A brief look at real-time analytics
Before we dive deeper, let’s discuss what true real-time analytics is. You may assume that your data is updating in real-time, and may even be using a nifty visualization, but that doesn’t mean your analytics is real-time.
See, many finance organizations use Hadoop or parts of the Hadoop ecosystem, like HBase, Spark and others. These are great at processing big data affordably, by using batch processing (in which it processes all data at once, in a batch.) However, it wasn’t designed for analytics. Even though it has been retrofitted, running actual complex analytics on Hadoop is typically slow and cumbersome – a far cry from the near real-time requirements of modern finance.
Instead, data is updated in intervals – providing you with snapshots, but not a continuous stream. It’s a bit like trying to drive a car but only opening your eyes every so often. You can easily be caught off guard by quick changes in your environment.
Near real-time and real-time analytics, on the other hand, provides a near constant stream of data. This allows for immediate action in response to significant events. Think sales leads, orders, customer service calls, and even fraud prevention. The relative length of this update interval is what defines if it’s real-time, near real-time, or just “batch”.
Fraud is a$190 billion problem, and while financial institutions have taken steps to prevent it, it still prevails. While there are signs that fraud may be occurring, a financial institution often doesn’t have the data fast enough to address it.
Even with big data, fraud was cut down on, but one or two fraudulent charges can almost always go through before the data updates and reveals fraud.
With real-time analytics, however, fraud can be detected within seconds. This can help cut down on money lost, saving banks millions a year. Banks that don’t evolve to include faster analytics will see lower profit margins as a result.
With real-time big data analytics, a month-long fraud attempt can be detected by looking at a wider time-frame.
Determining whether a customer is qualified for a loan or credit card is still a chore. A number of factors need to be checked, and pulling in customer data can be a time consuming process. Not to mention that the data may also be outdated.
With real-time analytics, data including deposit information, customer service emails, credit card purchase history, and more can give you a holistic view of your customers. This in turn enables a bank to make a faster, more accurate assessment of risk.
Increased customer loyalty
Perhaps the most groundbreaking changes are occurring in customer facing environments like sales, marketing, and customer service. In these areas, real-time analytics makes a huge difference.
Service can be rendered faster and more effectively, new leads can be prioritized and upsold/cross sold on relevant products, and marketing can be more highly targeted.
All of this combines to result in anincrease in new customers and decrease in lost customers. In other words, banks that use real-time analytics are growing faster than those without.
So why the wait?
Until recently, the technology for big data real-time and near real-time wasn’t affordable enough to gain widespread adoption. The processing power and software simply weren’t there.
The big catalyst for near real-time analytics has been the creation of GPU databases which have made it feasible for small and large companies to analyze huge quantities of data in near real-time. SQream, for example, is a GPU database that can process hundreds of terabytes, which in-memory databases or frameworks like Spark can not do easily.
It’s time to adapt
Banks that don’t adopt real-time analytics will be sitting in the past, literally. They won’t have the data at their fingertips to make rapid decisions.
On the other hand, banks that do adopt real-time analytics will have the data necessary to be agile and nimble. They’ll respond to market changes nearly instantly, more rapidly respond to customer service issues, and acquire more customers (which are higher quality even) than their competitors. This all leads to faster growth, and a better bottom line.
What are you waiting for?