Sponsored Financial services companies can't sit on their hands when it comes to technology. Challenger banks and other financial technology players are rushing ahead with new services that threaten the incumbents' competitive advantage. Established banks need to do more than build these new services; they need to make them fast and reliable enough to satisfy a picky client base that won't tolerate outages.
Modern development and database tools can help them achieve these goals. Redis Labs is positioning its in-memory NoSQL database as a go-to tool for banks planning a digital refresh. It offers several qualities that support financial institutions' technical requirements as they prepare their digital transformation strategies.
Financial services are at a turning point
Financial services companies face a simple challenge: transform or die. Their customer bases are changing and their relationships are moving increasingly online. Branch-based banking is dying away to be replaced by mobile apps. Non-profit financial services think tank BAI Banking Strategies found last year that millennials' preference for online banking jumped from 18 per cent in 2017 to 28 per cent in 2018.
Banks are struggling to keep up by as they follow their customers online. In its survey, BAI found that 57 per cent of customers wanted to open accounts online, but only 34 per cent of them did so. That's because almost half of the financial institutions it surveyed couldn't open a customer's first account online.
As banks flounder online, several factors are forcing their hand. One of them is regulatory. The EU's revised Directive on Payment Services promoted the concept of open banking, in which large financial services companies must open up access to customer data for licensed service providers. The UK followed through on these rules with an order of its own, ushering in a new era of open banking in the UK. Banks must open up their data to others, and do it securely.
Another factor propelling digital transformation is the pandemic, which has further accelerated this rush to online banking. J.D. Power found that 29 per cent of consumers were using mobile banking apps more in early April, during the first couple of weeks after lockdown. By the end of September, that had grown to 50 per cent.
Don't use digital duct tape
Banks can embrace digital transformation in two ways. The first is reactive (what Redis calls the duct tape approach). In this scenario, they piece together front-end solutions that offer some subset of banking functionality to users. That works - up to a point.
Banks relying on digital duct tape might be able to offer some services online, but not all. Perhaps customers can open an account but can't arrange a loan. Another danger is that they can't connect data from these front-end systems to the back office, making it difficult to handle things like analytics and fraud detection at scale. Or, banks do connect their front- and back-office systems, but the legacy infrastructure can't keep up.
The alternative approach is more strategic. Rather than taping together a quick front-end solution and making a best-effort, piecemeal attempt at back-end integration, banks can take the extra step to coordinate their entire infrastructure. Rethinking digital services at this foundational level is what digital transformation really means.
Operating in real time
Redis Labs argues that financial services companies need three technical qualities to pursue this ideal. The first is real-time operations.
Banking customers are used to modern interactive applications that deliver data instantly. They demand the same from their financial institutions.
Banks also need real-time operations to support urgent use cases such as fraud detection, which is becoming a more urgent issue. Lexis Nexis found that the cost of financial crime compliance rose by a third in North America between 2019 and 2020.
Banks must verify client IDs across multiple systems in real time. They must do that in the face of synthetic fraud, in which criminals craft completely fictional identities. That activity is hard for traditional fraud detection to track down because it is more complex with less empirical data to draw upon.
The duct tape approach to digital transformation, modernising front-end systems while leaving the back end to chunter along as best it can, is bound to fail. Legacy systems keep latency high, making it hard to keep up with transaction volumes. Those 30 year-old COBOL applications probably aren't going to cut it - at least, not on their own.
As an in-memory database, Redis provides high-write throughput with subsecond latency, making real time ID verification more tractable.
Modern data models
The other technical quality that banks need for digital transformation is the ability to support multiple data models. As it stands, many legacy banking systems are not digitised.
Even if back-office banking processes are digitised, they hold data in their own formats. Many of them don't talk directly to each other, making it difficult to create the end-to-end workflows that support digital transformation. Banks resort to relaying that information manually or using brittle ad hoc integrations that make systems inflexible in the face of unprecedented change.
Legacy data models also have their roots in last-century tabular relational models. Today's business applications see the world differently. Graph databases map relationships between people, while geospatial data maps trends across different regions. And time-series data enables you to analyse data using time bounds, for example, to see happened between January 1st and January 3rd, Kyle Davis, Technical Marketing Manager at Redis Labs, explains. “You can also get granular, into seconds, sometimes even milliseconds. You can also separate your data into time units, to see what happened on an hourly basis.
Banks can modernise their infrastructures to support these new data and application models in two ways. The first is to rip and replace their existing applications, starting again. That wasn't especially attractive even before the economy reached its current post-pandemic crossroad. Now, financial institutions will be holding onto their budgets tighter than ever.
Surround and extend
The second approach is to surround and extend, building new functionality around the existing applications and extending their functionality, decoupling monolithic apps by creating a layer of microservices around them.
Redis Labs has spent the last few years building native support for different data models into its database engine. This modern data model support covers graph and time-series data among many other kinds, including JSON-based key-value pairs, bloom filters for probabilistic analysis, and search for full-text search capabilities.
The company's Redis Gears data processing framework enables developers to access these data structures and modules together, transforming data between them across their applications. This positions the product as a unifying layer. It can interact with existing data structures in a bank's legacy application base and integrate them with new data structures in modern applications, enabling developers to build new functionality that scales.
One of the data modules underpinning Redis is RedisAI. Together with graph-based data, this enables developers to create more modern fraud detection systems that move beyond rules-bases algorithms. They can use AI-based models to spot not only the fraud that they already know about, but the fraud that they were not aware of.
Graph-based models and tabular AI models including Bayesian and linear regression create explainable results. Banks can use these to understand why algorithms highlight applications as potentially fraudulent.
Financial institutions and the fraud detection companies that serve them must analyse transactions across many locations in real time. Companies that use Redis to serve those needs include PayPal's Simility, which provides cloud-based machine learning-driven fraud detection to clients. It reduced IT costs by up to 30 per cent and delivered applications up to 30 per cent faster by using the NoSQL database with its embedded data modules.
Hardening the enterprise
The final quality that financial services organisations need as they level up their digital capabilities is reliability. In 2019, regulators told the UK's Treasury Select Committee that banks risked causing more outages as they rush to catch up with fintech companies.
Back-end systems are letting front-end systems down, explained Lyndon Nelson, deputy chief executive at the Bank of England's Prudential Regulation Authority, at the time. "They still use code back from the 1970s on some of these systems, and they've just built on top of them," he said.
Redis Labs solves that problem with a writable in-memory database that buffers increased demand, protecting back-end legacy systems. Redis also uses a peer-to-peer replication system that supports an active-active model, meaning that financial services companies can fail over very quickly between nodes, even when located at alternative sites. This contrasts with the active-passive model, in which a primary server replicates to a backup system that is on standby. It takes longer for active-passive systems to recover from disruption.
Redis also supports scalability through partitioning. This splits data across multiple cores and computers, allowing for very large in-memory computational loads.
The financial sector has embraced this open-source database project as it tackles these digital transformation challenges. Three of the four major credit card providers are Redis Enterprise customers, and other financial institutions such as Deutsche Börse are using the database in production to support their performance and functionality needs.
Financial institutions might not be ready to abandon applications built decades ago, but they can expand and enhance their existing portfolio to offer new customer experiences. A resilient, responsive, and feature-right infrastructure is an asset you can take to the bank.
Sponsored by Redis Labs