By Stuart Tarmy, Global Director of Financial Services Industry Solutions, Aerospike
It’s a generally accepted truism that data is a key driver of business success. In fact, the phrase ‘Data is the new oil’ was first coined in 2006 by Clive Robert Humby, a British mathematician and entrepreneur. However, this phrase is incomplete, as oil by itself is not useful until it is captured, refined and put to purpose, e.g., changed into gas, plastic or chemicals. Similarly, to be fully exploited, data must also be ‘captured, refined and put to purpose’. In today’s competitive environment that requires an enterprise grade data platform that can operate in real-time, at petabyte scale, be reliable, available in a hybrid model (multi-cloud, on-prem) and provide a low total cost of ownership.
Real-time, and how to achieve this, is critically important. A key concept here is that to develop best-in-class, real-time applications such as fraud, customer360, compliance or risk management, you need to balance the competing needs of utilising the most sophisticated algorithms (often based on AI or more complex neural nets) across the largest amount of relevant, non-correlated data available, and process this in real-time (often less than 30 milliseconds) for a pleasing customer experience.
Given that customers now expect a rapid response and a personalised approach from the financial companies they interact with, real-time data has never been so important. A good example of this is how real-time data is being used to detect fraudulent customer transactions and develop models to predict credit risk.
Processing large amounts of data in real time and delivering insightful analysis cannot be done without a modern data architecture. As part of digital transformation processes, investment needs to be made in the appropriate technologies and systems.
Graph analytics for dealing with fraud
Combating fraudulent activity is a constant, and growing, challenge for financial institutions and one that is high on boardroom agendas if they are to reduce the risk of financial and reputational damage. Some organisations have opted to adopt graph databases, each one of which consists of data elements and the connections between them. The data elements represent a customer or an account, and the connections are the relationships between these entities, which could be social connections, identity or transactions. A graph database works with a real-time data platform, which allows the company to analyse the relationships between the data elements and identify unusual, or suspicious patterns, such as multiple accounts being opened under different names, but with the same IP address.
PayPal is a great example of how to use graph analytics to prevent fraud. It has a bespoke solution which is capable of analysing millions of records within just 20 milliseconds. This can identify fraud risk, allowing the company to put in place prevention processes, thus saving itself and its customers millions in fraud losses.
Document data stores and credit risk management
For the kind of unstructured data that occurs in credit risk management, document data stores are gaining popularity. These document databases collect data from credit bureaus, financial institutions and social media to name a few, and can then provide a detailed overview about whether a borrower is credit worthy. The data can be analysed in real time using machine learning algorithms to identify patterns, trends, and potential risks, so action can be taken to mitigate against them. Risk models are created which will assess a potential creditor’s ability to pay based on their credit history, income and current employment status. If a customer is experiencing financial hardship, a financial services company can act before they default on a payment. Predictive analytics can also be used to develop models that identify potential credit risks before they materialise, which allows credit limits to be adjusted or alternative payment plans to be put in place.
Document data store for powerful personalisation
Any customer-facing operation understands the importance of personalisation when it comes to building strong customer relationships. Financial services companies are striving to enhance personalisation by aggregating data from various sources in real time, including mobile and location-based services.
A document data store is optimised to manage this data in real time and analyse it to build an accurate picture of their customers’ financial behaviour. Using AI and machine learning they can offer tailored product recommendations, personalised financial advice, and targeted marketing campaigns.
Every day the financial services industry generates massive volumes of data. A modern real-time data architecture is essential to help them build best-in-class customer solutions. By analysing customer habits and preferences, personalised product recommendations can be made that better suit their needs and preferences. Personalisation can also lead to customised pricing, credit scoring, interest rates, and loyalty programs, speed up customer onboarding, and predict and prevent customer churn. By using these techniques, financial institutions can beat their competition, enhance the customer experience, improve revenue and grow market share. The alternative is to become less competitive and less relevant to your customers.
