Last week at his post-half-year results market briefing, NAB chief Cameron Clyne was asked if the bank’s air-conditioners were up to scratch. The market was still digesting the news that a faulty air-conditioning unit managed to send Westpac’s entire ATM, Eftpos and online banking network down for six hours.
After he realised the question was serious, Clyne’s response was to point to the banks’ “extensive business continuity and disaster recovery plans”, and the work the bank has done improving its systems on the back of its own payment systems glitches.
There’s a problem with that argument, however, and it has to do with how the bank manages disaster recovery and processes payments.
NAB, and as was revealed last week, Westpac, are loath to cut over to disaster-recovery mode. Significant downtime would have to be a definite probability before they would even contemplate the switch.
The reason is most Australian banks still run once-a-day settlement with overnight batch processing, and in disaster recovery mode that can get complicated.
Elsewhere, such as the UK, payments are processed in real-time. Australians still face the inconvenience of their funds magically disappearing into the Bulk Electronic Clearing System before being settled to the recipient’s account, a disappearing act that can take days when a weekend is involved.
It’s time the industry lifted its game, says Jost Stollmann, a former politician who, after moving to Australia from Germany, decided to take on the banks with Tyro, the first new Eftpos solution to emerge in 14 years.
Tyro counts Atlassian co-founder Michael Cannon-Brookes among its board members, and the company is something of a marketing machine, but it’s had a tough ride after being forced to fight the banks tooth and nail for access to the payments system.
Stollman says the issue is not about disaster recovery, but disaster avoidance.
“The settlement processes have to move from once a day overnight to multiple intra-day settlements and ultimately real-time,” says Stollman. “This would de-risk the payment system and offer a level playing field for new entrants.”
A level playing field for new entrants? That’s not the kind of language the big four banks have responded well to in the past.
Meanwhile, Clyne boasts of the increased number of compliment letters he received as a result of how the bank handled recent payments glitches that left thousands of customers out of pocket, in some cases for days, compounded by settlement problems within the bank.
Glitches are a product of 40-year-old legacy systems that industry players are all working to replace, says Clyne, refusing to rule out further downtime.
Customers understand that “things happen in big companies”, says Clyne, it’s how you handle it that matters. There’s an understatement.
One of the things that happens is successive CIOs find ways to avoid spending money, leading to decades of under-investment in technology.
Another thing that happens is outsourcing of IT management to cheaper locations, and contracting out of staff with critical skills to technology giants.
Why do these things happen? Because without new entrants forcing change, banks see no imperative to invest for the future.
NAB is now three years in to its Next Gen project, and last week’s results reveal the bank spent $202 million on IT in the half-year. Clyne calls it a journey. The problem is it’s a journey that should have started a lot sooner.
*This article was originally published at Technology Spectator
I think we were originally promised that I.T. would deliver two things to financial instruments: 1) certainty and 2) speed.
To wind up in a world where the banks use their I.T. systems to deliver the same 3-day cheque clearing behaviour, and lock you out across public holidays is simply insane. I.N.S.A.N.E. We got certainty: Certainly, things are not that much better speed wise.
There is a third thing. we weren’t promised it, but there is no doubt we were led (by the nose?) to beleive it: the I.T. revolution in banking would SAVE US MONEY. And sure, the RBA would be confident that overall, the cost of financials has dropped. And, the cost in bank fee on account terms, has dropped. Win eh?
And where did we wind up on that one? Paying the same costs, but in a different bucket. Every Cardswipe, every EFT, every non-local branch withdraw. We cannot micropay, we can only withdraw cash and are recommended to withdraw wads and wads of it, because if we try and pay $0.15 we’re going to be screwed on a 50c transaction cost, and unknown amounts of pain otherwise.
The cost of electronic funds transfer now bears no relationship at ALL to the price of delivery of that service. none.
Those of you who haven’t worked in telecommunications might be surprised to know that the network behind ATMs, X.25, had a feature in the ‘early’ days which allowed banks to do the entire ATM transaction in what was for those days, NO CHARGE because a call-refused sequence was unbilled, and had enough data space to carry the entire transaction. Thus was born the ‘minicall’ to avoid the pain of admitting the Banks had hacked a free ride off X.25. Takes a bank to learn a trick like that!
How many bits and bytes flow to do a transaction? Amazingly close to the 140 characters in an SMS message. How much does that cost the telco? Well.. if you pay the 50c/SMS charge, those are some of the most expensive bits you buy. Imagine paying 50c/150 chars for a 4gigabyte movie. Same bits. Different charge. By movie scale, by ADSL scale, 140 bytes has vanishingly little cost. Line rental? its all noise.
Right now, the banks are being ripped off, and are ripping us off in turn. Some things they pay for have cost: dual point connect to independent exchanges, dual-trunking, SLA. But you better believe those costs are met from other places. The money we pay, that the RBA tracks, is not paying those costs: Its lining the shareholders pockets.
Its understood that cost and price are disjoint, but there has to be a relationship for at least part of the pair, or we’re in a very bad place. Which we are.
-G
Very interesting article. For more on bank excuses, google ‘account number portability’ and look at the excuses for why the banks don’t want to do it. Legacy computer systems, for which read mainframes, and recoding the likes of which you haven’t seen since y2k.
(yes, I’m a coder, so am not completely unbiased about this 🙂