Chris Skinner argues the case for more open source development in financial services.
Many banks are attracted to Open Source but, in the words of Will Smith, "Wiki, wiki, why, why?". Surely you give away your competitive advantage when you offer your ideas to all and sundry.
It amazes me the number of times I have encountered discussions about Service Oriented Architectures (SOA), Open Source, Websphere, Components, Wiki’s and so on. It amazes me for two reasons. One is that half the time I have no idea what these guys are talking about and two, many of us were talking about Object-based Developments over a decade ago. So what’s new? SOA, and its associated pieces, is nothing new. It is just that it is now mainstream, having been heavily promoted by IBM, Oracle and other tech players.
By way of example, Finextra recently published the results of a survey by BEA Systems of over 1000 developers across Europe with the headline that almost 75% of them are either developing, or expect to be developing, SOA applications in 2005 European developers say SOAs set to soar
When questioned about the most useful technologies in SOA, 28% said Web services, followed by 18% pointing to languages such as BPEL and BPELJ that help the automation of business processes.
Therefore, it is not surprising that JP Morgan Chase, Credit Suisse First Boston, Deutsche and a variety of other banks have recently been pushing out their open source developments.
What do they get out of doing this?
For most, faster, cheaper software development that is more resilient and reliable than anything they could develop internally. In addition, banks cannot afford to continue to employ masses of software developers to build bulk standard systems, when other markets are willing to give their skills for free in order to share the rewards.
By way of example, take the Openadaptor (www.openadaptor.org
) built by Dresdner Kleinwort Wasserstein.
Developed with a budget of around €5 million, Openadaptor was designed to reduce systems integration development times, typically from 5 days to as little as half a day. Launching Openadaptor as an open-source application in 2001, Dresdner soon found developers worldwide giving of their time for free, driving out errors and streamlining the coding to be as efficient as it could be. The result is the base investment of €5 million generating a raft of input worth exponentially more.
Now that is all well and good, but the question many ask is "doesn’t this diminish competitive advantage?", and that’s where it gets interesting.
For example, when you look at what most of these banks are doing with Open Source, they are deploying it for standard bits of work, like adaptors to connect to networks, file format converters and base security encryptions.
As we get into other areas, such as core systems developments, you rarely see an Open Source development in sight. After all, who wants to give their core systems to the marketplace?
Well, surprisingly, my answer would be "Why not?".
TowerGroup estimate that banks globally invested about $180 billion in core systems in 2004. Over 85% of that investment was on core systems maintenance. In other words, banks are spending over $150 billion a year just to keep their systems running. Nothing innovative or new – just keeping them running. And a lot of the costs involved in keeping them running are related to compliance with new regulatory mandates whilst coping with a cumbersome array of interfaces.
TowerGroup have also found a multiplier effect when banks rationalise these systems. Many banks find that, as they re-engineer legacy systems, they release up to $7 in business process expense savings for every legacy IT dollar saved. That is through the savings made in hardware and licensing, the efficiencies gained through new capabilities, including SOA, and the re-engineering of business processes associated with such change.
That is why banks will increase spending to $202 billion by 2010. In the process of increasing spend, core systems maintenance will reduce overall to less than 70% of spend, with 17% on core systems replacement (currently eight per cent of core systems spending) and 14% on new core systems (currently only six per cent of spend).
In this context of core systems renewal, we need to ask ourselves a basic question: "What is it about our systems that make us unique?"
There is very little about a bank’s core system that provides uniqueness, or even competitive advantage, anymore. For example, if a core system made a bank that unique, then banks using Sungard, Misys, Infosys, i-Flex or other core systems solutions wouldn’t be able to compete. Yet they do. And banks using solutions claim that they cannot change them too much because it would stop them from migrating and upgrading to future releases.
So, we have banks using virtually vanilla back office core systems from solutions firms. Where is the competitive advantage in that? And yet, the same banks are concerned about moving to open source systems.
Why? Typically it is either for the reason stated – it will damage their competitive differentiation – or because of other factors, such as concerns about a lack of reliable support services, a lack of regular updates, a loss of control, the unpredictable total cost of ownership and intellectual property concerns.
A common concern is around security, for example, and that using externally sourced open coding may subject the bank to easier hacking through Trojan horses or similar viral attacks.
These are the normal concerns for any technology that is new and many of these concerns will disappear as leading banks prove the technologies work and that these concerns are unwarranted. For example, the security issues can be alleviated through suitably diligent code inspection and certification, aligned with inline security tools.
So that brings us back to our base question: "What is it about our systems that make us unique?"
The answer is virtually nothing. There is very little these days that is truly unique to a bank operation, and certainly it is not their systems (unless you are Citigroup’s bond trading team).
What really makes a bank unique is its organisational structure, products, services, breadth, depth and culture. And that is all about people, not systems.
For example, banks are so deconstructed – through sourcing strategies to drive out cost through outsourcing, by moving development offshore, by buying vanilla solutions to run operations, by using lowest cost approaches to systems – that the real uniqueness lies in how they put the pieces of the puzzle together. Core systems, networks, infrastructures, front office, hardware and people are all part of the puzzle. And open source is going to be a critical area for future developments in driving cost out of that puzzle. A bank’s ability to use open source as part of their puzzle will be a part of their competitive differentiation therefore.
The message is that you will reconstruct systems over the next few years. Systems will be deconstructed and then placed through a sourcing strategy for renewal. The sourcing strategy may well move core systems from insourced to outsourced to open sourced. The reason is that banks will need to do whatever they can to create the highest margin for the lowest cost. By all means, therefore, treat your systems as objects but treat your people as more than just objects. After all, it is your people that make the bank unique. Chris Skinner is a director of TowerGroup and founder of ShapingTomorrow.com.
Web links: www.towergroup.com
Author's email: Chris Skinner