Software defects…a fact of life.

The Badger recently used a bank’s online processes to establish formalised ‘power of attorney’ control over someone else’s accounts. Formalising the ‘power of attorney’ and setting up the associated internet banking facilities was pleasingly easy. Everything went smoothly. This week, however, the Badger encountered a problem. Not a major one, more an inconvenience. The Badger, as ‘power of attorney’, set up a new payee in order to pay a small invoice the same day. However, a ‘technical failure’ error message appeared every time the Badger tried to send the payment. Grr! The Badger called the bank, who were very helpful. It was a known problem – a software defect. If you are a ‘power of attorney’ and click the ‘send payment immediately’ box, the software won’t let you send a payment! The solution? Click the ‘send at a future date’ box – i.e. tomorrow – instead. The solution worked perfectly.

The Badger wondered why this ‘software defect’ hadn’t been picked up in pre-release testing. The experience was also a reminder of how reliant we are on software and on it working correctly. It was also a reminder that software will always contain defects even when the best design, development and testing practices have been used. While the Badger cogitated on this, he saw last week’s reports from the US about the software for Boeing’s reusable spaceship, Starliner. The reports, here for example, highlight a review following the unsuccessful Starliner test flight to the International Space Station(ISS) in December 2019, which has exposed ‘process’ failings in the software design, development, testing and assurance oversight of the ~1 million lines of code. Oh dear. There are obviously many more defects in the software than the ones that impacted the mission in the first place. The Badger raised his eyebrows in surprise. After all, well-established engineering disciplines and processes for producing quality software have been around for a long time and are there for a reason.

Software runs the modern world. It’s everywhere. Its scale and complexity have risen dramatically in recent decades, and when software goes wrong it can have wide ranging, unwelcome, and sometimes disastrous impacts. You can get a sense of the scale of some codebases here and you’ll find some of the software failures that have wreaked havoc and disruption in recent years here. Without software, modern civilization would grind to a halt.

Years ago, the Badger was told ‘Never expect software to be perfect’. Wise words still relevant in today. AI, autonomous vehicles, robots – and so on – are not immune to having software defects, so when you go about your daily life just remember that a software defect is always lurking somewhere, and that it will manifest itself at the most inconvenient time. That’s just a fact of life in today’s world!

The Sillybilly Bank (TSB)…

Mainstream IT services companies wouldn’t be around today if they hadn’t learned lessons from poor delivery over the years. That doesn’t mean their delivery machinery is perfect – far from it – but it does mean they’re generally good at identifying and addressing risk. With 35 years IT delivery under the belt, the Badger’s nose still twitches when he sees, hears, or reads about IT delivery that’s gone wrong. Recently the nose twitched uncontrollably as the Badger caught up on past material about the 2018 TSB IT migration debacle and assimilated TSB’s independent review by Slaughter & May into their disastrous migration from Lloyds to their own systems. The latter has attracted lots of media comment – see, here, here, here and here, for example.

The Badger’s quietly followed the TSB debacle since it happened, labelling the bank as the ‘The Sillybilly Bank’ for the catalogue of failings. Throughout the last 18 months the Badger has always felt the debacle was unlikely to have just a single root cause. There’s been enough signals to suggest that corporate dynamics, financial pressure, poor planning, poor Go-live decision processes, lack of a solid fallback strategy, IT delivery expertise, and – as picked up in the media – poor common sense, all played a part. Future reports from the UK Banking Regulators will hopefully add more colour into the mix and provide more certainty.

In mulling things over, three impressions have come to the fore in the Badger’s mind. Firstly, that TSB’s parent Banco Sabadell – Europe’s ~36th largest bank and ~ 100th in the world – might be guilty of an ‘arrogance of acquisition, we know best’ attitude. They knew the migration was more complex than anything they’d attempted previously and they were warned in 2015 the migration budget was aggressive. Secondly, that Banco Sabadell appears keen to direct all the responsibility for the debacle onto TSB. This smacks of ‘responsibility denial‘ because Banco Sabadell must have endorsed the migration decisions and it was their own IT arm, SABIS, doing the IT. If they didn’t endorse decisions, then surely their corporate governance failed?  The third impression is that the Abilene Paradox was most likely rampant!

One recent piece of commentary neatly says ‘no one comes out of the TSB debacle smelling of roses’, and ‘the whole sorry episode is an example of how not to behave in an overseas takeover’. It’s hard to disagree. So, here’s a question. Would you trust a bank and its parent where there seems to have been governance, risk management, decision- making, and IT failures and the parent points the finger wholly at its subsidiary? You’ll have your own answer. One thing’s certain. When confidence is lost, customers overcome their lethargy and move elsewhere, which, if you look at the switching statistics, is exactly what TSB’s customers have been doing…