The riskiness of bank tech failure
We all caught sight of RBS NatWest’s failure with their computer glitch over the past few weeks (Ulster Bank customers are still offline – that’s damning).
If you didn’t, the gist of the story is that the bank’s core payment processing system is written with CA-7 software . A software update was applied on 19th June 2012 but the update was corrupted and, as a result, the bank has spent the last three weeks trying to sort out the system and get it working properly.
For more, read the Wikipedia page (now we have Wikipedia pages for the latest news!) or a few other articles:
- RBS boss blames software upgrade - BBC
- RBS computer glitch: Ulster Bank customers still hit by crisis - The Telegraph
- RBS-NatWest proves that technology IS the bank – The Finanser
Interestingly, the Daily Wail said that the problems were all caused by inexperienced operatives in India, although the bank claimed this was not the case. The Register and other media continue to debate the fine details.
Now I’m not going to write much more about this here, as there’s enough of this out there, but rather take a different angle that was kicked off when I received the latest McKinsey Quarterly.
One of this month’s articles caught my eye: “Automating the bank’s back office”.
The article talks about research that reveals a large amount of manual, people processing operations at the heart of banking that can be automated.
It talks about one large universal bank which categorised its processes into three ideal states: fully automated, partially automated, and “lean” manual (McKinsey’s lean program strikes again).
The bank found that 85% of its operations, accounting for 80% of the staff, could be fully or partially automated even though, at the time of the analysis, fewer than 50% of these processes were automated at all. If an ideal level of automation were reached, then almost 50% of staff could be taken out of the system.
Now I’m not one to question McKinsey’s lean IT processing. Plenty of other articles do that (this one is a particularly good read The McKinsey Heresy: Is big bank complexity irreversible?).
I do want to take issue with one thing however, and it stems from the RBS issue but is reflective of every part of our lives today.
The issue is that it is all well and good to automate processes, operations and services.
It is all well and good to have customers do all the work for themselves and self-serve rather than being served.
But where a bank needs to focus its energy is on the contingency plans.
Not just for when the systems go wrong but for when the customer gets it wrong.
The RBS glitch is just a sign of incompetency in internal systems management and governance.
It should not happen, but it did. Now get over it and move on (sorry, Ulster Bank customers, but this will get sorted sometime soon).
However, a bank that does not effectively manage its contingencies gets into the sort of mess that RBS has seen recently (and others have in the past).
Equally, the only time customers get mad is when they do something wrong online and then cannot unravel the error.
They paid money into the wrong account?
Sorry, your problem.
They set up a payment on the wrong day?
Sorry, your problem.
They didn’t read the small print?
Sorry, your problem.
These days, the issue is that everything that can be automated has been, is or will be automated. That means that the human error is now with the customer or with the internal operations lack of effective governance of the system.
Both are important.
The latter as it is the banks reason for existence – to be operationally excellent (again, RBS take note).
The former is just as important however, as this is where a bank can excel.
When the customer messes up, it is the chance for the bank to make the customer feel amazed.
You paid money into the wrong account?
No problem, we’ll sort it out for you.
You paid out on the wrong day?
No problem, we’ll waive the charges this time.
You didn’t read that this could only be cancelled within 14 days?
No problem. I am surprised you don’t read our contracts but, on this one occasion, we will cancel this for you as it’s only 21 days since you signed the agreement.
Too often, banks automate so the only human contact is the customer with the machine and the rules.
It should be the other way around, that when the customer’s contact with the machine has failed that the human contact will change the rules.
This is why first direct consistently get voted the #1 for customer service in UK banking.
Not because they automate everything, but because they have a human approach to automation.
They recognise that their bank is designed for non-physical contact, and therefore any human contact, albeit by telephone, should exceed expectations.
Something that all banks can learn from as they automate their back office service operations.
So to McKinsey’s article: yes, automate the back office customer service, take out the staff, but leave the human ability to recognise when things are not black and white and allow customers a little bit of grey.
And, to RBS’s problems: this is a one-off disaster that’s been bad for the bank. If it ever happened again in the near future, the bank would find customers leaving faster than rats on a sinking ship. That’s why operational excellence is critical and can only be achieved by having contingency plans for contingency plans.
I know the latter sounds preachy, but it’s the core of most other bank disaster recovery and business continuity planning.
Not sure what happened this time, but let’s hope it never happens again.