Tuesday 26 June 2012

The RBS fiasco

Well, it's been a week now, and things are still not back to normal. What was the problem, and how should it be fixed?

My qualification to make these comments are, 1) I don't know anything at all about the RBS systems and 2) I've spent a long time working with computers.


They did an update to their software. The update had a big problem. This meant they couldn't do processing, and after they fixed the problem (at least, I hope they have) it's taking them a long time to catch up with their processing.


First, what is a bank? Essentially, it's a data processing system with a bunch of managers who think they understand about money and risk. There was a time when all banking was done on paper; that's long past. It's all computing now. And so the only, the *ONLY* key activity of the bank, is data processing. So you'd better get that right.

And now, I'd guess that a lot of RBS and Natwest customers are saying "Never again" and moving their accounts to other banks; it's pretty easy to do that. Meanwhile, all the other banks are saying "could this happen here?" and looking into what weaknesses they have in their IT systems.


1) Before you do an update to a key system, test that it works, and have a fall back position, so that you can roll back to the previous version that works.

2) When you do the update, test to see that it's working, don't wait until customers tell you.

How do you do this testing? You have another computer system, with all the same data that you can use for testing. This doubles the cost of your hardware, of course, and the sort of size of hardware that banks use, don't come cheap. But it's cheaper than the RBS fiasco will be.

3) Have competent staff. People are making a bit of a fuss about the fact that RBS was outsourcing to India, and paying a fifth of the salaries. In itself, that's not a problem. What would be a problem, is if the staff employed in this manner are less competent than the more expensive staff they replace.

4) Have three times the processing capacity than you actually need. Again, this means a significant hardware cost, but you'll be glad you have that capacity when you get a big bulge in processing requirements, which is what's happening now.

5) Sack the CEO of the bank, and the head of IT. The converse of getting highly paid for doing a highly skilled and important job, is that you don't get paid when they job is done wrong.

The buck stops at the top.


  1. "The buck stops at the top."

    I agree, but sadly I think those days are behind us. Even if they do get sacked, they do so with their golden parachutes.

    What's Tony Hayward (BP) doing these days, for instance? Enjoying his wealth somewhere, despite the massive fiasco he oversaw. Hell, he went yachting DURING the fiasco. That's the level of accountability at the top these days, I'm afraid.

  2. I think you're right. Barclays seem to have been doing something so naughty, they've been fined £300 million. In this case, it seems to me that not merely the man at the top should get the push, there shoujld be a medium-sized pile of skulls.