute power within these organizations, there should be plenty of processing power to handle data analytics processing. Collectively, Georgia-based payment processors handle more than 85 billion, or two-thirds of all worldwide payment card transactions per year.
But it’s not that simple. Data analytics processing very often requires a different technology to analyze a combination of structured and unstructured data. Traditional data warehouses (commonly used in business intelligence) cannot scale as needed to meet growing demands. This is especially true with big-data analysis (the analysis of very large datasets). Traditional relational database management systems, for example, are not designed to analyze petabytes and petabytes of data, especially unstructured data.
As a result, a combination of new analytic software tools provided by commercial software vendors such as SAS and IBM and public domain analytics tools such as Apache Hadoop have revolutionized the way large volumes of data can be analyzed. New analytics hardware platforms often use massively parallel processing and in-memory databases to deliver remarkable speed at a reasonable price, often delivered in the cloud.
But, when asked whether cloud-based platforms were being used in fintech organizations, 78% of our research respondents indicated that cloud-based technology is not being used to augment the IT infrastructure. Fifty-four percent of the respondents indicated that management is not comfortable with the potential data security issues with cloud-based systems.
Data security is a sturdy guard rail as financial services and payment processing organizations look for ways to repurpose the vast quantities of data surrounding their customers’ transactions. Meeting or exceeding regulatory obligations and implementing strong privacy and security policies are very much top-of-mind for every fintech executive.
According to our research participants, data privacy and security will only get tighter over the next five years. When we asked our responders to look ahead to assess the data privacy and security landscape, here are some of the verbatim responses they shared:
- Increasingly more stringent data security and privacy requirements
- More sharing of data and therefore more obfuscation of data and its origins
- More frequent requests for security/regulatory audits from customers
- Strong information security will be increasingly demanded by consumers
- Introduction of Federal PII standard to streamline business compliance nationwide
- PCI compliance of Hadoop
- Better designed applications with a security focus
- Stricter, more expansive use of encryption and tokenization
In contrast to the state of data science in 2014, the vast majority of respondents were very optimistic that their organizations would evolve to become data-driven by 2020. They would move from being early adopters today to effectively competing with data as central player. Here are some of the key drivers shared by respondents:
- Competition will mandate that financial services organizations become data driven.
- Organizations will increasingly become purveyors of data to their customers, as 71 percent of respondents indicated that their source of data would be generated internally within the organization and be gathered from customers. Staying behind the firewall, so to speak, eases challenges faced by data security/privacy.
- While senior management commitment is a prominent issue in 2014, commitment will strengthen as the competitive landscape drives data science adoption.
- Fintech organizations will invest more in technology and the IT infrastructure to support data science. Investments will likely stay in-house as well, rather than using third-party or cloud providers.
- Much needed training programs will empower employees to nurture and mine data they handle daily to drive operational improvements, improve customer satisfaction, and define potential new products.
- More trained data scientists will be added to the organization.
Adding trained data scientists may be more difficult than it appears. McKinsey and Company has predicted that by 2018, there will be a shortage of 190,000 professionals with deep analytical skills. Potentially more significant will be the 1.5 million-person shortage of managers and analysts with the know-how to use data analytics to make effective decisions. To emphasize, it’s not just having the right technology, IT infrastructure, and software tools -- it takes the right training, knowledge, and expertise to interpret the results of the data.
Part 3 of this series, coming next week, will focus on defining a roadmap that should help guide financial services executives to develop data-driven organizations.
To view the results of this research study, please visit the Technology of Georgia website: https://www.tagonline.org/chapters-and-societies/fintech/. There are two whitepapers: One details the research results; the other provides a management summary and roadmap for how to become data-driven.
With extensive experience in the high technology industry, Don focuses on enterprise strategy development, operations execution, revenue generation, product definition, market/product positioning, and branding. RightCourse is a management consulting firm that helps FinTech ... View Full Bio