A decade of Cloud adoption in Banking: Living up to its promise?

About a decade ago banks started adopting cloud technology. At first, they were very careful, and it included lengthy discussions with conservative regulators and anxious security officers. Later banks took significant steps to leverage cloud technology. However, does the cloud deliver on its promise of lower infrastructure costs, better integration capabilities and improved time to market?   

The financial sector is expected to spend $500 billion globally on technology by 2021, and a signification portion of that will be funneled into cloud adoption as reported by Bloomberg, mainly through the support of cloud providers such as AWS, Azure and GCP.  

But reflecting on the last decade, many of the advisors and decisionmakers in banking I talk to are not necessarily satisfied with the current result. They see a slightly lower cost base, mainly as they outsourced hosting of their existing applications from the servers in their basements to a cloud provider. 

That said, the cloud promise is kept in the categories of more generic applications, such as collaboration tools like Microsoft Teams and Google Meet which saw an unprecedented rise in usage up to 30 times during the Covid-19 crisis. This scaling would not have been possible without cloud technology. But the main difference in these cases is that the new adopted technology was built for the cloud, instead of moving the existing technology to the cloud. 

Yet for mission critical technology like core banking that facilitates every transaction and financial product for a bank, cloud migration efforts hardly lead to improved functionality, faster upgrading processes or better integration capabilities. An important reason for this problem is that the technical foundations of these antiquated systems are from the pre-internet era. Back then changing products or expanding services was a multi-year process, which isn’t surprising considering that 43% of the banking systems are built on COBOL, a computer programming language that appeared in 1959.  

Today, new financial services are introduced faster than ever before. FinTech’s and big tech benefit from an absence of legacy systems, which allows them to invest in the latest technology and customer experiences rather than just keep existing systems ticking over, as acknowledged in a 2020 report on decreasing banking profitability by EY. In terms of time to market, perhaps weeks or months are available to follow and compete against the new players when they put new propositions or services in the market, but certainly not years. Here we see a crucial difference between organizations that use cloud native technology and the banks that are bound to their legacy systems.  

Within the legacy systems all changes are heavily dependent on each other. The effort to move existing technology to the cloud does not change how they were created: Heavily intertwined applications that can only be changed or upgraded when every relevant other system in the landscape is aligned. When these dependencies are not aligned well, change fails miserably, leading to failed projects, or even worse, data leaks or blackouts. And while required change grows rapidly every year, the ability to execute does not step up now that the legacy technology is hosted in the cloud.  

To approach this hurdle, adoption of cloud native technology is effective. It is shaped inherently different as every cloud native application is made up from different microservices, and these independent components interact with other microservices through APIs. This can be an API inside the bank’s technology landscape, or an external API to consume external services. Each microservice can be upgraded or changed by small autonomous development teams, without alignment. Not being dependent on other teams accelerates time to market significantly, served by continuous integration and continuous deployment (CI/CD) which is central to cloud native technology.  

Through CI/CD, teams can update their code up to multiple times per day supported by automated testing. This can and arguably should be carried out during business hours, as the proof is in eating the pudding. If testing does not turn out to be successful, it is a simple matter of rolling back changes. This is a crucial departure from big bang updates that often only happened once or twice a year, where entire IT departments work multiple weekends and CIOs have sleepless nights.  

Looking back at the first decade of cloud adoptions, banks are not to blame for missing out on the full cloud promise. Simply put, applications such as cloud native core banking were not available. And in this decade banks learned that updating legacy technology still equals less flexibility and vulnerability in the long run, whether it is adjusted for the cloud or not.   

Moreover, the future does not predict any simpler requirements in the technology agenda of banks as they are constantly pushed by new consumer expectations and increasing regulatory requirements. What gets simpler is the way to beat these challenges: Modern banking technology with modern functionality is becoming available in the form of cloud native systems.  

Therefore, investing in systems from previous decades seems rightly a dying trend. Banks frankly need to become as flexible, scalable, and accessible as Amazon, Netflix, or Uber since consumers are increasingly open to alternatives for their financial wellbeing and for banks to incorporate tools for regulation and risk requirements. Now that the first movers of banking specific technology are here to cater in cloud native fashion too, it is a blessing for banks that they spend their time and effort wisely.