Oracle’s three modes of Progressive Transformation

I was able to attend Oracle’s Open World at the end of September, and although it conflicted with Sibos, it was an extravaganza. While there I sat down with some of the folks involved with core systems; they outlined the interesting way they’re thinking about progressive transformation (briefly, how to migrate core systems gradually; the opposite of a “big bang” approach). Oracle agrees with the consensus that a big bang for any sizable bank is going to be problematic. What interested me was that they outlined three different approaches for progressive transformation:
  1. Replace a vertical slice
  2. Replace a horizontal slice
  3. Create a new target state architecture off to the side
Without going into great detail, I’ll describe how Oracle has at least started the journey in three different banks around the world.
  1. Vertical Slice. Suncorp in Australia has started the process of moving off its Hogan core by focusing on unsecured lending; its next stop will be secured lending.
  2. Horizontal slice. KeyBank, based in Cleveland, announced at Open World that it intends to use non-core systems components of Oracle Banking Platform (“OBP”) to enhance and modernize its mobile and online channels. To be clear, KeyBank has not committed to a core transformation. The project is in its very early stages; it’s one we’ll watch with interest
  3. Target architecture. National Australia Bank’s new entity, UBank, is a digital-only bank that NAB created as part of its bank transformation using OBP. Its goal is to change the customer experience, and uptake has surpassed initial expectations.
Celent’s perspective is that progressive transformation (or whatever various name different vendors use for the same basic concept) is a way to purchase a real option as banks think about how to modernize their systems and accommodate the increased demands that digital access place on their technology. It lets banks begin a journey without committing them to course of action that might not be appropriate down the road as the world changes. Time will, of course, tell how successful each of these projects will be, but thinking about the different ways to approach a phased core transformation is useful for any bank with core on its strategic agenda (which should be…almost any bank).

The more things (in core) change, the more they remain the same…

On September 10 Forbes contributor Tom Groenfeldt wrote an interesting piece titled “Core Banking Replacement Remains Locked in the Future.” The article started by citing a 2003 Celent report, Core Banking Replacement Strategies: The Time Has Come. In the intervening 10 years, there have been only a handful of successful Tier I core replacements. Tom’s point wasn’t to make Celent look foolish – he could have picked any number of analysts who’ve been delivering the same message over the last decade – but instead to highlight the slow pace of core migration. We stand by our rationale of ten years ago – the need for flexibility, customer-centricity, real time, and efficiency, together with the added demands of mobile, all point to the need for banks to upgrade their antiquated core systems.  The big difference is that today a wholesale rip and replace effort isn’t the only game in town.  Progressive migration, middleware additions, and even starting new banks to serve, in part, as a migration destination are all strategies to defer (not avoid) an eventual core replacement. In Celent’s most recent core report, Global Core Banking: Steady But Unspectacular Growth, we highlighted the challenges, both technological and cultural, that large banks face in making the decision. Very few large U.S. banks will act in the very near term: banking systems are complex, change is costly and time-consuming, and for many, the competitive push has yet to materialize.  It’s no accident that Australia has seen two Tier I core replacements – once one firm goes, particularly in a concentrated market, it raises the cost of inaction for its competitors. Further, as is the case with many cases of change and innovation, the downside is huge and the upside is limited. Core replacement is going to be much more realistic for smaller banks in the near future. Until large banks have role models of successful pioneers to follow, they’ll be reluctant to act.  In our conversations with banks around the world, however, we hear that the pressure’s building to tackle the core, rather than tinker around the edges.    

And then there were four…Fiserv acquires Open Solutions

On January 14 Fiserv announced that it closed the acquisition of the equity of Open Solutions for $55 million.  At this size the deal was not subject to regulatory review.  The result: there are now only four major U.S. based Core Solutions providers: Fiserv, FIS, Jack Henry and Harland. Fiserv touted three key strategic benefits to the transaction:
  1. An expanded base of account processing clients,
  2. The “high quality and innovative technology” of Open Solutions’ DNA platform, and
  3. The wise use of Fiserv’s capital.
Overall, Celent views this as a win-win transaction for the firms. Celent  sees two key benefits for Open Solutions customers, existing and potential.  First, questions about the company’s staying power are moot: under Fiserv’s umbrella, Open Solutions is now financially viable. Second, Open Solutions, at least theoretically, gains access to Fiserv’s proprietary  complementary products such as Corillian and Mobiliti.  If Fiserv executes to its historical standards, the implementation will go well. As for Fiserv, it gets access to a real-time relational database structured to be charter- and geography- agnostic, together with multi-currency capability.  This gives the company technology to accelerate its international expansion, particularly with smaller community-based banks (Fiserv estimates there to be 40,000 such institutions globally). Questions that we’re interested in:
  1. How much did defense factor into Fiserv’s strategic calculus – that is, did it pursue this acquisition to prevent a foreign entity from purchasing Open Solutions?
  2. How well will Fiserv be able to integrate DNA into its current Acumen offering, and will it meet its 24-month target?
  3. Fiserv spent a lot of time discussing the implications for Credit Unions; what should banks currently running on DNA expect to come out of this?
  4. What will happen with DNA vis a vis Fiserv Premier, Precision and Cleartouch?  Might DNA customers be encouraged to move to Premier or Precision?  And might Cleartouch users be migrated to DNA?
  5. What must Jack Henry and Harland be thinking?
Overall, this lifts a potential cloud for any potential Open Solutions customers worried about the firm’s long-term viability and creates new international opportunities for Fiserv.

Are Spanish Bankers More Farsighted?

When Citibank announced they were moving from an internally developed core system to Systematics, I wondered why a bank would take all the trouble of doing a core banking migration, but not moving to a modern real-time system. The answer was that there would be too much operational risk in both moving to a new core system and changing the operations of the bank from batch to real-time. I just attended an announcement from BBVA and Accenture describing the deployment of the Alnova core banking system at BBVA Compass, the US subsidiary. The news is stunning. BBVA Compass has gone live with an overseas, real-time, modern core banking system for all deposit products. While savings and loans and credit unions have been running real-time core systems for years, commercial banks have stayed on batch / memo post. This has huge implications on both the customer experience and back-office operations. Banks no longer have the ability to sort transactions because they are processed as soon as they arrive. This will change overdraft revenues for those banks that sort from largest to smallest transaction before processing in the overnight batch. Processing in real-time also dramatically reduces or eliminates back-office overhead. One and done is the rule. Celent has noticed that credit unions have dramatically lower efficiency ratios than banks of the same asset size, and one of the contributing factors is that the credit unions are running running real-time systems. Please see the Celent report EfficienCU: An Examination of Bank and Credit Union Efficiency Ratios by Asset Tier, November 2011 for more details. BBVA is thinking boldly about the US market. They believe that underinvestment by US banks in modern core banking technology has created an opportunity for them to exploit. Virtually every commercial bank in the United States has a batch memo post system with product siloed (as opposed to customer centric) architecture. Each channel is managed separately and plugs into the core independently. BBVA believes that by creating a business that is customer centric, multichannel, real-time, with straight through processing and related lower cost, they can make even greater inroads into the US banking market. The goal is to have an improvement in their efficiency ratios of 10% in the midterm. This is shy of the efficiency advantage credit unions have over banks:  

Credit unions enjoy an efficiency advantage over banks.

  They already have a top 25 US bank. If BBVA is able to reduce costs through real-time processing and better deployment of cost-saving channels, other banks will be in trouble. If BBVA can also develop customer centric pricing and use that to gain share, other banks will be in deep trouble. This announcement puts Accenture in the cat bird seat. They own the Alnova software asset and have demonstrated the ability to deploy the overseas real-time system in the United States at a commercial bank. This is not easy. There have been notable failures by other vendors in the recent past. Any bank looking to make the leap to a real-time customer centric system will likely look very hard at Accenture. Another bank, famous for its real-time core processing is Santander, BBVA’s Spanish competitor. They have an internally developed core system called Parthenon that they deployed quickly in the UK after acquiring of Abbey National, Alliance & Leicester and Bradford & Bingley. This rapidly drove down operating cost in the UK acquisitions. Santander owns Sovereign Bank in the US which is moving at least in part to the Parthenon system. It seems that these two Spanish banks understand the value that modern technology can play in making a bank cost effective and customer centric. They learn these lessons in Spain, but will be teaching them in the US.  

Oracle Announces a New Core System

Oracle today announced its new core system, Oracle Banking Platform. Why on earth would Oracle develop a new core system apart from its very successful FLEXCUBE? The answer is one size does not fit all. FLEXCUBE is the original bank-in-a-box solution, developed for Citi as an international branch solution, and later known as FLEXCUBE Universal. It has grown from that platform to a second code base FLEXCUBE Retail for mass retail markets. Oracle has been merging the retail and universal banking code bases to a single FLEXCUBE solution. While the solution is a very good fit for international branches and small and midsize banks it doesn’t really have the right architecture for a large bank in its home market. What Oracle has been working with National Australia Bank to develop is the Oracle banking platform, designed for a very large bank in its home market. What specific architectural features are required for such implementation? First and foremost the product needs to be exceptionally modular. A large bank will want to move stepwise with a major core project. That could mean that the bank only once replace a portion of their core system, perhaps customer information file or core deposits. This system will need to harmoniously coexist with the remaining infrastructure at the bank. Alternatively the bank may want to replace the entire core system, but not undertake the risk involved in migrating all in a single fell swoop. This also requires a very modular architecture. Large banks will not settle for best-of-suite solutions, so each module of the Oracle banking platform needs to be able to compete (or integrate) with a best-of-breed solution. Oracle also built this solution from the ground up to optimize implementation on the Oracle stack: Exadata storage, Exalogic processing, Fusion Middleware, and the Oracle database. Surprise! The Oracle Banking Platform is designed with SOA and integration in mind from the start. FLEXCUBE was not. This is clearly a play for the largest few hundred banks in the world in home markets, and was designed with that target in mind.

Foreign Systems in the US Market

Core banking migration is hard, and even harder if you don’t know the local market. Oracle learned that lesson at People’s Bank and Infosys/Finacle learned that at Union Bank. In both cases we had foreign systems being implemented by largely foreign teams into the complex US banking market. I was at an Infosys event where I heard that 99% of all Infosys implementations are on time and on budget. The one I was most familiar with, Infosys deploying Finacle at Union Bank, was not. Union Bank and Infosys crafted an official statement: Over the past several years, Union Bank has sought to drive strategic growth by upgrading and streamlining its enterprise-wide technology platform. Infosys and Union Bank have been working together on an integrated banking platform for the last two years. While the parties have made good progress, Union Bank’s changed business priorities have led the Bank to discontinue the program. Union Bank and Infosys have parted on amicable terms, and Union Bank wishes Infosys all the best in the US market. In both cases, management pulled the plug on the projects for reasons that are not in the public domain, but clearly not because the projects were on time and on budget. Banks don’t like to spend tens of millions of dollars to start projects they can’t finish. I do respect the solutions coming from overseas. As a rule they are far more modern that what the domestic suppliers offer. Yes, there are exceptions, but the vast majority of core banking solutions in the US market today were originally built thirty or forty years ago. Banks face a conundrum: modern systems from overseas or older systems from the US. There are a few exceptions to this and you can read about them in the Celent reports on core banking. Core Banking Solutions for Small Banks: A North American Perspective Core Banking Solutions for Midsize and Large Banks: A North American Perspective Core Banking Solutions for Small Banks: A Global Perspective Core Banking Solutions for Midsize Banks: A Global Perspective Core Banking Solutions for Large Banks: A Global Perspective An interesting upcoming migration is BBVA Compass going to Accenture’s Alnova. We have a modern foreign system, but the SI and owner of the system is Accenture, which has a huge US footprint. I am very hopeful that this will be a successful implementation. I take no joy in calling out core migrations that haven’t succeeded. It is far more pleasant and interesting to write about successful implementations. I wish Accenture and BBVA Compass a smooth and successful migration. It would open the doors to a wave of deferred demand.

Component or All-In-One?

As I work with banks across the globe I see two different philosophies of core banking design: The first is the component philosophy, where banks buy best of breed components and link them together, frequently with point to point integrations, more recently using SOA. The second philosophy is the universal or integrated system that does most of what a bank needs, but not necessarily everything, and not necessarily in the best way possible. I see the pendulum swinging towards integrated systems in many markets. Most of the core banking activities are in developing markets where banks may be starting from near green field installations or have simpler requirements than a top tier bank in a mature market. They choose integrated solutions for their easier implementation and integrations, as well as generally simpler and lower cost operations. These products have matured to offer much stronger functionality than in the past, growing with their existing customers and maturing with further investment. Most of the activity around the best-of-breed components is around the SOA or the glue that links these systems together. Managing point to point integrations is an exponential problem as the complexity of the systems and the number of systems grow. This solution is appropriate for the largest banks in home markets, where product complexity is high and the bank is already well along this path. In this type of environment, finding the best path to integration is the greatest challenge. SOA helps this process along, but making hard business decisions about standardizing data and processes at the bank are what really make this possible. The BIAN initiative and IFX attempt to create standards and frameworks as do a number of technology vendors. To summarize, most of the activity in the integrated space is around improving functionality in the universal system. Most of the activity in the component space is around more easily integrating using SOA and industry frameworks. The analogy to stereo components is not a bad one. Do you want the integrated system? It plays CDs, has a tuner and can fill a small room with decent sound. The features and functionality of these systems are improving smartly over time. If you need to fill a large room with high fidelity sound, integrated with your DVD player to create a home theater, the component system is appropriate. It will cost more, be harder to set up, but deliver superior sound. In these lean times, I think the integrated system will gain ground.

Citi’s core migration

FIS announced that Citi is going to be conducting a core migration and didn’t provide many details. I filled in some of the gaps in an article in American Banker. For those of you who don’t subscribe to American Banker, This is a big deal with Citi moving their entire deposits from an internally developed legacy system to licensed software. Many of the people who developed this internal platform are no longer with Citi, so maintenance and changes are a challenge. I am consistently surprised that banks that conduct a massive core migration are continuing to move to batch systems rather than move to real time systems. Please see the Celent report, Overcoming the Fear Factor: Migrating Core Banking Systems Citi is moving to Systematics, the 30 plus year old mainframe, COBOL, CICS, batch system that runs at Bank of America, RBS Citizens, and lots of other large banks. If a bank is going to all the trouble of moving core systems, why not go all the way and move to real time, to say FIS Profile? To answer my not so rhetorical question, it will involve changes in item processing and increase risk, but there is certainly some reward to go with that. It’s much easier to have such an opinion when I’m not responsible for the conversion and change management. American banks are clearly concerned about adopting the change that a real time system brings. Union Bank (formerly Union Bank of California) is moving to Infosys Finacle, which is a real time system. I eagerly await the results of this migration. Will this start the long-awaited avalanche of core banking migrations? Not yet, but pressure is building. The Citi migration is not a game changer. The Union Bank migration may well be. The Canadians are quite active and there should be some announcements soon. I think that if Union Bank is able to out-innovate the competitors with their new real-time platform and reduce costs, the mass migration will start.