The big news at the end of last week is that the Royal Bank of Scotland (RBS) had a computer crash, affecting all of the UK bank including NatWest and Ulster Bank.
It caught headlines worldwide:
- NatWest and RBS computer glitch stretches into third day - Telegraph.co.uk
- NatWest opens on Sunday as IT glitch causes chaos Reuters UK
- Natwest glitch: RBS chief Stephen Hester faces pressure to explain ... Telegraph.co.uk
- RBS Systems Failure Unlikely to Be Resolved Until Monday Bloomberg
- RBS branches open all weekend as customer anger rises - stv.tv
- RBS chief apologises for NatWest banking problems - BBC News
- Stephen Hester apologises to NatWest and RBS customers over ... - Telegraph.co.uk
- Natwest, RBS: When will bank glitch be fixed? Probably not today - Register
- Did RBS Hit the Greece Election Eurogeddon Bank Holiday Panic ... - The Market Oracle
- RBS: NatWest glitch 'is terrible for the bank' - Telegraph.co.uk
… and about 1600 other headlines so far.
It even made the front page of the newspapers:
The issue was caused by an upgrade to the core payments processing engine and a file became corrupted, although it’s interesting how people blamed it on legacy systems.
It’s not legacy systems that cause bank crashes, it’s legacy upgrades.
This is why banks don’t change core systems, because it can create issues and it just goes to show how important resilience and reliability of systems are in banking.
I’ve heard many illustrations of this, but the two best comparisons is that banks are like racing cars where the business is the driver but the technology is the engine, or that banks are like aircraft and charging the engines at 40,000 feet above the ground is hard.
I write about this often myself http://thefinanser.co.uk/fsclub/2011/05/how-the-new-bank-will-win.html:
Wholesale replacement of core systems is difficult you see. Everyone refers to it as being like changing the engines on an aircraft whilst flying at 40,000 feet, and they’re right. This is amply demonstrated by the migration of Abbey and Alliance & Leicester to Santander’s core systems over the past few years, and by the Australian bank debacle of NAB and others who are in migration mode as we speak.
And yet, some seem to achieve these upgrades without hitting the headlines.
For example, Lloyds migrated HBOS to their core system over the summer and Nationwide Building Society has been busy migrating al their systems over to SAP for the past few years.
Yes, they have issues – for a month many business customers of HBOS were unable to use online banking – but it doesn’t get the headlines like a total blackout vis-à-vis the RBS-NatWest issue this week.
This is because the migration is typically carried out over a weekend, as was the RBS upgrade, but during a quieter period – over a bank holiday or during the deep summer holiday – something that RBS could not wait for.
Even though the bank opened branches through the weekend and late into the evenings, and publicly apologised for the mess, the reputational hit is huge and this is where it really matters.
If banks are technology firms, which they are, then they have to realise that five 9’s availability isn’t enough.
It has to be 99.999999999% fault tolerant (or 100% in the ideal world).
Although tough to achieve, any bank that lets its systems fail for even a half a day is subject to ridicule in this world of high tech reliability.
Finally, it demonstrates how banking is different to telecoms and other industries.
If you let a call fail on the mobile network, it doesn’t matter.
The customer will just redial.
If you let a payment fail on the bank network, it always matter.
It brings me back to airlines as the best example of the RBS-NatWest challenge is that if a payment fails once, the customer will stay. A little like if a plane crashes, customers will still fly … they will just be a bit more cautious.
However, if a payment fails twice, three times or on a regular basis, customers leave.
Similarly, if an airline operates planes that regularly crash, customers stop flying.
And that’s the reason why we don’t change the engines on an aircraft at 40,000 feet above ground.