For financial markets firms, efficiency is becoming as important a differentiator as speed and scale. As a result, firms are delving deeper into predictive analytics to realize faster time to value and improve operational performance and decision outcomes.
Technologies that speed pattern recognition in ever-growing data sets – including big data – provide deeper insights into trends and behaviors that can translate into millions of dollars in opportunities and costs. Firms are increasing investments in predictive analytics amid an explosion in data sources and instruments to exploit them and a torrent of new regulations that are squeezing revenues, margins and profitability.
Four growing use cases where predictive analytics maps to return on investment (ROI) and risk management objectives include:
- Optimizing capital deployed. A deeper understanding of the link between portfolio risk/reward models and the impact of events on capital at risk and market liquidity can help market participants anticipate volatility and exposures under various scenarios. Without a real-time view of potential risk and return outcomes optimized for different scenarios, risk managers cannot understand or integrate the performance of capital across different trading desks or asset classes;
- Ensuring regulatory compliance. Predictive analytics are being used to anticipate, detect and prevent disruptions caused by trading errors, improper systems oversight, or other compliance violations. Firms also seek to become more predictive to root out rogue trading activity and improper client behavior, such as money laundering schemes. Predictive models can also supplement a firm’s compliance team to identify, monitor, and track regulatory changes across jurisdictions, enabling firms to anticipate risk and prepare relevant internal controls.
- Gaining operational intelligence. More firms are using predictive analytics to monitor, detect, and analyze events across all the enterprise’s operational data. The technology acts as an early warning system, helping financial institutions identify and reduce outages and avoid latency-impacting events, thereby mitigating risks, adverse impacts to revenue, and damage to reputations.
- Improving customer satisfaction and loyalty. As financial firms shift toward becoming more customer-centric from product-focused, they can go beyond understanding risks surrounding credit worthiness and spending patterns. New data types, such as online customer reviews, click streams and customer sentiment analysis can be used to help firms better understand customer behavior and how to engage with them. This data can then drive product innovation with tailored offerings.
Predictive analytics is an iterative process that begins with an understanding of the question the user wants to answer. By exploring the relationships among different variables using correlation analysis, users can build sophisticated mathematical models that can cut through the complexity of modern computing systems to uncover previously hidden patterns, identify classifications and make associations.
Success with predictive analytics relies on choosing the right data set, assuring the quality of the data, and validating models and algorithms used to analyze the data. But financial firms have been challenged by data management issues caused by incompatible systems and myriad data silos. As they now consolidate data centers and focus more on efficiency, a unified approach to data management is becoming a priority. Firms recognize the substantial opportunity to achieve a better return on data assets (RDA).
A Unified Approach to Predictive Analytics
Information sets have become too vast to understand all of the interactions and dependencies necessary to manage risk and return effectively. Advances in underlying technologies now allows for real-time analysis on massive sets of both structured and unstructured data. The ability to integrate historical data with newer sources of information to predict market trends and rapidly implement strategies is becoming the primary differentiator.
Firms are now incorporating text and social network analytics to discover relational patterns and sentiment trends. For example, more trading desks are beginning to include tweets into trends analysis, which are then incorporated into models and algorithms. Information gleaned can then be stored and used in predictive models to anticipate what might happen to various markets or securities when a similar trend occurs.
Other ways financial firms are harnessing big data are by correlating and analyzing data from ATMs, online banking, call centers, investment portfolios, market data feeds, mobile banking, and social media sites to better understand risk and pricing as well as customer opportunities and life cycle.
A formal data governance program that involves all relevant constituents – from IT and user groups to security and compliance – is the first step toward dynamic data integration. Since data quality, master and metadata management are integral to governance and data management, Tech-Tonics advocates that financial firms treat all of their data as big data.
Implementing a unified analytics platform that incorporates and aggregates all of the metrics financial firms rely on to run their businesses reduces computational complexity across instruments and data silos. It facilitates dynamic monitoring of positions, markets, events and capital flows.
Dynamic data integration and management allows calculations to be made and disseminated in real time. A unified analytics platform provides the tools to better monitor traders and portfolio managers, and can act as an early warning signal that informs decision-makers to act immediately. The objective is better, faster and compliant decisions that are based more on data and less on expert judgment.
As the amount of data financial markets firms are collecting and storing for real-time analysis continues to expand, related infrastructure and maintenance costs have also risen rapidly. Tech-Tonics expects that firms will begin to explore cloud-based data integration, management and analytics solutions over the next two years. Proprietary models and algorithms will remain in-house for the foreseeable future. However, with the scale, speed and cost benefits that cloud solutions offer, expect firms to take a hybrid approach to adopting a unified platform.
