Digital transformation in banking and brokerage

The core data that supports the business—such as accounts, positions, or client demographics—is generally stored in a monolithic back-end system. Data is freely accessible within the system, for instance, to any application on the mainframe. However, a change to any one component necessitates testing all aspects of the system, so changes are typically limited to once per quarter. This allows time for integrated testing and to involve multiple teams. Given this, and the advanced age of such systems, accessing their data from more modern systems within the enterprise can be strategically challenging for the business.

Agile, product-oriented transformation requires a fundamental shift to a new paradigm of development teams having autonomy to determine their own workload. They decide when they’ve completed a task (for example, product modernization) and when they’re ready to start the next task (for example, the next product or enhancement). This methodology is “pull” rather than “push.” This can cause cultural dislocation at many firms since it requires concerted effort, retraining, and perseverance to overcome cultural resistance.

Outsourced providers can reduce fixed costs and allow firms to focus on their bread and butter: building client relationships and offering financial services. However, to a modernized, agile product, they can seem like stones in the path of a swiftly flowing stream—everyone must adapt around them. This is because they themselves generally run cost-efficient yet cumbersome mainframe systems with limited scope for change or customization. An outsourced provider linked to the in-house mainframe for processing is a relatively inflexible arrangement that can be doubly challenging to pick apart when changes are required.

Because data and processes within a mainframe can interconnect freely, their interdependencies tend to only grow and become more complex over time. Given that these systems are 30+ years old, untangling their interlinked applications into discrete products can be quite intimidating. A change to one process can affect many others, necessitating laborious testing and creating a chicken-and-egg problem, with teams asking, “How can we modernize our product when we depend on product X, which is not yet modernized?”

With the advent of public cloud, the data center and its supporting staff—the ultimate horizontal organization—is no longer necessary. Now, a small team can make decisions to scale up and down its infrastructure on the cloud, tailored to the needs of its product alone. Depending on the efficiency of the data center and the computing capacity it provides versus business need, this may not represent a massive cost savings—but does benefit from converting from a large, fixed upfront investment to variable cost, and paying only for what is needed.

Even in the relatively conservative financial services industry, public cloud infrastructure is rapidly gaining traction as the method of choice. In 2023, 48% of banking and investment services CIOs surveyed by Gartner reported having deployed workloads to cloud, and 44% planned to deploy cloud workloads in the next 18 months (Gartner, June 2023). However, experience across financial services reveals that although organizations may have cloud deployments, the bulk of their processing remains on monolithic mainframe applications that date from the 1990s or earlier.

The cost of the foundational modernization required to address these systems is high, but the cloud migration trend is gaining steam: In 2021, 29% of banking CIOs surveyed stated that they intended to reduce data center technology funding over the coming year. By 2023, that number has risen to 47%; while in the same year, 53% of CIOs expected to increase cloud technology funding.