Today’s banks are rich in data. So much so that the Boston Consulting Group believes banks should treat “data-driven intellectual property and analytics as one of their most valuable assets.” Because the data is only as valuable as its analysis.
Problem is, many of today’s institutions are saddled with legacy technology that can’t process or analyze data at the speed of business. And for many, new use cases require technology they haven’t yet obtained or deployed.
Banks need to organize around innovation strategies that embrace new ways of using data to power customer engagement, while simultaneously fending off disruptive fintech upstarts trying to displace the traditional banking system.
In each case, asset managers, market analysis managers, and lending institutions alike can put their data to work with the help of an insight engine: a GPU (graphics processing unit)-accelerated database that performs advanced analytics to discover patterns, train models, measure risk, detect fraud, and develop upsell opportunities, among others. Let’s delve into a few common finance data use cases.
Financial institutions can use an insight engine to perform risk calculations on-demand using the most up-to-the-moment data with sub-second speed. This allows banks to make better informed investment decisions and react quickly to market events, while reducing credit risk.
A GPU-accelerated database is an ideal solution for complex risk calculations, as it removes the bottlenecks of data processing that arise using only CPUs (central processing units).By leveraging the thousands of cores of general purpose NVIDIA GPUs, data is processed in parallel. Parallel processing is a good fit for running risk queries multiple times a day to meet critical SLAs, as it can scale and deliver results in real time.
Engines that leverage high-density compute processors can also be used to accelerate complex market, liquidity, or counter-party risk calculations like liquidity coverage ratios, Monte Carlo simulations, xVA, and trade decisioning to obtain results intra-day rather than overnight. For example, a GPU-accelerated database can ingest data from any source and return instant insights, processing and analyzing billions of rows of data in milliseconds. It runs in-database analytics to handle time-sensitive, compute-intensive risk analysis at scale to project years into the future across hundreds of variables.
For trading, a GPU-powered engine can be used for tick data storage and analysis to deliver instant insight using time series analysis (time window analysis) for asset pricing. From there, investment professionals and traders can make more informed decisions at the speed of the market, not delayed by processing time.
Analysts can model risk by bringing together many sources of trading data including instrument IDs, product codes, metadata, values, and more. A GPU-accelerated database helps analysts to understand in real-time the failures of their models and where adjustments are needed, and provides them with greater insight into the data and a broader understanding of risk exposure. Increased transparency and risk exposure analysis drives better business outcomes.
Protecting against fraud goes hand in hand with customer experience. A GPU-accelerated database can be applied to fraud detection by combining location data with transaction information and other customer data in real-time. Further, the visual discovery component of an insight engine can be used for geo-fencing to identify fraud across regions, particularly for mobile transactions.
Additionally, fraud detection analysts can perform queries on large streaming datasets in order to uncover anomalies and patterns of behavior that signal potential fraud. By looking holistically at a customer’s activity and applying advanced analytics and machine learning, banks can understand when behavior deviates from the baseline and flag it for review.
UK fintech Monzo is growing a massive following through community-building efforts. Monzo analyzes and categorizes spending habits and allows users to freeze and unfreeze banking cards with a simple tap — capabilities that most traditional banks don’t offer. As a result, banks such as HSBC, Lloyds Banking Group, and the Royal Bank of Scotland are in a mad sprint to play “digital catch-up.”
To compound these challenges, new regulations such as the Revised Payment Services Directive (PSD2) in the European Union and the Open Banking initiative in the United Kingdom remove barriers to entry and require banks to provide open APIs to third-party providers to access customer bank data — essentially eliminating traditional bank monopolies over customer account information and payment services.
A GPU-accelerated insight engine can break down barriers for traditional banks and fintechs alike. It can help dismantle silos to connect transaction history with marketing and promotions by offering a single platform for data analysis. This helps banks establish a single view of the customer and their relationships across each channel. Financial services organizations can then use these insights to drive targeted marketing of loans and credit cards to select individuals, offer discounts based on geolocation or commute habits, identify high-value customers, and even look ahead to the future with predictive capabilities as well as cross-selling and upselling.
It is the actionable business insights that are derived from extreme data (that’s increasingly complex and unpredictable) that make data one of the bank’s most valuable assets. Modern data and analytics platforms that can address the Extreme Data Economy are a critical component of successful innovation strategies. Investing in an insight engine, with GPU-enabled advanced analytics and AI,is the place to start.
About the author
Dipti is the VP of product marketing at Kinetica and has over 15 years of experience in relational and non-relational database technology. Prior to Kinetica, Dipti was VP of product marketing at Couchbase where she also held several leadership positions, including head of global technical sales and head of product management. Previously Dipti was on the product team at MarkLogic, and at IBM where she managed development teams. Dipti holds a Master’s degree in Computer Science from the University of California, San Diego with a specialization in databases, and an MBA from the Haas School of Business at University of California, Berkeley.