Achieving the goal of monetising data assets through disruptive technologies such as blockchain, machine learning and data ontology standards requires thoughtful harnessing of these resources – and collaboration among units of firms, according to data management experts who spoke in a panel discussion on data innovation at the Data Management Summit hosted by A-Team Group in New York on November 17.
“In our organisations, for all the investments we put in, data is still a bottleneck for innovation, as opposed to being a driver for innovation,” says David Blaszkowsky, a former head of data governance at State Street. Citing Michael Stonebreaker, MIT professor and chief technology officer of Tamr, who delivered a keynote presentation at the event, Blaszkowsky mentioned that data scientists at large firms are spending 60% to 80% of their time fixing data in order to apply data science techniques.
“How can you spend time innovating, and be a forward-looking organisation? People doing data science don’t want to spend all their time doing scrubbing,” Blaszkowsky said. “Technologies like blockchain open opportunities for innovators to grab hold of the data content.”
New York-based Concur Reference Data applies blockchain protocols to source fixed-income reference data. Its co-founder and CEO, Tim Rice, noted that blockchain technology (which includes distributed ledger technology) facilitates leveraging of open standards for data, such as FIBO (Financial Industry Business Ontology).
“Blockchain technology … gives us a better opportunity to collect the source bond information in the same semantic framework that FIBO would require later on,” Rice said.
Chris Betz, a consultant and senior advisor at the EDM Council, which is the developer of FIBO, pointed to the need for more dynamic and faster data standard solutions that can cover the widest possible variety of assets and securities.
“A year ago, there were six blockchain proofs of concept (PoCs). Now there’s 70. There are a hundred different organisations involved and they’re all trying to figure out how to slice significant costs out of their infrastructure,” Betz said. “From a capability and innovation perspective, how quickly will new technology architectures be adopted? How can we use FIBO across asset classes, to define assets in a pure, de-materialised way? How does that accelerate business?
“There’s significant demand from a legal, regulatory and compliance perspective today,” he added. “Having watched blockchain and FIBO for the past year, and the funding model required for enterprise data management and best practices, the speed of delivery and the agility to deliver on what the industry is looking for is going to be a challenge. The difficulty of getting everyone’s agreement and consensus is no small thing.”
Machine learning capability has made it possible to conduct analytics on data at a greater scale, noted Tassos Sarbanes, data architect at Credit Suisse. Innovation in the form of distributed ledger technology or newer ontology standards could produce similar dividends, he suggested. Sarbanes and Rice both said the industry needs open collaboration about terms and conditions – such as FIBO, or otherwise – to drive data management for its business.
Blaszkowsky counseled firms to consider available innovations and not commit to a solution too early. “In data governance, blockchain and semantic data, there’s great opportunity to make better products,” he said. “The alternative, which I’ve seen firsthand, is letting the government require something, and then vendors show up. … That’s not the best way to do it.”