Data is everywhere; and in today's data-driven world, data management needs to be right. John Yelle, DTCC Executive Director, Enterprise Data Management, spoke with DTCC Connection about the necessity of a data management plan, and provided insight from his participation during a recent webinar, "Getting Data Management Right," sponsored by the Enterprise Data Management Council (EDMC) and Collibra.
Why is data management important?
Data touches upon the entire enterprise—from all aspects of sophisticated technology to spreadsheets as an end-user tool. And the data needs to be accurate. From a resiliency and recovery point of view, if you needed to failover to a different site—the first thing checked would be the data.
Data management is also necessary for an organization to be proactive. A data quality rule will help avoid the negative impact a quality issue could present by checking that data meets defined quality requirements. For example, from a controls and risk perspective, you want to be proactive instead of reactive. The earlier you find a problem, the better equipped you are to find a solution.
What are some challenges of data management?
We serve a critical role in the complex financial services industry—at the center of global trading activity, processing trillions of dollars of securities transactions on a daily basis. Every part of our job functions involves data. As a result, a key challenge is the sheer scale involved based on the volume of data along with it being an integral part of every aspect of our business. The complexity of the industry is manifested in the services we provide, and the data processed by each.
It’s important to carefully manage scope when establishing data management practices. That can be a challenge. Also, standing up a data management program requires extensive engagement and support. It’s not a project with a defined end date. Rather, when successful, a data management program changes the culture and establishes sustainable data management capabilities throughout the enterprise.
How are those challenges addressed?
A defined strategy supported by strong program management are used to set and manage expectations based on closely manage scope and priorities. The success of any attempt to change a corporate culture depends on communication, awareness and training that continually evolves as the program evolves. We also regularly measure our data management maturity levels using the EDMC’s Data Management Capability Assessment Model (DCAM) to track progress and adjust our roadmap and plans to address prioritized capabilities.
What is the future of data management?
Data is at the core of machine learning (ML) and artificial intelligence (AI). Both present a considerable opportunity as well as unique challenges.
If we look at a trade, it is made up of or described by data—the date, the number of shares, the price, the client. When processing is done through ML, the machine will examine the data and attempt to make the process more efficient by being trained and by learning. The machine then produces more data about the data and the process.
To incorporate ML or AI, an organization needs to understand their data and ensure that it is fit for purpose by being of the highest quality. Given the increased scale of data used and produced by ML, the implications of an error cause by the technology learning guided by bad data are exponential. It's even more of a reason that data management has to be a core capability of the firm.
How have past crises defined data management at DTCC?
Past challenges have taught us that we need to have the right data to make the right decisions. The crisis of 2008 drove a cultural change—both at DTCC and across the financial services industry—to greater data transparency. This change has enabled stakeholders to ask the right questions about data, including where the data originates and what controls are in place to ensure high quality.
Our continued investments in data ensure that we are prepared as best as we can be. At the height of the volatility and volumes seen in March, we couldn't wait to figure out what data was needed, or where to find it. Our past investments and preparations were rewarded. We had confidence in the data and that it was “fit for purpose.”
While DTCC and the industry have moved on from the height of the crisis, data continues to be central to risk and risk management. There is a continued evolution driving cultural change to have more transparency on legacy data. We can expand what has been in place and help data be sustainable and respond to change.
How does DTCC approach data management?
Data management is something we’ve always done though maybe not in a defined, comprehensive, consistent way across the enterprise. We use DCAM not only as a tool to assess our maturity levels but also as a framework used define practices supported by an operating model that works for DTCC. Clear roles and responsibilities have been established across business, operations and technology. We’ve invested in technology platforms that support our data management practices – data quality, data architecture, metadata management and data governance. The technology investment is necessary to enable effective and efficient scaling of our data management practices. We look at our strategy as a living document that continues to evolve. Doing so has positioned us well to be able to support forward looking efforts driven by strategic business objectives that will simplify our data architecture while also staying focused on the day to day table stakes.
How does an organization like DTCC maintain a data culture?
At DTCC, data is central to everything we do as an organization. To sustain the gains we’ve made, we continue to be laser focused on strategic priorities to determine scope and priorities for our data management program. While we have a central data management group acting as a control function, responsibility for data management is federated throughout the business, so we have an ever-expanding group of data quality stakeholders throughout the firm.
As data is so pivotal, we have a tremendous opportunity at the business unit level. When creating a new solution or application, it is essential to have data controls and data quality rules and measurements established at the same time. This proactiveness reduces not only the risk, but also the possibility of having to retrofit. The right level of understanding and communication is necessary for different stakeholders to see the benefit of investing in data.
A solid data management plan involves communication, awareness, and education to secure and maintain involvement across all enterprise levels.