Sungard likes to publish top ten trends lists, and the latest focuses on big data in financial services. We chatted to Neil Palmer, partner at SunGard’s consulting services’ advanced technology business, to find out some more behind the predictions.
Q: You just published SunGard’s view of theten big data trends for financial services in 2012. What’s the commonality in them? What’s the elevator pitch for big data trends over the next year?
A: Big data trends are being driven by a combination of an increased volume of data from internal and external system that needs to be rationalised and the drive to be more economical with existing hardware investments. The elevator pitch we see from software and platform vendors is to be able ‘to achieve more with less’.
Q: How would you say these trends – for financial services – fit in with more general big data trends? Where is the commonality? What is unique about financial services?
A: Financial services have hard requirements for compliance reporting as well as for data security. These requirements make the sector unique in how Big data solutions are applied. For example, other industries can mine structured and unstructured data without having to segment data for security and privacy concerns.
Q: You mention larger market data sets, longer histories, for predictive markets. How much? How long? What’s going to be the norm?
A: Predictive analysis requires very large sample sizes where the algorithms learn over historical data. The norms will depend on the nature of the algorithms but generally the longer the better training of the system.
Q: How much is enterprise risk a real big data issue, compared to one of simply connecting data silos and normalising data? Where big data technologies are going to play in this important area?
A: Enterprise risk is always a concern for the financial sector. Big data frameworks are going to have security layers wrapped into them over the next year or two as they get embraced more fully by the corporate enterprise.
Q: Is regulation and compliance the ‘killer app’ for big data approaches? And does that mean they will be driven by ‘have to do’, as opposed to ‘big opportunity’ investment and management buy in?
A: Regulation is always a driver for adoption and big data is the ‘must have’ for making sense of increased volumes of information being captured from ever increasing numbers of computer systems.
Q: What do you think might be on this list for 2013?
A: Compliance will be a big driver through 2013 but there is also opportunity for banks to arbitrage the insight they have on their data over their competitors. Technologies such as Hadoop and other MapReduce frameworks will be used to optimise ETL strategies but we will also see machine learning and predictive analytics platforms emerge to bolster existing intelligence systems. This will include market recommendations, operational risk detection and fraud detection.