What's more fun than a Monty Python sketch? Well, we kicked off the first in a series of Architects' Roundtables recently and... it came close!
Together with CapGemini we hosted a discussion with key banking players about the hot issues facing banks today. Top executives from Tier 1 and Tier 2 banks gathered alongside representatives of UK financial services regulators to take part in the discussion. Not a sequin vest in sight.
The conversation focussed on these key topics: the impact of a move to self-assisted transactions and straight through processing. Starting from the premise that retail banks, and to an extent wealth and universal banks, want to move to as near as possible 100% self-assisted transactions with as near as possible 100% straight through processing, we took a look at whats, whys and hows of this move.
First though, a here's an introduction to the topics.
Self-assisted Transactions (SAT)
Customers like self-assisted transactions, as the bank comes to them rather than them having to go to the bank. This continues the success of ATMs and PoS. Banks like it too as it is estimated that a transaction in the branch done by a teller costs the bank $5, whereas on the Internet it costs $0.50 and on mobile it only costs $0.05 (perhaps explaining the success of mobile banking in east Africa). However, there is another cost to moving to self-assisted transactions. When a teller does the transaction, the ratio of looks to books, that is the number of enquiries compared to the number of updates, is about 5 to 1. This was a rule of thumb for sizing enterprise systems in the 1980s. It applied to clerical workers, tellers and travel agents for instance (though not to Littlewoods Pools, one of my customers in the 80s). When travel agents got disintermediated by booking engines like Expedia and Travelocity, the ratio went from 5:1 to 500:1 and now to 1000:1. We estimate that the Internet bank is at about 50:1 and the mobile bank at about 500:1. If the Internet of Things (IoT) takes off and we start to get fridges ordering milk, the ratio might go to 1000:1. As the bank makes money from the books but not the looks, we can foresee a 100 fold increase in looks for no more books. So we have to re-architect our retail systems to support a 100 fold increase in enquiry for no extra cost to the business.
Straight Through Processing (STP)
STP is the (re-)architecture that potentially holds the answer to banks' and their customers' needs; but it requires a seismic shift in the banking model.
Customers and banks also both like straight through processing. However, they have a slightly different view of it. Both like the idea that there are no people involved in servicing the customers' needs. The banks like it because it dramatically lowers costs. Moving from 99.5% STP to 99.9% STP represents an 80% reduction in people costs in servicing customers!
Bank customers like it because it offers real time service, like they get from Amazon and Spotify.
To deliver STP, the bank has to remove bulk as well as remove people from their processes. It also requires all processes to be real time.
Moving to STP shifts risk from people to machines
In a straight through model of any commercial process, there are necessarily three steps. The first step is the taking of the order, like getting the order slip at Argos. The second step is to exchange risk, like paying for the item at Argos. The third step is change of ownership, like exchanging the till receipt for the item at Argos. Unfortunately, legacy retail systems, like legacy capital markets systems, only did the accounting part. That is why they are called 'branch accounting' systems. There is a difference between record keeping and book keeping.
For retail banks to move to STP they are going to have to fully accommodate record keeping rather than book keeping. The biggest problem for banks is the first step, recording the customer's order. Most banks don't do this at all today. They rely on the batch reconciliation of the accounts to ensure that all they were asked to do has been done. If you log onto the bank, it can't tell you what you asked it to do, when they did it, nor how much they charged. This is fine if you are dealing with tellers in a single branch, but those days are long gone. If you are competing with Google, Amazon, Facebook and Apple (GAFA) you have to remember the orders.
How does a large multi-regional bank optimise costs?
One large Temenos customer is moving from seven country wealth systems, each with its own front office and middle office (or core banking system) to a central core shared by the seven front offices. We expect to see this model more and more as banks attempt to get what another large Temenos customer calls 'value for scale'. The relationship of front office to middle office can easily become a many to many one as banks own lines of business globally and front offices regionally. So the French front office may take loans and deposits from say the German main bank, but take payments and cash management from a separate transaction facility.
This model already exists in wealth management and we expect to see it become the norm across retail and corporate banking too. The front office becomes very good at recording orders, doing round trip portfolio management, and providing high volume low cost access to data
Essentially we are seeing a split between the front office and the middle office, or more colloquially, between distribution and manufacturing. Essentially all banks will move to the wealth model with distribution separate from manufacturing and with a many to many relationship between the two.
Data and regulatory pressure play a major role
The role of data is rapidly changing in banking. Instead of being mainly for marketing, compliance, risk or management information, it is now the differentiator in the customer relationship.
This is as much a cultural and organisational change as it is a technological change. Banks are having to appoint Chief Data Officers to ensure they can balance efficient accommodation of data regulations such as BCBS 239 and Basel III and PSD2, with the implementation of information products as 'best of breed' data initiatives, as this is where the battle for the customer will be fought.
With new technology, a low cost, real time banking information system can be created that addresses both bank and customer needs. This information system can distribute products manufactured by multiple different banks and can consolidate information for many different holdings, making it possible to offer the mass affluent market levels of service that surpass what is currently offered to the ultra-high net worth market. This is the prize banks should be competing for.