Big data comes to FX
The analysis of structured, semi-structured and unstructured information from multiple sources, commonly referred to as ‘big data’, could improve FX pricing as well as reduce the potential for regulatory infringements, according to technology experts.
Foreign exchange markets generate large volumes of communication. Some of the information hidden within these interactions could be extremely valuable to market participants. However, in the past there were limitations on accessing some of this data. For example, e-mail messages, text documents, videos, photos and audio files are all unstructured data since the information they include does not fit into conventional databases.
However, in recent years analytics technology has developed to the point where companies are able to extract insights from such data.
Financial organizations need to use streaming analytics (which operates in real time) in conjunction with historical data from conventional business intelligence systems if they are to boost retention and identify potential clients, both retail and institutional, according to Giles Nelson, vice-president, intelligent business operations, at Software AG.
“This integration allows for customer profiling, not just to personalize offers but to achieve optimal timing. In foreign exchange markets, where an institution wants to check thousands of transactions per second, streaming analytics will enable it to narrow the focus and monitor the success rate in quoting foreign exchange prices to potential customers.”
Historical data on clients’ previous transactions and the types of products they have traded is important to inform recommendations and effectively match client requirements to products, says Qlik senior director of global financial services Duncan Ash. “With a detailed understanding of these trends, the institution would know which products to recommend in different market conditions and be in a position to advise clients when those conditions arose.”
Historical analysis will reveal a particular customer's usual transaction size, timing and individual price point relative to a median in the market, Nelson says. “However, dynamic pricing that ensures a profit on a specific transaction at a specific time, while meeting the customer's likely requirements based on their profile, requires the use of real-time analysis.”
The use of streamed real-time data is becoming increasingly common, agrees Tony Boobier, IBM’s business analytics insurance leader. “In the development and marketing of foreign exchange products, the typical foundational capabilities we might see include predictive customer intelligence and big-data analytics, coupled with enterprise marketing management.”
Of course, banks already have a lot of real-time data that they analyse for very specific needs as dictated by their current analytical capabilities, which have been designed to allow the users to carry out broadly predetermined analyses.
Deutsche Bank uses a big-data application called FiREapps to aggregate, validate and analyse underlying exposure data from corporate customers. Its FX trade execution services customers then use the results of this analysis to manage their overall currency exposures and ensure their trades are compliant with corporate foreign exchange policy.
Vincent Kilcoyne, capital markets specialist at SAS, suggests that the true potential of real-time big-data analytics will be identified when the user can combine their inherent knowledge and curiosity with data (both internal and external) to be able to deliver benefit to their existing customers, but also to identify new product offerings that will attract new customers. “This will require the organizations to be able to integrate across their entire business, from market and customer engagement through to marketing and new product design,” he says.
This analysis would enable the financial institution to offer corporate customers personalized foreign exchange products based on their typical transaction size and timing. Kilcoyne continues: “Banks will be able to engage with the client in a much more holistic way when they know more about them – through understanding the way in which the customer responds to market pressures – and by making it possible for the customer to know more about the bank.”
The right analytical tools can also compress the timeframe between receiving reports back from trading venues and responding to the needs of customers, says Ian Keldoulis, head of financial services marketing at TIBCO Spotfire. “Whether that is adjusting your size and pricing to improve fill rates or recognizing the most advantageous venue to execute a particular type of trade, the more you can automate your customers’ lower-margin trades, the greater the opportunity to focus on complex, higher-value trades.”
Nelson suggests that market scandals such as foreign exchange fixing indicate that traditional business intelligence technology alone is not enough to prevent reputational damage, heavy financial losses and fines. “Continually looking at the data stream flags up anomalous trading patterns or banking transactions, which can be nipped in the bud by human intervention before they develop into something that requires the intervention of a regulator,” he says. “Real-time monitoring is also a deterrent to unauthorized activity by employees.”