Like many industries, banking is under pressure from digital disruption and from digital disruptors (often referred to as 'fintechs').
From the 2017 state of Strategic Digital Banking report:
- According to Gartner and IDC, by 2018, banks and financial institutions’ clients will access and contact their banks mainly through mobile devices
- Banks and financial institutions CIOs’ main concern to 2017 and beyond is what will be done with their companies’ branches
- According to the 2016 MX Consumer Survey, clients find it more important to have an easy digital banking experience (67 percent) rather than a friendly teller or staff (33 percent) when choosing where to open an account
- According to the 2016 MX Consumer Survey, banks and financial Institutions clients are now more likely to access their banks in a more impersonal way, and only 19 percent use a personal method (branch or call center)
Furthermore, the report emphasises that the 3 main priorities for banking IT executives in 2017 are:
- Redesign or enhance the digital experience - 71%
- Enhance data analytics capabilities to identify customer needs - 50%
- Find ways to reduce operating costs - 41%
We think that there are 2 particular challenges to focus in on in the banking industry based on the conversations we have with our banking customers:
Some of the banks we work with are hundreds of years old - the information technology in an organisation will not be hundreds of years old, but is likely to be tens of years old, and will include very critical systems. Integration issues also come into place with older code and some of these legacy systems have been architected in a very monolithic manner - creating a technology issue, which will make something like DevOps very hard to do.
Problems float around people and technology, but there are additional issues centered around people related to culture. Banking organisations are very large and they're really trying to stay up to date with fintechs, but are having a hard time because there are a lot of people in these organisations and at the end of the day they are human - they have behaviors, have learned to do things in certain ways and follow certain processes. As humans, we sometimes build systems that aren't as functional or as healthy as they could be and have to deal with work cultures driven by fear and blame.
An additional factor related to the culture problem, particularly in bigger banks, is outsourcing, which seems like a great idea in saving lots of money - we've seen cycles of this where things has been outsourced and then another cycle has been made, and the outsourced provider ended being pushed down to the ground even more, creating a supplier (rather than partner) relationship instead. Of course, a lot of outsourced companies are not in the place where the original company staff are (popular places for outsourcing include India and Eastern Europe), creating additional geographical and timezone problems.
Regulations strongly affect companies and lots of people are facing these problems. However, the biggest problem is around segregation of duties - we've now got DevOps and want to erase the lines, bringing people back together, and it might mean that our developers will have to have access to production systems, which segregation of duties prevents. We shouldn't panic because of this, as it doesn't mean we can't do DevOps - it means we have to think carefully about how we do it and understand the processes before we try to automate (a key part of the solution to this problem) them.
What are the leaders in banking doing with DevOps?
Topo Pal, Director & Platform Engineering Fellow regularly talks his DevOps experiences at Capital One and is a frequent speaker at DevOps Enterprise Summit in San Francisco. He says that:
“At Capital One, we believe that while tools, automation and collaboration are very important, a continuous feedback loop is critical to DevOps success.”
Pal describes the benefits of Capital One's DevOps initiatives as:
- 100s of code commits per day, from a group of applications. It should make you think about the frequency that YOU have at the moment, what frequency YOU need (strengthened by value outcomes through your customers)
- Integration from once a month to every 15 minutes (large amounts of changes infrequently will significantly increase the chances of things going wrong)
- QA from once per month to 4 times per day (shifting left and getting tests done early)
- Deployment from manual to completely automated
- Production release from monthly/quarterly to once per sprint
In short, Capital One set up a web deployment tool, which automated the whole process of deploying a web server in the correct configuration by pressing a single button – saving a lot of time and reducing risk significantly.
Ranger4 are DevOps evolution experts based in central London.