Capital One closes its data centres and goes all in with AWS

Capital One has exited all of its data centres, moving all applications and systems to the public cloud with AWS. The change to the company's operating model is particularly significant for a financial institution, where many banks still struggle to move away from mainframes and legacy architectures. 

Capital One was founded 25 years ago by Richard Fairbank, who still runs the company today. The bank specialises in credit cards, auto loans, banking, and savings accounts, and has close to 50,000 employees. It is one of America's largest financial institutions and brought in close to $30 billion in revenues last year. 

Speaking at AWS' virtual re:Invent event this week, Capital One's SVP of Technology, Chris Nims, explained that almost 8 years ago the bank recognised that it needed to change how it delivered services to its customers if it was going to compete in a digital-first world. He said: 

Capital One recognised early on that the winners in banking of the future will be great tech companies with the risk management skills of a leading bank. They will build great customer experiences, delivered in real time, through software, data, and algorithms. We sought to completely redefine who we are as a company, to build a technology company that does banking, instead of a bank that just uses technology. We needed to become great at building software. And we needed the top engineering talents to do it.

Nims said that Capital One sought to reinvent its technology operating model, to enable agility and innovation, so as not to just keep up with the pace of the market, but to set the pace of the market. This involved adopting agile delivery practices, using modern architectural standards (RESTful APIs), making use of micro services, and building an open source foundation. Cloud first was the new mantra. He added: 

Today, much of the innovation at Capital One is homegrown and built first as a proof of concept. We're working on unique challenging technology problems that will ultimately benefit millions of customers.

The challenge is banks aren't built to deliver those types of capabilities. We believe that delivering transformative machine learning products across the business will allow us to interact with our customers in a more natural seamless and accessible way, by moving to the cloud. We've built the foundation for the bank of the future. We can now build experiences that are powered by more data available in real time, with algorithms and AI.

Talent is key

Central to Capital One's transformation over recent years has been its thinking around talent acquisition and development. Nims said that the bank had to comprehensively reimagine its technology, how employees work, but also its people. 

Capital One knew that by bringing the development of its applications in house, it had to become great at building software, which meant recruiting top engineering talent. Nims explained: 

We took a multifaceted approach to raising our talent bar, our recruiting efforts are focused on both experienced and college hires. We recruit from the top engineering schools across the country. We have large pools of summer interns, with the hope that many of them will receive offers before they return to school for their senior year. We more than doubled our technology team over the last eight years, today we have more than 11,000 technology associates. 85% are engineers.

Capital One has also adopted a strategy of continuously investing in its people, according to Nims, where it has launched a tech college inside the company, to support associates and "sharpen their skills". He added: 

Our focus areas include software engineering, security, cloud, mobile data, agile and machine learning. In addition to hard skills tech, the college provides training around soft skills as well.

In addition, we're rapidly moving towards every software engineer having a certified secure software engineering certification. We have an AWS navigator initiative, where we have AWS account team members that lead tech talks, and workshops tailored to Capital One, in order to train up our technology associates in using AWS. 

A new operating model

As noted above, after years of intentional effort across the whole company, Capital One has now reached the milestone of closing all of its data centres and now operates fully using AWS' public cloud architecture. During this migration process, the bank has also rebuilt 80% of its applications to be cloud native, using modern cloud architectures, such as micro services and RESTful APIs. Nims said: 

We are truly all in on the cloud with AWS, as our strategic partner, and for us there's no turning back. 

When we decided to go all in on the cloud, we chose to solve the hardest problems first. One key early step we took was to establish a cloud governance function, consisting of risk managers and cloud engineers to establish capabilities and controls that would keep us well managed, as we migrated our applications. 

Some of the other primary enablers for our transformation included long term strategic planning, and a focus on education. The most challenging part of a large scale transformation for a company in a highly regulated industry is building the technology governance structure and the culture. 

On the technology front, many financial institutions have structural barriers, such as legacy systems that have been in use for over 20 years, and governance is foundational. It takes time to develop processes that enable efficient code reviews, reviews that need to be done by other developers, by legal, by information security, and other stakeholders. Also, consider culture, traditional banks need to challenge the perceived risks, such as loss of intellectual property competitiveness, or productivity.

Nims said that with proper discipline and optimisation, applications are able to run in the public cloud for less cost than a traditional data centre. Not only this, but it has enabled Capital One to take advantage of the innovation and new services being introduced by AWS, at scale. 

By adopting DevOps and operating in the cloud, Capital One has also enabled more frequent releases compared to when it was operating on premise. For example, the bank has gone from monthly or quarterly application releases, to releasing code into production multiple times a day. 

Between 2016 and 2019, Capital One has increased application changes by more than 300%. It has also seen a reduction in transaction errors and critical incident resolution time both cut by half. The bank can now build a new development environment in minutes, compared to months previously. 

Nims said: 

We comprehensively reimagined our talent, our technology and how we work. This transformation was a significant effort that required every part of the business to come along on the journey. This is really about how we change our entire business model, and our DNA as a company. 

We declared the cloud as our destination, which began with the declaration that anything new, could only be built in the cloud. We didn't just do it in pockets, but instead we made the declaration that we go all in.